Liujun Li

Cropland resources are experiencing substantial pressure due to the combined effects of global climatic shifts and population growth. Meanwhile, the extensive trend of “cropland conversion” has intensified the loss of cropland, threatening food stability. To tackle this issue, remote sensing has emerged as a crucial approach for observing changes in cropland. However, recognizing changes in cropland using high-resolution remote sensing imagery is still a difficult task because of challenges such as large intra-class variation and small inter-class variation of cropland in complex contexts, insufficient extraction of detail at the edges of cropland, and insufficient optimization stability under imbalanced sample conditions. In response to these challenges, we propose an innovative cropland change detection network, IMEA-Net. Firstly, an intelligent weighted wavelet signal extractor (IWWSE) is introduced to extract low-frequency global background and high-frequency local detail features of cropland from remote sensing images using wavelet transform. Secondly, a novel edge-sensitive mamba (ESMamba) is proposed to enhance the detection of fuzzy boundaries of cropland through spatial modeling and structural adjustment. Finally, an adaptive fusion optimization algorithm (AFOA) is introduced to enhance training stability by dynamically adjusting the learning rate. The experimental evaluation confirms that IMEA-Net outperforms seven state-of-the-art (SOTA) approaches on six benchmark datasets: CLCD, LuojiaSET-CLCD, Hi-CNA, JL1, Fuzhou and PX-CLCD. Furthermore, in practical applications, 50 high-resolution image pairs are collected by the Google Earth API. The results demonstrate that IMEA-Net effectively captures cropland changes and shows promising generalization performance.

Tomato, as a globally important economic crop, requires precise and timely disease management to secure yield and quality. Yet segmentation robustness is often limited by weak semantic understanding from single-modality images, narrow receptive fields of convolutional structures, and discontinuous boundary predictions. To address these issues, we propose the Multi-scale Linear Cross-modal Fusion Architecture for Tomato Leaf Disease Segmentation (MS-LCFNet). We construct a real-world field dataset covering five major tomato leaf diseases, annotated by experts with detailed textual descriptions to enable multimodal learning. MS-LCFNet strengthens semantic representation via cross-modal fusion, captures local and global context through an Adaptive Long-short Distance Perception module, and improves boundary continuity with a Physics-informed Smoothness-constrained Loss. Experiments show that MS-LCFNet achieves 87.13 % mIoU on our dataset and 90.78 % on PlantVillage, improving over previous state-of-the-art methods by + 4.62 % and + 4.48 %, respectively, and demonstrating superior accuracy and robustness in complex agricultural scenarios.

This work presents an AI-driven modularized low-cost spot spraying system that uses threshold-based decision-making to dynamically adjust herbicide application in real-time. The system integrates OAK-D for weed detection, Jetson Orin Nano for plant classification, and Arduino Mega for spray control. Simulations demonstrate that this approach significantly reduces herbicide consumption, improves spray precision, and effectively adapts to varying environmental conditions, making it a scalable solution for precision agriculture.

Remote sensing (RS) technology is crucial for monitoring global changes, plays an important role in urban planning, disaster management, and environmental monitoring through building change detection (BCD). High‐resolution RS images used in BCD tasks face challenges such as overlooked derived information, complex backgrounds, sample imbalances, and the selection of an optimal learning rate, complicating their effective utilization. Consequently, the ACSPNet, a Siamese‐architecture BCD network, is introduced. Firstly, an adaptive edge visual feature extraction algorithm is designed to effectively capture architectural edge features, provide important a priori information, and reduce data redundancy and background noise problems. Secondly, coordinated context threshold‐awareness is proposed to enhance the convolutional feature representation through cross‐attention and threshold‐awareness strategies to improve the sensitivity of the model to discriminative features and effectively cope with complex background interference. Subsequently, the self‐calibrating visual field‐enhanced convolution is developed to expand the perceptual range of input features, significantly enhancing the detection of foreground information. This approach sharpens the network's focus on the foreground region and effectively addresses the issue of sample imbalance. Finally, a particle chameleon algorithm is designed to search for the optimal learning rate, thereby accelerating convergence and improving training efficiency. Comparative experiments highlight ACSPNet's superior performance over six state‐of‐the‐art BCD methods across the self‐built dataset (CSUFT‐CD) and three public datasets: Google‐CD, WHU‐CD, and LEVIR‐CD.

Accurate building change detection (BCD) in high-resolution (HR) remote sensing imagery faces persistent challenges including edge misdetection, omission errors, and coarse segmentation. We propose an effective feature-enhanced detection network (DVM-Net) for BCD in HR remote sensing images. This framework addresses these limitations through three key innovations: a dual-channel structure for comprehensive detail extraction, a multi-scale down-sampling module to preserve features and reduce errors, and a visual attention mechanism for refined edge detection. Evaluations include ablation studies on the Wuhan University Building Change Detection Dataset (WHU_CD) and cross-dataset validation using the Google Earth Image Change Detection Dataset (Google_CD) and Large-Scale Earth Vision Infrastructure Change Detection Dataset (LEVIR_CD). DVM-Net achieves state-of-the-art performance, surpassing 11 existing methods with mean Intersection over Union (mIoU) scores of 81.71 %, 79.77 %, and 86.53 % on these benchmarks, respectively. Analysis on the SEmantic Change detectiON Dataset (SECOND) identifies limitations in complex urban environments. The open-source implementation (https://github.com/AroundSeeYou/DVM_Net) provides an advanced solution for HR BCD tasks.

