Search Results (18)

Printed circuit board (PCB) defect detection faces challenges like small target feature loss and severe background interference. To address these issues, this paper proposes PCES-YOLO, an enhanced YOLOv11-based model. First, a developed Pre-convolution Receptive Field Enhancement (PRFE) module replaces C3k in the C3k2 module. The ConvNeXtBlock with inverted bottleneck is introduced in the P4 layer, greatly improving small-target feature capture and semantic understanding. The second key innovation lies in the creation of the Efficient Feature Fusion and Aggregation Network (EFAN), which integrates a lightweight Spatial-Channel Decoupled Downsampling (SCDown) module and three innovative fusion pathways. This achieves substantial parameter reduction while effectively integrating shallow detail features with deep semantic features, preserving critical defect information across different feature levels. Finally, the Shape-IoU loss function is incorporated, focusing on bounding box shape and scale for more accurate regression and enhanced defect localization precision. Experiments on the enhanced Peking University PCB defect dataset show that PCES-YOLO achieves a mAP50 of 97.3% and a mAP50–95 of 77.2%. Compared to YOLOv11n, it shows improvements of 3.6% in mAP50 and 15.2% in mAP50–95. When compared to YOLOv11s, it increases mAP50 by 1.0% and mAP50–95 by 5.6% while also significantly reducing the model parameters. The performance of PCES-YOLO is also evaluated against mainstream object detection algorithms, including Faster R-CNN, SSD, YOLOv8n, etc. These results indicate that PCES-YOLO outperforms these algorithms in terms of detection accuracy and efficiency, making it a promising high-precision and efficient solution for PCB defect detection in industrial settings. Full article

Individual tree segmentation is crucial to extract forest structural parameters, which is vital for forest resource management and ecological monitoring. Airborne LiDAR (ALS), with its ability to rapidly and accurately acquire three-dimensional forest structural information, has become an essential tool for large-scale forest monitoring. However, accurately locating individual trees and mapping canopy boundaries continues to be hindered by the overlapping nature of the tree canopies, especially in dense forests. To address these issues, this study introduces CCD-YOLO, a novel deep learning-based network for individual tree segmentation from the ALS point cloud. The proposed approach introduces key architectural enhancements to the YOLO framework, including (1) the integration of cross residual transformer network extended (CReToNeXt) backbone for feature extraction and multi-scale feature fusion, (2) the application of the convolutional block attention module (CBAM) to emphasize tree crown features while suppressing noise, and (3) a dynamic head for adaptive multi-layer feature fusion, enhancing boundary delineation accuracy. The proposed network was trained using a newly generated individual tree segmentation (ITS) dataset collected from a dense forest. A comprehensive evaluation of the experimental results was conducted across varying forest densities, encompassing a variety of both internal and external consistency assessments. The model outperforms the commonly used watershed algorithm and commercial LiDAR 360 software, achieving the highest indices (precision, F1, and recall) in both tree crown detection and boundary segmentation stages. This study highlights the potential of CCD-YOLO as an efficient and scalable solution for addressing the critical challenges of accuracy segmentation in complex forests. In the future, we will focus on enhancing the model’s performance and application. Full article

Although existing 3D object-detection methods have achieved promising results on conventional datasets, it is still challenging to detect objects in data collected under adverse weather conditions. Data distortion from LiDAR and cameras in such conditions leads to poor performance of traditional single-sensor detection methods. Multi-modal data-fusion methods struggle with data distortion and low alignment accuracy, making accurate target detection difficult. To address this, we propose a multi-modal object-detection algorithm, Snow-CLOCs, specifically for snowy conditions. In image detection, we improved the YOLOv5 algorithm by integrating the InceptionNeXt network to enhance feature extraction and using the Wise-IoU algorithm to reduce dependency on high-quality data. For LiDAR point-cloud detection, we built upon the SECOND algorithm and employed the DROR filter to remove noise, enhancing detection accuracy. We combined the detection results from the camera and LiDAR into a unified detection set, represented using a sparse tensor, and extracted features through a 2D convolutional neural network to achieve object detection and localization. Snow-CLOCs achieved a detection accuracy of 86.61% for vehicle detection in snowy conditions. Full article

