Exploring the Effectiveness of Fusing Synchronous/Asynchronous Airborne Hyperspectral and LiDAR Data for Plant Species Classification in Semi-Arid Mining Areas

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript presents a study on the effectiveness of fusing synchronous/asynchronous airborne hyperspectral and Lidar data on the plant species classification in semi-arid mining areas. Considering that the classification of plant species in semi-arid mining areas is of great significance for assessing the environmental impacts of coal mining and the effect of ecological restoration, I think that the manuscript presents helpful conclusions from the research objectives have been achieved to an extent. Overall, the manuscript is worthy of publication to an extent. However, I have some concerns that should be addressed before the paper could be published.

1. In the Abstract part, it is better for the authors to supplement the concluding remarks in the end.
2. In the Introduction part, it is better for the authors to show the novelty in a clearer way by reviewing and supplementing the related references in the research field.
3. In the Discussion part, the authors must verify the research results by comparing with the others’ research results and supplementing the related references. And the suitability of the results and conclusions of the research should be highlighted and discussed.
4. In the Conclusion part, it is better for the authors to supplement the concluding remarks and the future prospects.
5. It is better for the authors to improve the English language thoroughly by a language company.

Comments on the Quality of English Language

1. It is better for the authors to improve the English language thoroughly by a language company.

Author Response

The manuscript presents a study on the effectiveness of fusing synchronous/asynchronous airborne hyperspectral and Lidar data on the plant species classification in semi-arid mining areas. Considering that the classification of plant species in semi-arid mining areas is of great significance for assessing the environmental impacts of coal mining and the effect of ecological restoration, I think that the manuscript presents helpful conclusions from the research objectives have been achieved to an extent. Overall, the manuscript is worthy of publication to an extent. However, I have some concerns that should be addressed before the paper could be published.

1. In the Abstract part, it is better for the authors to supplement the concluding remarks in the end.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

Quantitative statistical results show that the classification accuracy of plant species in semi-arid mining areas after the fusion of HSI and LiDAR is substantially improved compared to the use of HSI dataset only, and the synchronized collection of airborne HSI and LiDAR data has higher plant classification accuracy compared to asynchronous collection. This study explores the performance differences between airborne HSI and LiDAR fusion for synchronous/asynchronous acquisitions in semi-arid plant classification, which can provide a guarantee for the classification accuracy of HSI and LiDAR fusion features for similar scenes, as well as a reference for error analysis.

These changes are reflected in the revised manuscript on Page 1, Lines 36-44.

2. In the Introduction part, it is better for the authors to show the novelty in a clearer way by reviewing and supplementing the related references in the research field.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

Satellite remote sensing technology has the ability of large-scale coverage, can ob-tain the vegetation information of large mining areas in a relatively short period of time, and is suitable for long-term and large-scale ecological monitoring and assess-ment, but the data resolution is limited and the classification error is large[6]. In recent years, unmanned air vehicles （UAV） have achieved great success in many applica-tion scenarios including vegetation classification in semi-arid areas, and this success is mainly attributed to the flexibility of their mounted sensors and the high resolution of their data [7, 8]. Airborne multispectral sensors are able to capture information about the differences in plant responses in different bands, which can help to distinguish between different species and their health status [9]. Jeong et al.[10] support vector machine model based on multispectral data was able to effectively classify cabbage, rice, and soybean, in which the recursive feature elimination (RFE) model achieved an overall accuracy (OA) of 88.64% with a Kappa coefficient of 0.84. Mao et al. [11] cap-tured images during the flowering period of maize by an airborne Sequoia multispec-tral sensor, which accurately reflected the changes in chlorophyll content, providing technical support for field crop growth monitoring and accurate fertilization decisions. However, the number of bands in multispectral remote sensing is small, which makes it difficult to distinguish vegetation species with similar spectral characteristics. HSI which has richer spectral data in bands, shows great potential in semi-arid areas, and can distinguish the differences in reflectance characteristics of plants in different nar-row bands, improving the accuracy of species identification [12]. Zhan et al. [13] used HSI combined with support vector machine algorithm to classify the vegetation in the target area, and distinguished the vegetation types such as pine forest, broadleaf forest and shrub based on spectral features, which could still maintain high robustness in the case of insufficient samples. Liang et al.[14] proposed a spectral-spatial parallel con-volutional neural network to classify the forest species based on the HSI imagery of UAV and achieved 97.91% OA. neural network to classify forest species, and the OA reached 97.91%. In recent years, the synergistic application of HSI and machine learn-ing algorithms has shown significant technical advantages and great application po-tential in the field of feature classification [15-17]. However, HSI data cannot acquire ground-object vertical distribution information [18]. More and more studies have shown that data fusion of LiDAR data, which can acquire spatial three-dimensional information of ground objects, with HSI data with complementary advantages can achieve better classification accuracy [19-22]. Daeyeol Kim et al. [23] conducted a study on urban-scale tree classification. They extracted the vegetation spectral index using HSI. This approach effectively characterizes the spectral response of chlorophyll, carotenoids, and other biochemical components. Additionally, they used LiDAR to ex-tract the maximum tree height, leaf area index, and symmetry difference index of the canopy structure. These metrics were used to quantitatively characterize the hetero-geneity of the canopy structure and seasonal volume change. Li et al. [24] addressed the issue of uneven distribution of taxonomic demands caused by different wetland mangrove areas. They used HSI data to select key areas, such as the red edge and near-infrared bands, to characterize the spectral feature differences of different plant leaves. Meanwhile, LiDAR data utilized multidimensional features, such as the canopy height model, vertical quartile, and structural heterogeneity, to describe the vertical structure and spatial heterogeneity distribution of vegetation communities. The LiDAR data employed multidimensional features, like the canopy height model, vertical quar-tile, and structural heterogeneity, to characterize the vertical structure and spatial heterogeneity of vegetation communities. The selection of HSI and LiDAR fusion clas-sification features varies greatly for the fine classification needs of features in different regions. For the semi-arid mining areas in China, where trees, shrubs and grasses are mixed and the heterogeneity of spatial distribution is high, how to better integrate the HSI and LiDAR data to realize the fine classification of plants is a hot issue worth studying.

