The Gaofen-1 (GF-1) wide-field-view (WFV) sensor provides valuable Earth observation (EO) data, but its limited geometric accuracy hinders applications requiring precise geolocation. This study presents a novel and practical on-orbit geometric recalibration method specifically designed for satellite images lacking a rigorous sensor model. Our approach uses a rowwise rational polynomial coefficient (RPC) refinement model and leverages well-distributed ground control points (GCPs) within single rows of calibration images to estimate and correct charge-coupled device (CCD) distortions using thin plate spline (TPS) interpolation. Validation results demonstrate a significant improvement in geometric accuracy, achieving a circular error 90% (CE90) of approximately 0.4 pixels for GF-1 WFV, highlighting the method’s effectiveness and broad applicability in remote sensing.

Sharpening a single satellite image remains challenging due to low computational efficiency, complexity of multiparameters, unphysical modeling, and the potential for radiometric consistency loss. To address these issues, this article introduces a modulation transfer function (MTF)-based sharpening method that is fast, has a single tunable parameter, and effectively suppresses noise and over-enhancement. This article also proposes an automatic method for extracting edge objects with any angle for MTF calculation, without relying on ideal edge objects. The improved slanted-edge method is more robust against noise by incorporating the logistic function and employing the random sample consensus (RANSAC) algorithm to remove deflected edges. The new 2-D MTF estimation method provides precise and stable sharpening results. This article extends the proposed method to single image super-resolution (SISR) for satellite images. The proposed approach outperforms state-of-the-art SISR methods, including 11 deep learning-based methods, across three public datasets and raw images (water, city, and building) acquired from three satellites. The utmost correlation to the histogram of raw image proves the proposed method’s superiority in preserving radiometric information compared to other methods. In addition, the successful application of the one-time estimated 2-D MTF for raw satellite images over a year and its capability to improve edge sharpness uniformity across cameras within the sensor system further solidify the method’s universality and reliability. More comparison results and code are available athttps://github.com/RSingKK/Any-angle-MTF.

Forests are the most widely distributed terrestrial vegetation type and play a significant role in the global carbon cycle and ecological diversity. Accurate and timely forest detection provides essential data for forest management and development. Current forest-related products differ in definition, accuracy, and spatial consistency, making them difficult to use. Therefore, it is necessary to map forest cover under a unified framework. However, detecting forests on a large scale requires high-quality and representative samples, which can be challenging. This study proposes a weakly supervised forest classification framework (WSFCF) that uses noisy labels. The WSFCF is designed to address label generation, correction, and sample location optimization. We employ a spectral-spatial network to extract forest cover accurately for medium-resolution forest classification. The experimental results show that the proposed method outperforms the compared methods, achieving an accuracy of 91.76% OA and 88.28% F1 score on 110 GF-1 WFV images. This supports the subsequent extraction of national-scale forest cover and encourages the mapping of China’s forest cover using GF-1 WFV images. Moreover, the proposed method produces satisfactory outcomes for objects such as water, farmland, and built-up areas within the study area, demonstrating its effectiveness and potential for transferability.

Plastic greenhouses (PGs), as an important part of modernized facility-based agriculture, are widely used worldwide to improve crop yield. To facilitate precision agricultural management and environment assessment, there has been widespread research attention that focuses on accurately mapping the spatial distribution and exploring coverage of PGs via remote sensing and deep learning techniques. However, limited by PGs’ poor intraclass similarity and interclass separability in high-resolution remote sensing images (HRRSIs), existing methods for PGs extraction are hard to achieve promising performance. Besides, their visualization results also suffer from adhesion problems in densely distributed areas, due to the lack of attention to geometric and boundary information. Therefore, this article proposes a novel edge-guided dual stream network (EDSNet), which consists of an edge detection branch, a body extraction branch, and an adaptive fusion module. In the body extraction branch, a new feature extractor, termed geometric-refined attention module (GRAM), is embedded to enhance the geometric information of PGs. Moreover, an edge-guided module (EGM) is introduced to guide the edge detection branch to gradually recover precise boundary details. Finally, the adaptive semantic fusion module (ASFM) adaptively fuses enhanced edge features and body features to maximally facilitate the information interaction between dual branches, thus improving interclass separability on both sides of edges while suppressing meaningless information outside bodies. The proposed EDSNet is validated on self-labeled and publicly available datasets, and the results demonstrate that this method performs well and is suitable for accurate extraction of PGs. The code will be available athttps://github.com/xchouzhang/EDSNet.