Selection of sunflower varieties with greater cadmium (Cd) tolerance and detecting physiological variation under different Cd concentrations are important to study the potential of sunflower (Helianthus annuus L.) in the phytoremediation of Cd. The aim of this study was to investigate the variation in the Cd tolerance among four sunflower varieties (79−79, 363, 8361, ADT). Photosynthesis was determined using a Li−6400 XT portable photosynthesis system. Inductively coupled plasma mass spectrometry was used to detect the accumulation of Cd in different plant parts (leaf, stem and root). Subsequently, the Cd amount per plant, bio−concentration factor (BCF), and translocation factor (TF) were calculated. Cd exposure caused a decline in photosynthesis in four sunflower varieties. The 79−79 species displayed the highest Cd concentrations in tissues and 363 displayed a higher BCF in aerial parts under Cd exposure among the four species. Under Cd stress, the total soluble sugars in roots remained unaffected in 363. Based on the results of this experiment, the cultivar of 79−79 and 363 were more tolerant to Cd when compared to the other sunflower cultivars ADT and 8361. The present investigation results indicate that 79−79 and 363 can be further applied in the field trials of phytoremediation practices in contaminated soil. Full article

The current study evaluated the canopy cover competition for light and heat in a jujube–cotton intercropping system to measure the growth and yield performance of cotton, and the optimal cotton planting configuration. In this study, a two-year field experiment (2020 and 2021) was studied with different spacing configuration modes designed as follows: two rows of cotton (CM1) planted 1.4 m apart, four rows of cotton (CM2) planted 1.0 m apart, and six rows of cotton (CM3) planted 0.5 m apart, spacing intercropped jujube trees, respectively. The control treatment consisted of monocultured cotton (CK). The light-response curve was plotted using an LI-6400 XT photosynthesis instrument. Based on the modified rectangular hyperbola model, the photosynthetic characteristics were fitted, and the dry matter distribution characteristics and yield were compared. The results showed that with the increase in photosynthetically active radiation, the net photosynthetic rate (Pn) of each growth phase decreased first and then increased rapidly in the range of 0–200 μmol·m⁻²·s⁻¹ and then decreased slightly after the inflection point (light saturation point). The light-response curves of stomatal conductance and transpiration rate showed a linear relationship. The trend in the intercellular CO₂ concentration response curve was opposite to that of Pn. The maximum Pn (Pmax) of intercropped cotton was significantly impacted by configuration modes, of which CM2 treatment generated 1.8% and 22.8% higher Pmax than the CM1 and CM3 treatments. The cotton yield in the two years ranked as CK > CM3 > CM2 > CM1, and the average land equivalent ratio of CM2 was significantly higher than that of CM3 (22.4%) and CM1 (95.9%). The six-row configuration resulted in greater competition with the trees, which affected the accumulation of below-ground dry matter, while the four-row configuration formed a reasonable canopy structure, which ensured that more photosynthetic substances were distributed to the generative organs. The reasonable four-rows configuration mode may improve the photosynthetic efficiency of intercropped cotton economic yield. Full article

Deep-learning-based multi-sensor hyperspectral image classification algorithms can automatically acquire the advanced features of multiple sensor images, enabling the classification model to better characterize the data and improve the classification accuracy. However, the currently available classification methods for feature representation in multi-sensor remote sensing data in their respective domains do not focus on the existence of bottlenecks in heterogeneous feature fusion due to different sensors. This problem directly limits the final collaborative classification performance. In this paper, to address the bottleneck problem of joint classification due to the difference in heterogeneous features, we innovatively combine self-supervised comparative learning while designing a robust and discriminative feature extraction network for multi-sensor data, using spectral–spatial information from hyperspectral images (HSIs) and elevation information from LiDAR. The advantages of multi-sensor data are realized. The dual encoders of the hyperspectral encoder by the ConvNeXt network (ConvNeXt-HSI) and the LiDAR encoder by Octave Convolution (OctaveConv-LiDAR) are also used. The adequate feature representation of spectral–spatial features and depth information obtained from different sensors is performed for the joint classification of hyperspectral images and LiDAR data. The multi-sensor joint classification performance of both HSI and LiDAR sensors is greatly improved. Finally, on the Houston2013 dataset and the Trento dataset, we demonstrate through a series of experiments that the dual-encoder model for hyperspectral and LiDAR joint classification via contrastive learning achieves state-of-the-art classification performance. Full article