Airborne HSI and LiDAR are mostly along track sensors, and reliable geo-localization of the acquired data presupposes known accurate position and atti-tude information during flight [25]. The position and attitude information of the sen-sors is usually obtained through data post-processing of a tightly combined GNSS/INS navigation system [26-28]. Existing airborne HSI and LiDAR data acquisition can be categorized into two modes, i.e., synchronous and asynchronous acquisition [29]. Syn-chronized HSI and LiDAR acquisition refers to the rigid integration of HSI and LiDAR sensors on the same UAV. For each UAV operation, both HSI and LiDAR data are col-lected simultaneously, and both share the same set of positional data. HSI and LiDAR asynchronous acquisition means that the HSI and LiDAR sensors are mounted on the UAV respectively. Two UAV operations are required to acquire both HSI and LiDAR data separately, each with a separate set of positional data. Due to the continuous high-frequency vibration of the UAV acquisition platform and frequent large-attitude braking during the actual data acquisition process, there are dynamic model perturba-tion errors, observation coarseness, and random sensor errors during the sensor posi-tion data solving process. These factors lead to large differences in the position solving results between multiple flights[30, 31]. Wu et al.[32] showed that the UAV will be af-fected by the aerodynamic force generated by the rotating paddles and its own mass during the flight process, which leads to nonlinear changes in the position parameters of the UAV, and then generates cumulative errors in the navigation solving system. Feng et al.[33] showed that the vibration noise of paddle disk during the flight process of the co-axial UAV will interfere with the accelerometers and gyroscopes, and is prone to drift and spin during the flight process, which leads to errors in data acquisi-tion. In the process of flight, it is easy to appear offset, spin and other phenomena and lead to data acquisition error. Theoretically, the data discrepancy of the position sen-sors will directly lead to the inability to strictly align homonymous features between the HSI and LiDAR data acquired by multiple sorties. At the same time, when collect-ing vegetation information, the HSI and LiDAR data acquired by multiple sorties will also introduce multiple environmental errors. These errors are caused by changes in meteorological factors during the radiometric correction process. This will affect the accuracy of the fusion application of HSI and LiDAR [34].

These changes are reflected in the revised manuscript on Page 2, Lines 59-139.

3. In the Discussion part, the authors must verify the research results by comparing with the others’ research results and supplementing the related references. And the suitability of the results and conclusions of the research should be highlighted and discussed.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

Clarifying the importance of HSI and LiDAR feature variables is crucial for vege-tation classification in semi-arid mining areas[49, 50]. In this study, the order of im-portance of feature variables in the synchronous/asynchronous acquisition group was basically the same, with only minor differences in the order of individual features. Among them, the spectral index features accounted for the majority, proving that the spectral index contains a large amount of information that can be used to differentiate species by category. Among all the spectral indices, the PRI and B30 features associated with the green band, accounted for a higher importance, while the near-infrared band, which is more capable of recognizing species in other studies, was rather insignificant in the classification of plants in this study area [51-53]. The possible reason is that the plant leaf morphology in the study area varies greatly. The sensitivity of the green band to chlorophyll absorption and scattering information is more likely to highlight the differences between vegetation and sparse vegetation, which can better reflect the small changes in the physiological state of vegetation. In contrast, the reflected signals of the near-infrared band are often significantly affected by background interference under sparse canopy conditions. This results in a reduced ability to discriminate be-tween different species. The vegetation in the study area covers a wide range of trees, shrubs and herbs, and the characteristics of canopy height are particularly important. Canopy height provides structural information in distinguishing different vegetation, which can directly reflect the vertical distribution and growth status of vegetation. The experimental data showed that the fusion of HSI and LiDAR improved the classifica-tion accuracy of each vegetation compared with using only HSI data. Among these, the classification accuracy of trees showed the greatest improvement (asynchronous: 15%, synchronous: 23%). Similar to the findings of Picos, J et al. [54]. The reason for this is that trees exhibit a distinct vertical structure, and the inclusion of canopy height fea-tures can effectively reduce the misclassification of trees as shrubs and herbs. In con-trast, the role of plant intensity features acquired by LiDAR was relatively weak, re-lated to the fact that most species in the study area have thinner branches and smaller leaves, and the difference in the reflected intensity of laser light is not obvious.

In this study, we quantified the difference in the classification accuracy of vegeta-tion in the study area between synchronized acquisition of HSI and LiDAR data and asynchronous acquisition of HSI and LiDAR data. As shown in Table 4, the largest dif-ference in identification accuracy occurred in the case of trees, with a difference of 8%, followed by herbaceous plants at 4.5%, and the smallest difference was in the case of shrubs at 3%. Trees show a discrete spatial distribution pattern in the study area, and their morphological characteristics are characterized by small crown widths and sig-nificant heights [55], which are extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. These methods are extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In asynchronous acquisition, an offset of 32.3 cm occurs between the HSI image and the LiDAR point cloud due to two flight position resolution errors. As a result, the spectral reflectance of the top of the tree canopy may be matched with the canopy height data of neighboring low-growth vegetation. This mismatch leads to the failure of feature fusion, reducing classification accuracy. In some cases, the correspondence between the point cloud and the HSI pixels may be en-tirely lost, which affects the correlation between the CHM and spectral features. In the study area, shrub vegetation has a simple and uniform structure with continuous spa-tial distribution. It exhibits low heterogeneity in HSI and LiDAR data, and its growth status and canopy structure changes are small in magnitude [56]. This results in a low difference between the two acquisition modes in terms of data characteristics, making the impact on classification accuracy relatively limited. The accuracy is reduced by only about 3%. Medicago sativa L. had the highest omission rate among the plant spe-cies classification results. The main reason is that Medicago sativa L. has a small canopy radius, the average canopy width is less than 0.5 m, and the branches are extremely thin and the leaf area index is very small, which makes it difficult to accurately cap-ture the canopy height and density information in LiDAR data [57, 58]. Based on the spatial distribution characteristics of Medicago sativa L. and the structural characteris-tics of the species, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data, and the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4% in the experimental data, while that of synchronous fusion of HSI and LiDAR data increased to 47.8% in the experimental data. Based on the spatial dis-tribution characteristics of Medicago sativa L. and the structural characteristics of the species, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In the experimental data, the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4%, while the PA of synchronous fusion of HSI and LiDAR data increased to 47.8%. The important influence of synchronous and asynchronous acquisition on the accuracy of remote sensing monitoring of Medicago sativa L. was re-vealed.

