A High-Precision Crop Classification Method Based on Time-Series UAV Images

Round 1

Reviewer 1 Report

Aim of this study ;presented a method for high-precision crop classification using time-series UAV images. Euclidian Distance  (ED) was utilized to calculate the separability of samples under various vegetation indices,  Co-occurrence Measures and Gray-Level Co-occurrence Matrix (GLCM) were employed to derive texture characteristics, and the spectral and texture features of the crops were successfully  fused, and Random Forest (RF) and other algorithms were utilized to classify crops, and the confusion matrix was applied to assess the accuracy

1-Please explain Euclidian Distance  (ED) ,Co-occurrence Measures and Gray-Level Co-occurrence Matrix (GLCM),Random Forest (RF) 

for this please use this papers ,

Avcı, C. , Budak, M. , Yağmur, N. & Balçık, F. (2023). Comparison between random forest and support vector machine algorithms for LULC classification . International Journal of Engineering and Geosciences , 8 (1) , 1-10 . DOI: 10.26833/ijeg.987605

Sarı, F. & Koyuncu, F. (2021). Multi criteria decision analysis to determine the suitability of agricultural crops for land consolidation areas . International Journal of Engineering and Geosciences , 6 (2) , 64-73 . DOI: 10.26833/ijeg.683754

ofrizal, A. Y. , Sonobe, R. , Hıroto, Y. , Morita, A. & Ikka, T. (2022). Estimating chlorophyll content of Zizania latifolia with hyperspectral data and random forest . International Journal of Engineering and Geosciences , 7 (3) , 221-228 . DOI: 10.26833/ijeg.953188

Kaya, Y. & Polat, N. (2023). A linear approach for wheat yield prediction by using different spectral vegetation indices . International Journal of Engineering and Geosciences , 8 (1) , 52-62 . DOI: 10.26833/ijeg.1035037

2- first please explain UAV and UAV clasification for agriculture aplication. Please add new swction for this before section 2 (before study area)

Sefercik, U. G. , Kavzoğlu, T. , Çölkesen, İ. , Nazar, M. , Öztürk, M. Y. , Adalı, S. & Dinç, S. (2023). 3D positioning accuracy and land cover classification performance of multispectral RTK UAVs . International Journal of Engineering and Geosciences , 8 (2) , 119-128 . DOI: 10.26833/ijeg.1074791

3- Figure 2 must be more clear

4-Figure 4 must be in English or remove 

5- in 2.2.3. Ground survey data . in this section please explain did you make any control point measurement ? how many?

6-Please what is Figure 9 please explain it in text

Author Response

Please see the attachment. All modification instructions are in the file.

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments for author File: Comments.pdf

Author Response

Please see the attachment. All modification instructions are in the file.

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

Author Response File: Author Response.pdf

Reviewer 3 Report

This study presented a method for high-precision crop classification using time-series UAV images. The Co-occurrence Measures and Gray-Level Co-occurrence Matrix (GLCM) were employed to derive texture characteristics, and the spectral and texture features of the crops were successfully fused. Random Forest (RF) and other algorithms were utilized to classify crops, and the confusion matrix was applied to assess the accuracy. There were few studies on crop classification utilizing time-series UAV remote sensing images. However, the classification methods were just common and ordinary ones. Besides, there were also some detailed problems to be solved in the manuscript:

1 Line 172: Two dots.

2 Line 224: “tim-series” should be “time-series”.

3 Line 321: “Pixels” should be deleted.

Author Response

Please see the attachment. All modification instructions are in the file.

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

Author Response File: Author Response.pdf

Reviewer 4 Report

The authors compared the accuracy of different classification algorithms for crop mapping using time-series data. The manuscript suffers from the following drawbacks:

It has no novelty. Comparison between different classification algorithms is a very well-studied subject in remote sensing literature.

A comparison between the UAV images and Sentinel images is not logical. These two modalities have very different spatial resolutions, and clearly, the UAV data will yield higher classification accuracy.

To sum up, all the conclusions drawn by the authors are well-established facts; thus, I cannot suggest this paper for publication.

Author Response

Please see the attachment. All modification instructions are in the file.

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

 manuscript has been sufficiently improved

Author Response

Thank you very much for the expert's approbation of the paper. Thanks again for the expert's professional comments and hard work.

Reviewer 2 Report

The manuscript was greatly improved. However, I still have some comments on the article and suggested some improvements before it is considered to be published.

1. Please standardize the form of author name when citing references. Both abbreviations ([11],[18] et al.) and full names ([12],[20] et al.) are used in the article. Please unify and correct.

2. Line 91-93: “Most of the previous studies were based on simple splicing of remote sensing images to construct time-series data and did not explore the impact of image spectral characteristics on it.” Please cite the appropriate references for the above.

3. The authors used the GLCM method to extract features, but there are no relevant experiments in the article to show the contribution of GLCM. Please add ablation experiments.

Author Response

Response 1: According to the suggestions of expert, the references of the full text have been checked and modified to standardize the author name.

Response 2: We took the expert's advice to the manuscript and quoted references for this statement. At the same time, the language has done a better expression. For details, please see Introduction.

Response 3: The expert advice is fully adopted, and the comparison experiment with single feature classification is added to show the effect of GLCM on classification accuracy. The test results are supplemented in Section 4.1. The experimental results show that the construction of multi-feature space is one of the effective methods to improve the accuracy of crop classification, which is the same as the previous research results cited in Section 3.2.

Thanks again for the expert's professional comments and hard work.

Reviewer 4 Report

I like to thank the authors for answering my previous comments. However, the manuscript still lacks the required novelty of a research paper. The problems of univariate time series classification and object-oriented classification are very well-studied topics in remote sensing literature, and the UAV image's applicability has already been proven. This manuscript, although very informative and valuable for practice, is not as much interesting for researchers in this field. 

Author Response

Thanks again to the expert for your valuable suggestions on the paper. Just as the expert said, the applicability of UAV images has been proven. It is a pity that this article does not meet the more novel criteria in the minds of experts. However, this paper achieves higher classification accuracy by mining image information and learning and screening of multiple features. This can provide some reference for researchers to select crop classification methods. According to the opinions of expert, we will also analyze the challenges of this research in the discussion section, and further research will be carried out in combination with deep learning methods in the future. Please see the discussion section for details. Thanks again for the expert's professional comments and hard work.