Tomatoes are an important global crop, and automating the segmentation of leaf diseases is essential for agricultural security. Effective segmentation of these diseases is vital for timely intervention, which can significantly enhance crop yield and reduce pesticide usage. However, challenges such as background interference, tiny diseases, and blurred disease edges pose immense obstacles to the segmentation of tomato leaf diseases. To address these issues effectively, we propose a fusion adversarial segmentation network for tomato disease segmentation named FATDNet. Firstly, to eliminate background interference effectively, we introduce a dual-path fusion adversarial algorithm (DFAA). This algorithm employs parallel dual-path convolution to extract features of leaf disease regions, merging and adversarially processing complex background noise and disease features. Secondly, to enhance the feature extraction of small lesion regions, we employ a multi-dimensional attention mechanism (MDAM). This mechanism allocates weights in both horizontal and vertical directions, subsequently calculating weights in different channels of the feature map. This enhances the dispersion of semantic information through the adoption of diverse weight calculation strategies. Furthermore, to improve the model’s ability to extract features at the edges of leaf diseases, we introduce a Gaussian weighted edge segmentation module (GWESM). This module calculates weight distribution through a Gaussian-weighted function, guiding the model to highlight features at different scales and reduce information loss caused by pooling operations. To demonstrate the superiority of FATDNet, we conduct comparative experiments using a self-built dataset and a public dataset. Experimental results show that FATDNet outperforms nine state-of-the-art segmentation networks. This validates that FATDNet provides a reliable solution for the automated segmentation of tomato leaf diseases.

Road cracks pose a serious threat to the stability of road structures and traffic safety. Therefore, this paper proposes an optimized accurate road crack segmentation network called MBGBNet, which can solve the problems of complex background, tiny cracks, and irregular edges in road segmentation. First, multi‐scale domain feature aggregation is proposed to address the interference of complex background. Second, bidirectional embedding fusion adaptive attention is proposed to capture the features of tiny cracks, and finally, Gaussian weighted edge segmentation algorithm is proposed to ensure the accuracy of crack edge segmentation. In addition, this paper uses the preheated bat optimization algorithm, which can quickly determine the optimal learning rate to converge the equilibrium. In the validation experiments on the self‐built dataset, mean intersection over union reaches 80.54% and precision of 86.38%. MBGBNet outperforms the other seven state‐of‐the‐art crack segmentation networks on the three classical crack datasets, highlighting its advanced segmentation capabilities. Therefore, MBGBNet is an effective auxiliary method for solving road safety problems.

Modern human productivity and daily life rely on identifying ground objects using remote sensing images (RSIs). Traditional remote sensing object detection (RSOD) techniques lack timeliness and accuracy and fail to meet practical demands. Existing deep-learning algorithms face continued challenges when processing RSIs because of the diverse shapes and extensive scale variations of objects, of which a significant proportion are small-scale. To address these challenges, we propose the PSWP-DETR, a Transformer-based network that leverages adaptive deformation-learning and multiscale integration for enhanced object detection in remote sensing. First, we propose PradatorConv (PdConv) to address the significant shape changes of objects because it adaptively learns the horizontal and vertical deformations to perceive the complex geometric features of RSIs. Secondly, we propose Scale-wise Differential Modules (SDM), which comprise multi-scale convolution and Edge Captor Convolution (ECC). SDM integrates features across various scales and captures edge characteristics and local textures. This is advantageous for detecting multi-scale objects, tiny objects with limited feature information. Finally, we propose the Whale Particle Optimization (WPO) algorithm for learning rate optimization, which improves convergence speed and accuracy. Experiments using the VisDrone2019-DET, DIOR, and AI-TOD datasets demonstrated that PSWP-DETR achieves the best accuracy benefits, offering significant insights for future RSOD efforts. The source code will be available at https://github.com/Get1star/PSWP-DETR.git.

Image denoising remains a classic and crucial issue in the field of image processing, significantly impacting the outcomes of subsequent image processing tasks. For instance, the denoising network depends on “noise-clean” image pairs to train network effectively. However, it is often hampered by issues such as useful information loss, low training efficiency, and poor blind denoising. To address these challenges, this study proposes a novel image denoising network that integrates the complementary strengths of model-based and learning-based approaches, specifically leveraging the capabilities of both SDDW-GAN and CHRNet. Firstly, SDDW-GAN is designed to estimate the noise distribution on the input noisy images, and a fast-smoothing noisy block sampling algorithm is proposed to extract the noise blocks in noisy images in SDDW-GAN. Secondly, a network with dual generators and dual discriminators based on W-GAN is designed to estimate the noise distribution on the input noisy images and generate noise sample pairs with the same noise distribution, which solves the problem of relying on “noise-clean” image pairs. Thirdly, CHRNet is designed to compute the mapping relationship between the double-noise samples and the single-noise samples. In order to further improve the denoising effect, the dual-channel residual attention module is proposed for fusion learning of global and local features. Experimental results show that the proposed method has a better denoising effect in complex environments and outperforms existing denoising methods. Specifically, in comparison with the stand-alone denoising methods BM3D, DnCNN, Noise2Noise, and Blind2Unblind, the proposed method improves the average peak signal-to-noise ratio (PSNR) by 0.23 dB to 0.78 dB on two benchmarking datasets crossing different noise levels. Its denoising effect is also greater than other competitive stand-alone and combination methods. The proposed method can also extend to low-light image enhancement, deblurring, and super-resolution.