In addition, due to the specificity of vegetation growth conditions in the study ar-ea, the canopy width of trees was generally low and did not differ much from that of shrubs, which could easily lead to the confusion between trees and shrubs [59]. The other vegetation misclassified as trees was dominated by shrubs, and the difference between the canopy widths of trees and shrubs was insignificant. This was the main reason for the decrease in the UA of trees. The tree structure is usually more stratified, and the local leaf area index values may be higher and show a clear vertical gradient, while the spatial distribution of shrub vegetation is continuous, and its leaf area dis-tribution is more uniform. Utilizing the distribution characteristics of LAI in different height strata can help to capture the differences in the internal structure of the two types of vegetation. The height distribution statistics of the point cloud data, such as selecting 90%- and 95%-point cloud data, can more accurately reflect the differences in the vertical structure of the plant canopy, and will be expected to further improve the classification accuracy of arboreal, shrubs and grasses, which is the focus of our next research.

These changes are reflected in the revised manuscript on Page 15, Lines 442-519.

4. In the Conclusion part, it is better for the authors to supplement the concluding remarks and the prospects.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

This study explores the accuracy difference between synchronous and asynchronous LiDAR and HSI data in the application scenario of plant classification in semi-arid mines. And the fusion application of the two has a wide range of application value in the fields of urban planning, agricultural monitoring, environmental protection and resource investigation. In our future research, we will continue to explore the accuracy difference between fused feature classification using synchronously acquired and asynchronously acquired HSI and LiDAR data. This exploration will focus on different application scenarios, monitoring objects, feature extraction methods (e.g., partial least squares), and fusion methods. And we hope that our study can provide valuable theoretical insights and practical frameworks for the scientific community engaged in multimodal remote sensing anal-ysis, particularly for researchers seeking to advance feature classification methodolo-gies through the effective integration of HSI and LiDAR technologies.

These changes are reflected in the revised manuscript on Page 16, Lines 538-549.

5. It is better for the authors to improve the English language thoroughly by a language company.

Response: Thank you for pointing this out. We agree with this comment. We have improved the English language thoroughly by a language company.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript (remotesensing-3547295) by Tian et al. is interesting and potentially suitable for publication in Remote Sensing, pending minor revisions. The hypothesis stated at the end of the introduction needs to be clarified and rephrased as a clear statement that should be either confirmed or refuted in the discussion section. The figures and their captions require improvement—both in terms of image quality and the clarity of the descriptions of elements illustrated. The results section should be strictly descriptive and must not include any elements of discussion. The discussion section is currently inadequate. The authors need to completely revise it, aligning it clearly with the results previously presented. Moreover, the discussion lacks references, which is unacceptable in a scientific context. It should be rewritten carefully and supported by relevant literature. Appropriate statistical analyses should be applied and reported, with significance values presented either in the figures or in a table format. The conclusion should include future perspectives, as well as highlight the main advances of the study and the limitations that still remain.

Author Response

The manuscript (remotesensing-3547295) by Tian et al. is interesting and potentially suitable for publication in Remote Sensing, pending minor revisions.The hypothesis stated at the end of the introduction needs to be clarified and rephrased as a clear statement that should be either confirmed or refuted in the discussion section. The figures and their captions require improvement—both in terms of image quality and the clarity of the descriptions of elements illustrated. The results section should be strictly descriptive and must not include any elements of discussion. The discussion section is currently inadequate. The authors need to completely revise it, aligning it clearly with the results previously presented. Moreover, the discussion lacks references, which is unacceptable in a scientific context. It should be rewritten carefully and supported by relevant literature. Appropriate statistical analyses should be applied and reported, with significance values presented either in the figures or in a table format. The conclusion should include future perspectives, as well as highlight the main advances of the study and the limitations that still remain.

Response: Thank you very much for your valuable comments, we will respond to each of your questions below.

1.The hypothesis stated at the end of the introduction needs to be clarified and rephrased as a clear statement that should be either confirmed or refuted in the discussion section.

Response: Thank you for pointing this out. We agree with this comment. We added a clear statement at the end of the introduction, which is confirmed in the discussion section.

We have made the following revisions in the introduction: Therefore, in this study, we designed two sets of experiments to acquire HSI and LiDAR data synchronously and asynchronously. These experiments were conducted using an airborne HSI and LiDAR integrated monitoring system that shares a combined GNSS/INS navigation module. We also carried out a fusion test of airborne HSI and LiDAR data for the classification of plant species in semi-arid mining areas. The goal was to quantitatively evaluate the impact of fusing synchronous and asynchronous airborne HSI and LiDAR data on the fine classification of plant species in these areas. The impact of fusing synchronized and asynchronous airborne HSI and LiDAR data on fine classification of plant species in semi-arid mining areas was quantitatively evaluated. This paper can provide a scientific basis for the subsequent application of the fusion of airborne HSI and LiDAR data in the classification of plant species in semi-arid mining areas, which has certain theoretical and practical significance.

These changes are reflected in the revised manuscript on Page 3, Lines 140-151.

We have made the following revisions in the discussion: In this study, we quantified the difference in the classification accuracy of vegeta-tion in the study area between synchronized acquisition of HSI and LiDAR data and asynchronous acquisition of HSI and LiDAR data. The largest difference in identifica-tion accuracy occurred in the case of trees, with a difference of 8%, followed by herba-ceous plants at 4.5%, and the smallest difference was in the case of shrubs at 3%. Trees show a discrete spatial distribution pattern in the study area, and their morphological characteristics are characterized by small crown widths and significant heights [55], which are extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. These methods are extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In asynchronous acquisition, an offset of 32.3 cm occurs between the HSI image and the LiDAR point cloud due to two flight position resolution errors. As a re-sult, the spectral reflectance of the top of the tree canopy may be matched with the canopy height data of neighboring low-growth vegetation. This mismatch leads to the failure of feature fusion, reducing classification accuracy. In some cases, the corre-spondence between the point cloud and the HSI pixels may be entirely lost, which af-fects the correlation between the CHM and spectral features. In the study area, shrub vegetation has a simple and uniform structure with continuous spatial distribution. It exhibits low heterogeneity in HSI and LiDAR data, and its growth status and canopy structure changes are small in magnitude [56]. This results in a low difference between the two acquisition modes in terms of data characteristics, making the impact on clas-sification accuracy relatively limited. The accuracy is reduced by only about 3%. Medi-cago sativa L. had the highest omission rate among the plant species classification re-sults. The main reason is that Medicago sativa L. has a small canopy radius, the average canopy width is less than 0.5 m, and the branches are extremely thin and the leaf area index is very small, which makes it difficult to accurately capture the canopy height and density information in LiDAR data [57, 58]. Based on the spatial distribution characteristics of Medicago sativa L. and the structural characteristics of the species, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data, and the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4% in the experimental data, while that of synchronous fusion of HSI and LiDAR data in-creased to 47.8% in the experimental data. Based on the spatial distribution character-istics of Medicago sativa L. and the structural characteristics of the species, it is ex-tremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In the ex-perimental data, the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4%, while the PA of synchronous fusion of HSI and LiDAR data in-creased to 47.8%. The important influence of synchronous and asynchronous acquisi-tion on the accuracy of remote sensing monitoring of Medicago sativa L. was revealed.