Liver segmentation is a prerequisite for various hepatic interventions and is a time-consuming manual task performed by radiology experts. Recently, various computationally expensive deep learning architectures tackled this aspect without considering the resource limitations of a real-life clinical setup. In this paper, we investigated the capabilities of a lightweight model, UNeXt, in comparison with the U-Net model. Moreover, we conduct a broad analysis at the micro and macro levels of these architectures by using two training loss functions: soft dice loss and unified focal loss, and by substituting the commonly used ReLU activation function, with the novel Funnel activation function. An automatic post-processing step that increases the overall performance of the models is also proposed. Model training and evaluation were performed on a public database—LiTS. The results show that the UNeXt model (Funnel activation, soft dice loss, post-processing step) achieved a 0.9902 dice similarity coefficient on the whole CT volumes in the test set, with 15× fewer parameters in nearly 4× less inference time, compared to its counterpart, U-Net. Thus, lightweight models can become the new standard in medical segmentation, and when implemented thoroughly can alleviate the computational burden while preserving the capabilities of a parameter-heavy architecture. Full article

The availability of commercially available portable photosynthesis systems has facilitated widespread photosynthetic research. Our study aimed to evaluate the reliability of a recently developed inexpensive portable photosynthesis system, Yaxin-1105, in measuring gas exchange in rice (Oryza sativa L.) leaves. Gas exchange parameters, including net photosynthetic rate (A_net), stomatal conductance (g_s), intercellular CO₂ concentration (C_i), and transpiration rate (E), were measured on 88 rice leaves across seven rice cultivars and three growth stages (panicle initiation, heading, and early ripening), using both Yaxin-1105 and LI-6400XT. There were significant difference between each gas exchange parameter at each growth stage measured by Yaxin-1105 and LI-6400XT, except C_i at the heading stage. Mean A_net, g_s, and E measured by Yaxin-1105 were 26–66% lower than those measured by LI-6400XT at panicle initiation, heading, and early ripening stages. Mean C_i measured by Yaxin-1105 was 13–22% higher than that measured by LI-6400XT at panicle initiation and early ripening stages. The coefficients of determination between each gas exchange parameter measured by Yaxin-1105 and by LI-6400XT at panicle initiation, heading, and early ripening stages ranged from only 0.0007 to 0.1889. These results indicate that the Yaxin-1105 is not a reliable tool for measuring gas exchange in rice leaves. Full article

Given its essential role in body functions, liver cancer is the third most common cause of death from cancer, despite being the sixth most common type of cancer worldwide. Following advancements in medicine and image processing, medical image segmentation methods are receiving a great deal of attention. As a novelty, the paper proposes an intelligent decision system for segmenting liver and hepatic tumors by integrating four efficient neural networks (ResNet152, ResNeXt101, DenseNet201, and InceptionV3). Images from computed tomography for training, validation, and testing were taken from the public LiTS17 database and preprocessed to better highlight liver tissue and tumors. Global segmentation is done by separately training individual classifiers and the global system of merging individual decisions. For the aforementioned application, classification neural networks have been modified for semantic segmentation. After segmentation based on the neural network system, the images were postprocessed to eliminate artifacts. The segmentation results obtained by the system were better, from the point of view of the Dice coefficient, than those obtained by the individual networks, and comparable with those reported in recent works. Full article