Rapid progress has been made in the research of high-resolution remote sensing road extraction tasks in the past years, but due to the diversity of road types and the complexity of road context, extracting the perfect road network is still fraught with difficulties and challenges. Many Convolutional Neural Networks (CNNs) based on encoder-decoder structures have demonstrated their effectiveness. Transformer’s self-attention mechanism shows more powerful performance than CNNs in modeling global feature dependencies. In this paper, we propose a Multi-scale Deformable Transformer Network (MDTNet) based on encoder-decoder structure to extract road networks from remote sensing images. The core of MDTNet is our proposed Multi-scale Deformable Self-Attention (MDSA) mechanism. MDSA can capture more comprehensive features than conventional self-attention. In addition, roads are not present in certain blocks of areas like other objects, but are interwoven throughout the image in such a long, linear fashion that information about certain road segments may be overlooked. To minimize residual errors in road segmentations, our MDSA incorporates a deformable design on feature maps, which effectively enhances the salience of road features relative to their surroundings. Extensive experiments on several public remote sensing road datasets show that our MDTNet achieves higher segmentation [F1 score and Intersection over Union (IoU)] and connectivity [Average Path Length Similarity (APLS)] accuracy, which verifies the effectiveness of our approach.

The development of deep learning in remote sensing (RS) visual tasks has led to remarkable progress in RS image change detection (CD). However, RS bi-temporal images cover complex and confusing scenes due to natural environmental factors, which presents challenges for CD task. How to effectively exploit long-range dependencies and sensitively discriminate real-changes with various scales from pseudo-changes are urgent problems. It is especially obvious for the changes of building structures man-made. This paper presents a CD approach named CSViG, which utilizes Siamese Vision Graph neural network (SViG) with crossed feature fusion. SViG acts as a feature extractor to capture richer short- and long-range dependencies. Crossed feature fusion consists of a horizontal feature fusion module (HFFM) and a vertical feature fusion module (VFFM). HFFM designs cross-concatenation (CC) way to reveal real-changes from pseudo-change in the same horizontal stage, after which global and local features are extracted by using attention mechanism and multi-scale depth-wise separable convolution. VFFM further fuses complementary content from vertical multiple stages to effectively represent change regions of different sizes (tiny or huge) by using attention mechanism. Extensive comparative experiments conducted on three available building change detection datasets demonstrate that the proposed method achieves better CD performance than previous counterparts.

The impervious surface (IS) cover and its dynamics are key parameters in research about urban and ecology. This letter proposed an automatic framework to map the IS growth end-to-end based on the temporal deep learning (DL) model and long time-series Landsat imagery. First, the training and validating datasets were auto-generated by a joint strategy. Then, a DL network was designed, and the IS growth was predicted in temporal windows. Finally, the results from multi-temporal windows are combined to generate the IS growth map. The data around the core of Beijing, China, is tested, and the result shows that the proposed method could: 1) efficiently model the IS growth; 2) map IS growth with less salt-and-pepper noise and false alarm compared to existing products; and 3) be extended to future data easily.

Building detection from high-resolution remote sensing images is one of the tasks of remote sensing information extraction. Due to the particularity of the radiation quantization value and spatial resolution of Gaofen-2 satellite, the shape of buildings cannot be clearly represented in many scenes, so it's difficult to realize building detection in complex scenes. This paper made improvements based on the instance segmentation network CenterMask. Because the shapes of the building is irregular, traditional convolution cannot express spatial deformation well. So we used deformable convolution to replace the traditional convolution. In the attention mechanism, channel-dimensional attention can highlight useful channels and weaken redundant channels. The spatial attention can capture long-distance dependence through global context information. We added a channel attention mechanism to the segmentation branch. The feature map passes through the channel attention mechanism and the spatial attention mechanism successively. Experimental results show that the improved model can achieve good results in building detection tasks under complex scenes.

As the direct geo-locating accuracy of spaceborne optical images is limited by the uncertainty of the exterior orientation parameters, precise ground control points (GCPs), which are difficult or expensive to obtain, are commonly required to improve the geometric accuracy in practical applications. In this article, we propose a novel block adjustment (BA) method to make use of the GCPs automatically collected from reference images of coarse resolution (Landsat-8 or Sentinel-2), which are publicly available. Different from the conventional BA methods, the proposed one treats the GCPs of low accuracy as relaxed constraints instead of directly minimizing the error between the geometric models and GCPs, and only guarantees that the GCPs are satisfied by the geometric models with a prior accuracy. In addition, an automated method is introduced to estimate the proper GCP accuracy for the proposed BA. The experimental results of three testing sites in China using three different types of spaceborne images, i.e., Gaofen-1 (GF-1) panchromatic (PAN), ZY-3 nadir (NAD), and SPOT-5 high resolution geometric (HRG) whose spatial resolutions are around 2 m, show that the accuracy of 1-2 pixel can be achieved for these high-resolution images when only coarse reference images (spatial resolution of 15 and 10 m) were used as ground control. The results also show that the inaccurate GCPs are not likely to undermine the geometric consistency of images in the proposed BA, and BA with relaxed constraints can even achieve better tie points (TPs) accuracy than BA without ground control. This article provides a practical way to utilize inaccurate ground control and balance the tradeoff between GCPs and TPs.