These changes are reflected in the revised manuscript on Page 15, Lines 470-505.

2.The figures and their captions require improvement—both in terms of image quality and the clarity of the descriptions of elements illustrated.

Response: Thank you for pointing this out. We agree with this comment. We have improved image quality and the clarity of the descriptions for all illustrations in this paper.

3.The results section should be strictly descriptive and must not include any elements of discussion.

Response: Thank you for pointing this out. We agree with this comment. We removed the discussion element from the results section to ensure that the results section is strictly descriptive.

4.The discussion section is currently inadequate. The authors need to completely revise it, aligning it clearly with the results previously presented. Moreover, the discussion lacks references, which is unacceptable in a scientific context. It should be rewritten carefully and supported by relevant literature.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

Clarifying the importance of HSI and LiDAR feature variables is crucial for vege-tation classification in semi-arid mining areas[49, 50]. In this study, the order of im-portance of feature variables in the synchronous/asynchronous acquisition group was basically the same, with only minor differences in the order of individual features. Among them, the spectral index features accounted for the majority, proving that the spectral index contains a large amount of information that can be used to differentiate species by category. Among all the spectral indices, the PRI and B30 features associated with the green band, accounted for a higher importance, while the near-infrared band, which is more capable of recognizing species in other studies, was rather insignificant in the classification of plants in this study area [51-53]. The possible reason is that the plant leaf morphology in the study area varies greatly. The sensitivity of the green band to chlorophyll absorption and scattering information is more likely to highlight the differences between vegetation and sparse vegetation, which can better reflect the small changes in the physiological state of vegetation. In contrast, the reflected signals of the near-infrared band are often significantly affected by background interference under sparse canopy conditions. This results in a reduced ability to discriminate be-tween different species. The vegetation in the study area covers a wide range of trees, shrubs and herbs, and the characteristics of canopy height are particularly important. Canopy height provides structural information in distinguishing different vegetation, which can directly reflect the vertical distribution and growth status of vegetation. The experimental data showed that the fusion of HSI and LiDAR improved the classifica-tion accuracy of each vegetation compared with using only HSI data. Among these, the classification accuracy of trees showed the greatest improvement (asynchronous: 15%, synchronous: 23%). Similar to the findings of Picos, J et al. [54]. The reason for this is that trees exhibit a distinct vertical structure, and the inclusion of canopy height fea-tures can effectively reduce the misclassification of trees as shrubs and herbs. In con-trast, the role of plant intensity features acquired by LiDAR was relatively weak, re-lated to the fact that most species in the study area have thinner branches and smaller leaves, and the difference in the reflected intensity of laser light is not obvious.

In this study, we quantified the difference in the classification accuracy of vegeta-tion in the study area between synchronized acquisition of HSI and LiDAR data and asynchronous acquisition of HSI and LiDAR data. As shown in Table 4, the largest dif-ference in identification accuracy occurred in the case of trees, with a difference of 8%, followed by herbaceous plants at 4.5%, and the smallest difference was in the case of shrubs at 3%. Trees show a discrete spatial distribution pattern in the study area, and their morphological characteristics are characterized by small crown widths and sig-nificant heights [55], which are extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. These methods are extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In asynchronous acquisition, an offset of 32.3 cm occurs between the HSI image and the LiDAR point cloud due to two flight position resolution errors. As a result, the spectral reflectance of the top of the tree canopy may be matched with the canopy height data of neighboring low-growth vegetation. This mismatch leads to the failure of feature fusion, reducing classification accuracy. In some cases, the correspondence between the point cloud and the HSI pixels may be en-tirely lost, which affects the correlation between the CHM and spectral features. In the study area, shrub vegetation has a simple and uniform structure with continuous spa-tial distribution. It exhibits low heterogeneity in HSI and LiDAR data, and its growth status and canopy structure changes are small in magnitude [56]. This results in a low difference between the two acquisition modes in terms of data characteristics, making the impact on classification accuracy relatively limited. The accuracy is reduced by only about 3%. Medicago sativa L. had the highest omission rate among the plant spe-cies classification results. The main reason is that Medicago sativa L. has a small canopy radius, the average canopy width is less than 0.5 m, and the branches are extremely thin and the leaf area index is very small, which makes it difficult to accurately cap-ture the canopy height and density information in LiDAR data [57, 58]. Based on the spatial distribution characteristics of Medicago sativa L. and the structural characteris-tics of the species, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data, and the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4% in the experimental data, while that of synchronous fusion of HSI and LiDAR data increased to 47.8% in the experimental data. Based on the spatial dis-tribution characteristics of Medicago sativa L. and the structural characteristics of the species, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In the experimental data, the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4%, while the PA of synchronous fusion of HSI and LiDAR data increased to 47.8%. The important influence of synchronous and asynchronous acquisition on the accuracy of remote sensing monitoring of Medicago sativa L. was re-vealed.

