Cloud detection is a key step in optical remote sensing image processing, and the cloud-free image is of great significance for land use classification, change detection, and long time-series landcover monitoring. Traditional cloud detection methods based on spectral and texture features have acquired certain effects in complex scenarios, such as cloud–snow mixing, but there is still a large room for improvement in terms of generation ability. In recent years, cloud detection with deep-learning methods has significantly improved the accuracy in complex regions such as high-brightness feature mixing areas. However, the existing deep learning-based cloud detection methods still have certain limitations. For instance, a few omission alarms and commission alarms still exist in cloud edge regions. At present, the cloud detection methods based on deep learning are gradually converted from a pure convolutional structure to a global feature extraction perspective, such as attention modules, but the computational burden is also increased, which is difficult to meet for the rapidly developing time-sensitive tasks, such as onboard real-time cloud detection in optical remote sensing imagery. To address the above problems, this manuscript proposes a high-precision cloud detection network fusing a self-attention module and spatial pyramidal pooling. Firstly, we use the DenseNet network as the backbone, then the deep semantic features are extracted by combining a global self-attention module and spatial pyramid pooling module. Secondly, to solve the problem of unbalanced training samples, we design a weighted cross-entropy loss function to optimize it. Finally, cloud detection accuracy is assessed. With the quantitative comparison experiments on different images, such as Landsat8, Landsat9, GF-2, and Beijing-2, the results indicate that, compared with the feature-based methods, the deep learning network can effectively distinguish in the cloud–snow confusion-prone region using only visible three-channel images, which significantly reduces the number of required image bands. Compared with other deep learning methods, the accuracy at the edge of the cloud region is higher and the overall computational efficiency is relatively optimal.

Forests play a critical role in global carbon cycling, and continuous forest destruction together with other factors has accelerated global warming. Furthermore, continued decline of forest areas will critically hinder the accomplishment of carbon neutrality goals. Although the geographical location of deforestation can now be rapidly and accurately detected with remote sensing technology, current forest change products are still not fine-grained, especially from the perspective of carbon trading. Here, we used a deep learning method to detect deforestation in large regions based on 2 m high-resolution optical remote sensing images. Firstly, we proposed a new deforestation detection dataset, which was generated from 11 provincial regions in the Yangtze River Economic Zone of China, containing a total number of 8330 samples (the size of each sample being 512 × 512 pixels). Then, a new deforestation detection model, SiamHRnet-OCR, was designed, based on this dataset. Compared with other deep learning models, SiamHRnet-OCR achieves better results in terms of precision, F1-score, and OA indicator: 0.6482, 0.6892, and 0.9898, respectively. Finally, two large-scale scenarios of deforestation experiments in Southern China were further tested; the deforestation detection results demonstrate that SiamHRnet-OCR can not only detect deforestation effectively but also capture the accurate boundary of the changing area.

The quality of optical remote sensing images is largely affected by clouds and haze. In addition, the mosaicking image of multiple remote sensing images, due to objective factors such as acquiring time or climate conditions, will lead to large spectral differences in the area around the seamline. The aforementioned scenarios will seriously affect the accuracy of change detection models based on deep learning. However, there is still a lack of methods to address such issues. To solve these problems, from the perspective of training samples, this manuscript proposed a simple but effective data augmentation method to improve the generalization ability of the deep change detection model in the region of haze cover and the seamline. First, from the characteristics of optical remote sensing image itself, two image simulation methods are designed to conduct data augmentation, named mosaic simulation and haze image simulation. Then, the newly augmented training samples are mixed with the original training samples and then input into a deep learning model for model training. Finally, the change detection results indicate that the proposed data augmentation method can effectively improve the generalization ability of the change detection model in the region of haze cover and seamline, which has high practical value for improving the deep learning model's performance in real-world scenarios and also provides a simple but effective algorithm reference for other intelligent interpretation tasks from the perspective of training data.

Deep-learning-based ship detection methods in Synthetic Aperture Radar (SAR) imagery are a current research hotspot. However, these methods rely on high-quality images as input, and in practical applications, SAR images are interfered with by speckle noise, leading to a decrease in image quality and thus affecting detection accuracy. To address this problem, we propose a unified framework for ship detection that incorporates a despeckling module into the object detection network. This integration is designed to enhance the detection performance, even with low-quality SAR images that are affected by speckle noise. Secondly, we propose a Multi-Scale Window Swin Transformer module. This module is adept at improving image quality by effectively capturing both global and local features of the SAR images. Additionally, recognizing the challenges associated with the scarcity of labeled data in practical scenarios, we employ an unlabeled distillation learning method to train our despeckling module. This technique avoids the need for extensive manual labeling and making efficient use of unlabeled data. We have tested the robustness of our method using public SAR datasets, including SSDD and HRSID, as well as a newly constructed dataset, the RSSDD. The results demonstrate that our method not only achieves a state-of-the-art performance but also excels in conditions with low signal-to-noise ratios.

Multi-class geospatial object detection in high-resolution remote sensing images has significant potential in various domains such as industrial production, military warning, disaster monitoring, and urban planning. However, the traditional process of remote sensing object detection involves several time-consuming steps, including image acquisition, image download, ground processing, and object detection. These steps may not be suitable for tasks with shorter timeliness requirements, such as military warning and disaster monitoring. Additionally, the transmission of massive data from satellites to the ground is limited by bandwidth, resulting in time delays and redundant information, such as cloud coverage images. To address these challenges and achieve efficient utilization of information, this paper proposes a comprehensive on-board multi-class geospatial object detection scheme. The proposed scheme consists of several steps. Firstly, the satellite imagery is sliced, and the PID-Net (Proportional-Integral-Derivative Network) method is employed to detect and filter out cloud-covered tiles. Subsequently, our Manhattan Intersection over Union (MIOU) loss-based YOLO (You Only Look Once) v7-Tiny method is used to detect remote-sensing objects in the remaining tiles. Finally, the detection results are mapped back to the original image, and the truncated NMS (Non-Maximum Suppression) method is utilized to filter out repeated and noisy boxes. To validate the reliability of the scheme, this paper creates a new dataset called DOTA-CD (Dataset for Object Detection in Aerial Images-Cloud Detection). Experiments were conducted on both ground and on-board equipment using the AIR-CD dataset, DOTA dataset, and DOTA-CD dataset. The results demonstrate the effectiveness of our method.

Aircraft detection in SAR images of airports remains crucial for continuous ground observation and aviation transportation scheduling in all weather conditions, but low resolution and complex scenes pose unique challenges. Existing methods struggle with accuracy, overlapping detections, and missed targets. We propose You Only Look Once-SAR Aircraft Detector (YOLO-SAD), a novel detector that tackles these issues. YOLO-SAD leverages the Attention-Efficient Layer Aggregation Network-Head (A-ELAN-H) module to prioritize essential features for improved accuracy. Additionally, the SAR Aircraft Detection-Feature Pyramid Network (SAD-FPN) optimizes multi-scale feature fusion, boosting detection speed. Finally, Enhanced Non-Maximum Suppression (EH-NMS) eliminates overlapping detections. On the SAR Aircraft Detection Dataset (SADD), YOLO-SAD achieved 91.9% AP(0.5) and 57.1% AP(0.5:0.95), surpassing the baseline by 2.1% and 1.9%, respectively. Extensive comparisons on SADD further demonstrate YOLO-SAD’s superiority over five state-of-the-art methods in both AP(0.5) and AP(0.5:0.95). The outcomes of further comparative experiments on the SAR-AIRcraft-1.0 dataset confirm the robust generalization capability of YOLO-SAD, demonstrating its potential use in aircraft detection with SAR.