Kaffir lime is leaf-oriented minor citrus that required extra attention to study. This study aimed to (i) comparatively analyze the young and mature leaf morpho-ecophysiological characters; and (ii) perform a correlation analysis for revealing the relationship among the physiological characters. Plants were ten one-year-old kaffir lime trees cultured under full sun condition. Leaf size was measured by using a specific allometric model. The Li-6400XT portable photosynthesis system was used to observe the leaf ecophysiological characters. The statistical analysis revealed significant differences in leaf size and physiology as the effect of leaf age. A significant size enlargement in mature leaves was noticed, especially in terms of leaf length, area, and weight, of about 77%, 177%, and 196%, respectively. Young leaves experienced a significant improvement in photosynthetic rate and actual water use efficiency for about 39% and 53%, respectively. Additionally, a strong, significant, and positive correlation between leaf chlorophyll, carotenoid content, and photosynthetic rate was found in the present study. Further studies using a multi-omics approach may enrich the science between kaffir lime leaf maturation as the basis of agricultural modification practice. Full article

The effects of low-intensity and high-frequency prescribed burning on the soil CO₂ emissions from Pinus yunnanensis forestland should be explored to achieve sustainable operation and management under fire disturbance. A Li-6400XT portable photosynthesis meter (equipped with a Li-6400-09 soil respiration chamber) and a TRIME^®-PICO 64/32 soil temperature and moisture meter were used to measure the soil CO₂ flux, soil temperature, and soil moisture at fixed observation sites in two treatments (i.e., unburned (UB) and after prescribed burning (AB)) in a Pinus yunnanensis forest of Zhaobi Mountain, Xinping County, Yunnan, China from March 2019 to February 2021. We also determined the relationships between the soil CO₂ flux and soil hydrothermal factors. The results showed that (1) the soil CO₂ flux in both UB and AB plots exhibited a significant unimodal trend of seasonal variations. In 2020, the highest soil CO₂ fluxes occurred in September; they were 7.08 μmol CO₂·m⁻²·s⁻¹ in the morning and 7.63 μmol CO₂·m⁻²·s⁻¹ in the afternoon in the AB treatment, which was significantly lower than those in the UB treatment (p < 0.05). The AB and the UB treatment showed no significant differences in annual soil carbon flux (p > 0.05). (2) The relationship between the soil CO₂ flux and moisture in the AB and UB plots was best fitted by a quadratic function, with a degree of fitting between 0.435 and 0.753. The soil CO₂ flux and soil moisture showed an inverted U-shaped correlation in the UB plot (p < 0.05) but a positive correlation in the AB plot (p < 0.05). Soil moisture was the key factor affecting the soil CO₂ flux (p < 0.05), while soil temperature showed no significant effect on soil CO₂ flux in this area (p > 0.05). Therefore, the application of low-intensity prescribed burning for fire hazard reduction in this region achieved the objective without causing a persistent and drastic increase in the soil CO₂ emissions. The results could provide important theoretical support for scientific implementation of prescribed burning, as well as scientific evaluation of ecological and environmental effects after prescribed burning. Full article

Ex situ conservation plays an important role in maintaining global plant biodiversity and protects thousands of wild plants. Plant conservation in botanical gardens is an important part of ex situ conservation; however, little attention has been given to whether plant ecophysiological traits change and whether plant conservation goals are reached following ex situ conservation. In this study, tree and shrub plants were selected from Shanxi, Beijing of China and from Beijing Botanical Garden, and plants with good growth and similar ages were randomly selected to measure their light response curves, CO₂ response curves with a portable photosynthesis system (Li-6400XT), relative chlorophyll contents using a chlorophyll meter (SPAD-502) and leaf water potential using a dew point water potential meter (WP4C). In comparison with cultivated plants, wild plants had higher water use efficiencies among all plants considered (by 92–337%) and greater light use efficiencies among some of plants considered (by 107–181%), while light response curves and CO₂ response curves for wild plants were either higher or lower compared with cultivated plants. Ecological traits of wild and cultivated plants changed more as a result of habitat factors than due to plant factors. The initial slope of the light response curve, net photosynthetic rate at light saturation, light saturation point, maximum light energy utilization efficiency, maximum water use efficiency, leaf water content, and the leaf water potential of wild plants were larger or equal to those of cultivated plants, while dark respiration rate (by 63–583%) and light compensation point (by 150–607%) of cultivated plants were higher than those of wild plants. This research compared the ecophysiological traits of common green space plants cultivated in botanical gardens and distributed in different areas in wild environments. The response of plant ecophysiological traits to the changing environment has important theoretical and practical significance for wild plant conservation and urban green space system construction. Full article