In addition, due to the specificity of vegetation growth conditions in the study ar-ea, the canopy width of trees was generally low and did not differ much from that of shrubs, which could easily lead to the confusion between trees and shrubs [59]. The other vegetation misclassified as trees was dominated by shrubs, and the difference between the canopy widths of trees and shrubs was insignificant. This was the main reason for the decrease in the UA of trees. The tree structure is usually more stratified, and the local leaf area index values may be higher and show a clear vertical gradient, while the spatial distribution of shrub vegetation is continuous, and its leaf area dis-tribution is more uniform. Utilizing the distribution characteristics of LAI in different height strata can help to capture the differences in the internal structure of the two types of vegetation. The height distribution statistics of the point cloud data, such as selecting 90%- and 95%-point cloud data, can more accurately reflect the differences in the vertical structure of the plant canopy, and will be expected to further improve the classification accuracy of arboreal, shrubs and grasses, which is the focus of our next research.

These changes are reflected in the revised manuscript on Page 15, Lines 442-519.

5.Appropriate statistical analyses should be applied and reported, with significance values presented either in the figures or in a table format.

Response: Thank you for pointing this out. We agree with this comment. We have incorporated quantitative analysis statistics into the manuscript and presented the significance values in chart format.

6.The conclusion should include future perspectives, as well as highlight the main advances of the study and the limitations that remain.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

This study explores the accuracy difference between synchronous and asynchro-nous LiDAR and HSI data in the application scenario of plant classification in semi-arid mines. And the fusion application of the two has a wide range of application value in the fields of urban planning, agricultural monitoring, environmental protec-tion and resource investigation. In our future research, we will continue to explore the accuracy difference between fused feature classification using synchronously acquired and asynchronously acquired HSI and LiDAR data. This exploration will focus on dif-ferent application scenarios, monitoring objects, feature extraction methods (e.g., par-tial least squares), and fusion methods. And we hope that our study can provide valu-able theoretical insights and practical frameworks for the scientific community en-gaged in multimodal remote sensing analysis, particularly for researchers seeking to advance feature classification methodologies through the effective integration of HSI and LiDAR technologies.

These changes are reflected in the revised manuscript on Page 16, Lines 538-549.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study addresses the issue of plant classification in semi-arid mining areas and proposes a synchronous and asynchronous fusion classification method using airborne HSI and LiDAR data. The classification accuracy under the two data acquisition modes is compared. The paper demonstrates strong innovation and practical value, with an in-depth analysis of data fusion, feature extraction, and classification accuracy evaluation. The research results indicate that integrating HSI and LiDAR data significantly improves plant classification accuracy, especially when using the synchronous data acquisition method, which outperforms the asynchronous method. This study provides effective remote sensing technology support for ecological monitoring and vegetation restoration assessment. The research design is reasonable, the experimental methods are clear, and the data processing is well-regulated. However, some improvements are needed:

1. The Latin names of all plants in the paper contain errors. The complete Latin names should be provided. For example, Prunus pseudocerasus should be written as Prunus pseudocerasus Lindl.. It is recommended to verify the Latin names using this website: https://www.cfh.ac.cn/default.html.

2. The image quality is problematic. Regarding Figure 1, I suggest replacing it with a clear image of an herbaceous plant or shrub. Your paper mentions the classification of Medicago sativa L.. However, based on our past field survey experience, it is difficult to distinguish this species using remote sensing technology unless it is in large patches or during the flowering season.

3. The clarity of all images and the legends is insufficient. In Figure 3, the legend's first letter should be capitalized. In Figures 4, 5, and 6, the legend fonts are unclear, and the figure captions have issues. I recommend that the authors thoroughly adjust and modify the images to improve their clarity and readability.

Author Response

This study addresses the issue of plant classification in semi-arid mining areas and proposes a synchronous and asynchronous fusion classification method using airborne HSI and LiDAR data. The classification accuracy under the two data acquisition modes is compared. The paper demonstrates strong innovation and practical value, with an in-depth analysis of data fusion, feature extraction, and classification accuracy evaluation. The research results indicate that integrating HSI and LiDAR data significantly improves plant classification accuracy, especially when using the synchronous data acquisition method, which outperforms the asynchronous method. This study provides effective remote sensing technology support for ecological monitoring and vegetation restoration assessment. The research design is reasonable, the experimental methods are clear, and the data processing is well-regulated. However, some improvements are needed:

1. The Latin names of all plants in the paper contain errors. The complete Latin names should be provided. For example, Prunus pseudocerasus should be written as Prunus pseudocerasus Lindl. It is recommended to verify the Latin names using this website: https://www.cfh.ac.cn/default.html.

Response: Thank you for pointing this out. We have scrutinized https://www.cfh.ac.cn/default.html and corrected the Latin names of all plants in the paper.

These changes are reflected in the revised manuscript on Page 4, Lines 161-164.

2. The image quality is problematic. Regarding Figure 1, I suggest replacing it with a clear image of an herbaceous plant or shrub. Your paper mentions the classification of Medicago sativa L.. However, based on our past field survey experience, it is difficult to distinguish this species using remote sensing technology unless it is in large patches or during the flowering season.

Response: Thank you for pointing this out. We have improved image quality and the clarity of the descriptions for all illustrations in this paper. Additionally, we specifically modified Figure 1 by incorporating field photographs of dominant shrubs and herbaceous plants in the study area. Regarding the remote sensing identification of Medicago sativa L. raised in your comments, we collected samples of flowering Medicago sativa L. during August in the ecological restoration demonstration area of China's Shendong mining area, which improves its identification potential (as shown in Figure 1(k)). The optical remote sensing data employed consists of airborne hyperspectral data with a ground resolution of 20 cm, significantly contributing to Medicago sativa detection. However, as noted, remote sensing identification of Medicago sativa L. remains challenging. As illustrated in Figure 7, the classification accuracy for Medicago sativa remains suboptimal, with a Producer's Accuracy (PA) of merely 34.8% when using hyperspectral data alone.

These changes are reflected in the revised manuscript on Page 4, Lines 167.

3. The clarity of all images and the legends is insufficient. In Figure 3, the legend's first letter should be capitalized. In Figures 4, 5, and 6, the legend fonts are unclear, and the figure captions have issues. I recommend that the authors thoroughly adjust and modify the images to improve their clarity and readability.

Response: Thank you for pointing this out. We have improved image quality and the clarity of the descriptions for all illustrations in this paper.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

Good attempt but you clearly need to put more work on the value of the hyperspectral data given that it could also cause some artifacts (investigate this and include zoomed-in maps!) in the map output if not properly calibrated esp. mounted in the UAV. I was expecting a higher accuracy given the high quality of remote sensing data (hyperspectral + LiDAR). Another uncertainty source could be the sample geolocation error given that your sample looks dense and the GPS is not RTK. You can try Partial Least Squares and compare to PCA also.  Below my other comments:

 

ou should clearly define Synchronous vs Asynchronous

Line 17-18 is a big statement! How certain are you?

Line 24-25 unnecessary or at least rephrase

What does "fusion based classification test" means?

Why stop at the results in the Abstract? What are the implications of these? and concluding remarks/outlooks?

Line 61 you say the objective at the end of the introduction , not in the middle

Line 258-260 no need to repeat in the results

You didn't tell much about your sampling points and this is very important! Judging by the map (Figure 1) they are not based on probability sampling and this has implications in your mapping. 

Can you include zoomed-in photos of species-level maps? Do you expect to have spatial heterogeneity? 

Looking closely at the maps, I see artifacts (diagonal lines and strong transition between species) most evident in the HSI. Are you sure you calibrated well the hyperspectral sensor?

Principal component sounds good for dimension reduction but have you tried also hyperspectral data-friendly ones like Partial Least Squares? 

I'd appreciate a diagram of the study.

I find it weird that you have used two remote sensing sensors/data (LiDAR and hyperspectral) and yet your accuracy is relatively moderate. 

Which species is most tricky to classify i.e. high omission-commission errors and why

how "synced" your hyperspectral data and LiDAR is in terms of survey date, spatial resolution, altitude etc - how aligned should they be - look for others who did the same "fusion"

Figure 3. better to put in one graph with different colors

Comments on the Quality of English Language

writing style can be improved.. at least they did not use GPT-based text in most parts 

Author Response

Good attempt but you clearly need to put more work on the value of the hyperspectral data given that it could also cause some artifacts (investigate this and include zoomed-in maps!) in the map output if not properly calibrated esp. mounted in the UAV. I was expecting a higher accuracy given the high quality of remote sensing data (hyperspectral + LiDAR). Another uncertainty source could be the sample geolocation error given that your sample looks dense, and the GPS is not RTK. You can try Partial Least Squares and compare it to PCA also. Below are my other comments:

Response: Thank you very much for your valuable comments. We will respond to your review comments one by one in the following content.

1.you should clearly define Synchronous vs Asynchronous.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

Hyperspectral and LiDAR synchronous acquisition refers to the rigid integration of both sensors on a single unmanned aerial vehicle (UAV). During each UAV mission, hyperspectral and LiDAR data are simultaneously collected while sharing identical position and attitude datasets. Asynchronous acquisition involves separate mounting of hyperspectral and LiDAR sensors on UAV platforms, requiring two distinct UAV missions to independently collect hyperspectral and LiDAR data, each accompanied by independent position and attitude datasets.

These changes are reflected in the revised manuscript on Page 3, Lines 117-120.

2. Line 17-18 is a big statement! How certain are you?

Response: Thank you for pointing this out. Our statement is not rigorous enough. The declaration specifically pertains to our study area and cannot be generalized to all semi-arid mining regions. Therefore, we have made the following revisions in the manuscript:

However, in semi-arid mining areas characterized by mixed arbor-shrub-herb vegetation, the complex vegetation distribution patterns and spectral features render single-sensor approaches inadequate for achieving fine classification of plant species in such environments.

These changes are reflected in the revised manuscript on Page 1, Lines 15-18.

3. Line 24-25 unnecessary or at least rephrase.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

There is a lack of precise evaluation regarding how these two data collection approaches impact the accuracy of fine-scale plant species classification in semi-arid mining environments.

These changes are reflected in the revised manuscript on Page 1, Lines 23-25.

4. What does "fusion based classification test" mean?

Response: Thank you for pointing this out. We apologize for the lack of clarity in our original wording. Therefore, we have made the following revisions in the manuscript:

This study designed two experimental scenarios involving synchronous and asynchronous acquisition of HSI and LiDAR data, respectively. We conducted comparative tests on airborne HSI+LiDAR fusion classification for plant species identification in semi-arid mining regions under both acquisition modes.

These changes are reflected in the revised manuscript on Page 1, Lines 25-28.

5. Why stop at the results in the Abstract? What are the implications of these? and concluding remarks/outlooks?

Response: Thank you for pointing this out. We have made these changes to the abstract below in the manuscript:

Quantitative statistical results demonstrate that the fusion of HSI and LiDAR data sig-nificantly enhances plant species classification accuracy in semi-arid mining areas compared to using HSI data alone. Furthermore, synchronously acquired airborne HSI and LiDAR data exhibit superior plant classification accuracy over asynchronously acquired datasets. This study investigates the performance disparities between syn-chronous and asynchronous acquisitions of airborne HSI and LiDAR data in semi-arid plant fusion classification, providing critical references for accuracy assurance and error analysis in HSI-LiDAR fusion-based land cover classification under analogous scenarios.

These changes are reflected in the revised manuscript on Page 1, Lines 36-44.

6. Line 61 you say the objective at the end of the introduction, not in the middle.

Response: Thank you for pointing this out. We have rewritten the introduction in the manuscript and stated the research objectives of this paper at the end of the introduction. Therefore, we have made the following revisions in the manuscript:

Therefore, in this study, we designed two sets of experiments to acquire HSI and LiDAR data synchronously and asynchronously. These experiments were conducted using an airborne HSI and LiDAR integrated monitoring system that shares a com-bined GNSS/INS navigation module. We also carried out a fusion test of airborne HSI and LiDAR data for the classification of plant species in semi-arid mining areas. The goal was to quantitatively evaluate the impact of fusing synchronous and asynchro-nous airborne HSI and LiDAR data on the fine classification of plant species in these areas. The impact of fusing synchronized and asynchronous airborne HSI and LiDAR data on fine classification of plant species in semi-arid mining areas was quantitative-ly evaluated. This paper can provide a scientific basis for the subsequent application of the fusion of airborne HSI and LiDAR data in the classification of plant species in semi-arid mining areas, which has certain theoretical and practical significance.

These changes are reflected in the revised manuscript on Page 3 Lines 140-151.

7. Line 258-260 no need to repeat in the results.

Response: Thank you for pointing this out. We have deleted lines 258-260.

These changes are reflected in the revised manuscript on Page 9, Lines 321-324.