Segmenting medical images is a necessary prerequisite for disease diagnosis and treatment planning. Among various medical image segmentation tasks, U-Net-based variants have been widely used in liver tumor segmentation tasks. In view of the highly variable shape and size of tumors, in order to improve the accuracy of segmentation, this paper proposes a U-Net-based hybrid variable structure—RDCTrans U-Net for liver tumor segmentation in computed tomography (CT) examinations. We design a backbone network dominated by ResNeXt50 and supplemented by dilated convolution to increase the network depth, expand the perceptual field, and improve the efficiency of feature extraction without increasing the parameters. At the same time, Transformer is introduced in down-sampling to increase the network’s overall perception and global understanding of the image and to improve the accuracy of liver tumor segmentation. The method proposed in this paper tests the segmentation performance of liver tumors on the LiTS (Liver Tumor Segmentation) dataset. It obtained 89.22% mIoU and 98.91% Acc, for liver and tumor segmentation. The proposed model also achieved 93.38% Dice and 89.87% Dice, respectively. Compared with the original U-Net and the U-Net model that introduces dense connection, attention mechanism, and Transformer, respectively, the method proposed in this paper achieves SOTA (state of art) results. Full article

Despite the positive impact of no-tillage (NT) on soil organic carbon (SOC), its potential to reduce soil CO₂ emission still needs enhancing for climate change mitigation. Combining NT with controlled-grazing of crop residues is known to increase nutrient cycling; however, the impacts on soil CO₂ effluxes require further exploration. This study compared soil CO₂ effluxes and SOC stocks from conventional tillage with free grazing (CTFG), NT with free grazing (NTFG), NT without grazing (NTNG), NT without crop residues (NTNR) and NT with controlled-grazing (NTCG), in South Africa. Soil CO₂ effluxes were measured 1512 times over two years using LI-COR 6400XT, once to thrice a month. Baseline SOCs data were compared against values obtained at the end of the trial. Overall, NTCG decreased soil CO₂ fluxes by 55 and 29% compared to CTFG and NTNR, respectively. NTCG increased SOCs by 3.5-fold compared to NTFG, the other treatments resulted in SOC depletion. The increase in SOCs under NTCG was attributed to high C input and also low soil temperature, which reduce the SOC mineralization rate. Combining NT with postharvest controlled-grazing showed high potential to increase SOCs, which would help to mitigate climate change. However, it was associated with topsoil compaction. Therefore, long-term assessment under different environmental, crop, and soil conditions is still required. Full article

Knowledge of the sunshine requirements of landscape plants is important information for the adaptive selection and configuration of plants for urban greening, and is also a basic attribute of plant databases. In the existing studies, the light compensation point (LCP) and light saturation point (LSP) have been commonly used to indicate the shade tolerance for a specific plant; however, these values are difficult to adopt in practice because the landscape architect does not always know what range of solar radiation is the best for maintaining plant health, i.e., normal growth and reproduction. In this paper, to bridge the gap, we present a novel digital framework to predict the sunshine requirements of landscape plants. First, the research introduces the proposed framework, which is composed of a black-box model, solar radiation simulation, and a health standard system for plants. Then, the data fitting between solar radiation and plant growth response is used to obtain the value of solar radiation at different health levels. Finally, we adopt the LI-6400XT Portable Photosynthetic System (Li-Cor Inc., Lincoln, NE, USA) to verify the stability and accuracy of the digital framework through 15 landscape plant species of a residential area in the city of Wuhan, China, and also compared and analyzed the results of other researchers on the same plant species. The results show that the digital framework can robustly obtain the values of the healthy, sub-healthy, and unhealthy levels for the 15 landscape plant species. The purpose of this study is to provide an efficient forecasting tool for large-scale surveys of plant sunshine requirements. The proposed framework will be beneficial for the adaptive selection and configuration of urban plants and will facilitate the construction of landscape plant databases in future studies. Full article