8. You didn't tell much about your sampling points and this is very important! Judging by the map (Figure 1) they are not based on probability sampling, and this has implications in your mapping.

Response: Thank you for pointing this out. Sorry we missed the description of the sample point information. We randomly selected samples of various plant types in the field and accurately positioned each of them using the RTK working mode of Huace i70, with a horizontal positioning accuracy of better than 3 cm. As the study area constitutes a key ecological restoration demonstration zone in China's Shendong mining region, vegetation planting exhibits distinct spatial zoning characteristics, resulting in clustered distribution of sampling points for identical plant species. The absence of sampling points in the central region of Figure 1 is attributed to this area being a concentrated Hippophae rhamnoides Linn. plantation zone, characterized by dense shrubland where manual pedestrian access is infeasible.

These changes are reflected in the revised manuscript on Page 5, Lines 186-189.

9. Can you include zoomed-in photos of species-level maps? Do you expect to have spatial heterogeneity?

Response: Thank you for pointing this out. We have modified Figure 1 to reflect the heterogeneity of the spatial distribution of vegetation in the region as well as species-level photographs of the main plants in the region.

These changes are reflected in the revised manuscript on Page 4, Lines 167.

10. Looking closely at the maps, I see artifacts (diagonal lines and strong transition between species) most evident in the HSI. Are you sure you calculated well the hyperspectral sensor?

Response: Thank you for pointing this out. We have conducted comprehensive calibration of the hyperspectral data in this area, including geometric correction and radiometric correction. The calibration methodology for this regional hyperspectral dataset has been published in Remote Sensing ^[1]。In this article, we proposed a comprehensive radiation distortion correction method that integrates radiation attenuation difference correction, topographic correction, and multi-strip images consistency adjustment (RA-TOC-CA). Artifacts in the hyperspectral imagery primarily originate from two causes. First, as our study area represents a typical ecological restoration demonstration zone in Chinese mining regions, vegetation exhibits spatial zoning in plantation patterns, particularly for arbor species and large shrubs (Figure S1). Second, the push-broom hyperspectral sensor employed lacks a stabilized gimbal, rendering imaging quality highly dependent on the computational accuracy of sensor position and attitude data. Although commercial software was utilized for high-precision position and attitude data processing, sudden acceleration/deceleration of the multi-rotor UAV caused by strong winds during data acquisition introduced localized errors in the resolved sensor position and attitude data. These residual errors ultimately propagate into the hyperspectral remote sensing imagery (Figure S2).

[1] Zhao, Y.B.; Tian, Y.; Lei, S.G.; Li, Y.Y.; Hua, X.; Guo, D.; Ji, C.N. A Comprehensive Correction Method for Radiation Distortion of Multi-Strip Airborne Hyperspectral Images[J]. Remote Sensing. 2023, 15:1828.

11. Principal component sounds good for dimension reduction but have you tried also hyperspectral data-friendly ones like Partial Least Squares?

Response: Thank you for pointing this out. We have previously implemented the Competitive Adaptive Reweighted Sampling (CARS) for hyperspectral data dimensionality reduction in other studies ^[2]. Partial Least Squares has been successfully applied to hyperspectral data processing in alternative regions, demonstrating superior performance. Future investigations will include a comparative analysis of multiple hyperspectral dimensionality reduction methodologies.

[2] Zhao, Y.B.; Lei, S.G.; Yang, X.C.; Gong, C.G.; Wang, C.J.; Cheng, W.; Li, H.; She, C.C. Study on Spectral Response and Estimation of Grassland Plants Dust Retention Based on Hyperspectral Data. Remote Sensing. 2020, 12, 2019.

These changes are reflected in the revised manuscript on Page 16, Lines 544-546.

12. I'd appreciate a diagram of the study.

Response: Thank you for your recognition. We will continue to contribute our efforts in the relevant fields.

13. I find it weird that you have used two remote sensing sensors/data (LiDAR and hyperspectral) and yet your accuracy is relatively moderate.

Response: Thank you for pointing this out. The moderate accuracy observed in our study primarily stems from three factors. First, the position and attitude solution errors mentioned in Question 10 compromised hyperspectral and LiDAR data quality, which will be addressed through integration of three-axis stabilized gimbals in future work. Second, insufficient field sampling points for specific plant species and limited classifier training samples contributed to this limitation. Third, our research focused on quantitatively comparing classification accuracy differences between synchronous/asynchronous HSI-LiDAR fusion, without in-depth exploration of shared dimensionality reduction and classification methodologies. Future investigations will develop optimized dimensionality reduction and fusion classification approaches to fully exploit complementary advantages of HSI and LiDAR data for enhanced plant classification performance.

These changes are reflected in the revised manuscript on Page 16, Lines 544-546.

14. Which species is most tricky to classify i.e. high omission-commission errors and why.

Response: Thank you for pointing this out. Medicago sativa L. had the highest omission rate among the plant species classification results. The main reason is that Medicago sativa L. has a small canopy radius, the aver-age canopy width is less than 0.5 m, and the branches are extremely thin and the leaf area index is very small, which makes it difficult to accurately capture the canopy height and density information in LiDAR data. Based on the spatial distribu-tion characteristics of Medicago sativa L. and the structural characteristics of the spe-cies, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data, and the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4% in the experimental data, while that of synchronous fusion of HSI and LiDAR data increased to 47.8% in the experimental data. Based on the spatial distribution characteristics of Medicago sativa L. and the structural characteristics of the species, it is extremely sensitive to the spatial matching accuracy of HSI and LiDAR data. In the experimental data, the PA of asynchronous fusion of HSI and LiDAR data of Medicago sativa L. was 30.4%, while the PA of synchronous fusion of HSI and LiDAR data in-creased to 47.8%. The important influence of synchronous and asynchronous acquisi-tion on the accuracy of remote sensing monitoring of Medicago sativa L. was revealed.

These changes are reflected in the revised manuscript on Page 15, Lines 490-505.

15. how "synced" your hyperspectral data and LiDAR is in terms of survey date, spatial resolution, altitude etc.- how aligned should they be - look for others who did the same "fusion".

Response: Thank you for pointing this out. In this study, hyperspectral-LiDAR synchronous acquisition denotes rigid integration of both sensors on a single UAV platform, enabling simultaneous data collection with shared position-attitude parameters during each mission. Consequently, both datasets share identical acquisition dates and altitude parameters. Spatial resolution standardization requires manual definition of a unified value followed by individual resampling. For researchers implementing hyperspectral-LiDAR fusion classification, we recommend rigid co-registration of sensors on UAVs for concurrent data acquisition. If unfeasible, high-precision GNSS/IMU integrated navigation modules should be employed to minimize inter-flight position-attitude solution errors. Alternatively, operating at reduced flight altitudes can effectively mitigate horizontal displacement between hyperspectral and LiDAR datasets.

16. Figure 3. better to put in one graph with different colors.

Response: Thank you for pointing this out. We have modified Figure 3.

These changes are reflected in the revised manuscript on Page 5, Lines 198.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have revised the manuscript largely. It is better for the authors to revise the Discussion part of the manuscript by adopting the passive intense rather than such intenses as "we will ". 

Author Response

The authors have revised the manuscript largely. It is better for the authors to revise the Discussion part of the manuscript by adopting the passive intense rather than such intenses as "we will ".

Response: Thank you for pointing this out. We agree with this comment. We have made revision to the discussion of some of the tense issues. Therefore, we have made the following revisions in the manuscript:

In this study, the difference in the classification accuracy of vegetation in the study area between synchronized acquisition of HSI and LiDAR data and asynchronous acquisition of HSI and LiDAR data was quantified.

It is expected that the classification accuracy of arboreal, shrubs, and grasses will be further improved through these methods, which is the focus of our next research.

These changes are reflected in the revised manuscript on Page 15, Lines 460-462, and Page 16, Lines 508-506, respectively.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I thank the authors for addressing my comments. The manuscript has improved considerably and can now be considered for publication.

Author Response

Thank you very much for your recognition of our research work. We wish you good health and success in your endeavors.

Reviewer 3 Report

Comments and Suggestions for Authors

I believe this paper has strong innovative points and is well-written can be accepted.

Author Response

Thank you very much for your recognition of our research work. We wish you good health and success in your endeavors.

Reviewer 4 Report

Comments and Suggestions for Authors

Good improvement, pls attend to these minor suggestions: 

1. Abstract now is relatively long please compress.
2. Figure 7 why red color? Add total accuracy as text in the graph and include in the caption what are the units of the values
3. Discuss more not only advantages but also limitations especially the seldomness of having both data streams (hyperspectral and LiDAR).. When it is very advisable (only for species mapping?) How about practical approaches e.g. UAV-based? Deep learning utility for species classification?  Most researchers lack the budget to have hyperspec + LiDAR

Author Response

Good improvement, pls attend to these minor suggestions:

Response: Thank you very much for your recognition of our research work. We wish you good health and success in your endeavors.

1. Abstract now is relatively long please compress.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

Plant species classification in semi-arid mining areas is of great significance in as-sessing the environmental impacts and ecological restoration effects of coal mining. However, in semi-arid mining areas characterized by mixed arbor-shrub-herb vegeta-tion, the complex vegetation distribution patterns and spectral features render sin-gle-sensor approaches inadequate for achieving fine classification of plant species in such environments. How to effectively fuse hyperspectral images(HSI) data with Light Detection and Ranging (LiDAR) to achieve better accuracy in classifying vegetation in semi-arid mining areas is worth exploring. There is a lack of precise evaluation re-garding how these two data collection ap-proaches impact the accuracy of fine-scale plant species classification in semi-arid mining environments. This study established two experimental scenarios involving synchronous and asynchronous acquisition of HSI and LiDAR data. The results demonstrate that integrating LiDAR data, whether synchronously or asynchronously acquired, significantly enhances classification accu-racy compared to using HSI data alone. The overall classification accuracy for target vegetation increased from 71.7% to 84.7% (synchronous) and 80.2% (asynchronous) respectively. In addition, the synchronous acquisition mode achieved a 4.5 % higher overall accuracy than asynchronous acquisition, with particularly pronounced im-provements observed in classifying vegetation with smaller canopies (Medicago sativa L.:17.4%, Pinus sylvestris var. mongholica Litv.:11.7%, Artemisia ordosica Krasch.:7.5%). This study can provide important references for ensuring classification accuracy and error analysis of land cover based on HSI-LiDAR fusion in similar scenarios.

These changes are reflected in the revised manuscript on Page 1, Lines 13-33.

2. Figure 7 why red color? Add total accuracy as text in the graph and include in the caption what are the units of the values.

Response: Thank you for pointing this out. We agree with this comment. We have made changes to Figure 7.

These changes are reflected in the revised manuscript on Page 13-14, Lines 425-428.

3.Discuss more not only advantages but also limitations especially the seldomness of having both data streams (hyperspectral and LiDAR).. When it is very advisable (only for species mapping?) How about practical approaches e.g. UAV-based? Deep learning utility for species classification?  Most researchers lack the budget to have hyperspec + LiDAR.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made the following revisions in the manuscript:

The airborne hyperspectral and LiDAR used in this study are two cutting-edge and complementary Perceptual sensors. Their integration can not only be applied to plant classification but also holds extensive potential in fields such as precision agriculture and geological and mineral exploration. However, it is regrettable that their substantial costs prevent many researchers from acquiring both sensors simultaneously. This practical limitation reduces the referential value of our research for other scholars.  Additionally, it is important to note that while our findings demonstrate the significant impact of fusing synchronously/asynchronously acquired hyperspectral and LiDAR data on plant classification outcomes, the classification accuracy of their fusion applications remains influenced by numerous factors. These include complex data preprocessing workflows and fusion-classification methodologies. Studies indicate that rigorous geometric correction, radiometric calibration, and point cloud filtering of airborne hyperspectral and LiDAR data can effectively enhance data reliability [60,61]. Moreover, when the plant features extracted from two sensors are sufficiently abundant, fusing them through deep learning methods may achieve higher classification accuracy compared to machine learning. Incorporating the convolutional block attention module (CBAM) into deep learning methods could enhance classifier performance by enabling the model to focus on critical plant classification-related features while suppressing irrelevant noise [62]. Therefore, we recommend that researchers systematically consider potential error sources when working with these costly hyperspectral and LiDAR datasets to ensure the accuracy of their results.

These changes are reflected in the revised manuscript on Page 16, Lines 510-529.

Author Response File: Author Response.pdf