Due to advances in remote sensing satellite imaging and image processing technologies and their wide applications, intelligent remote sensing satellites are facing an opportunity for rapid development. The key technologies, standards, and laws of intelligent remote sensing satellites are also experiencing a series of new challenges. Novel concepts and key technologies in the intelligent hyperspectral remote sensing satellite system have been proposed since 2011. The aim of these intelligent remote sensing satellites is to provide real-time, accurate, and personalized remote sensing information services. This article reviews the current developments in new-generation intelligent remote sensing satellite systems, with a focus on intelligent remote sensing satellite platforms, imaging payloads, onboard processing systems, and other key technological chains. The technological breakthroughs and current defects of intelligence-oriented designs are also analyzed. Intelligent remote sensing satellites collect personalized remote sensing data and information, with real-time data features and information interaction between remote sensing satellites or between satellites and the ground. Such developments will expand the use of remote sensing applications beyond government departments and industrial users to a massive number of individual users. However, this extension faces challenges regarding privacy protection, societal values, and laws regarding the sharing and distribution of data and information.

With the wide vision and high flexibility, unmanned aerial vehicle (UAV) has been widely used into object tracking in recent years. However, its limited computing capability poses a great challenges to tracking algorithms. On the other hand, Discriminative Correlation Filter (DCF) based trackers have attracted great attention due to their computational efficiency and superior accuracy. Many studies introduce spatial and temporal regularization into the DCF framework to achieve a more robust appearance model and further enhance the tracking performance. However, such algorithms generally set fixed spatial or temporal regularization parameters, which lack flexibility and adaptability under cluttered and challenging scenarios. To tackle such issue, in this letter, we propose a novel DCF tracking model by introducing dynamic spatial regularization weight, which encourage the filter focuses on more reliable region during training stage. Furthermore, our method could optimize the spatial and temporal regularization weight simultaneously using Alternative Direction Method of Multiplies (ADMM) technique method, where each sub-problem has closed-form solution. Through the joint optimization, our tracker could not only suppress the potential distractors but also construct robust target appearance on the basis of reliable historical information. Experiments on two UAV benchmarks have demonstrated that our tracker performs favorably against other state-of-the-art algorithms.

Discriminative correlation filters (DCFs) have recently achieved competitive performance in visual tracking benchmarks. However, most of the existing DCF trackers only consider the spatial features of the target and could hardly benefit from the inter-frame and historical information, which may degrade the tracking performance when occlusion and deformation occurs. To tackle the above issues, in this letter, by introducing the temporal constrain into the DCF tracker, we advocate our Spatial-Temporal Context-Aware tracker (STCAT). Through jointly modeling the spatial context and historical target information, our tracker could not only adapt the appearance change but also maintain a relatively stable filter due to the small target variation between inter-frames. Furthermore, we show that the proposed objective formula could be directly solved using ADMM technique with low computational cost. Experiments on the large-scale benchmark demonstrate that the proposed trackers perform favorably against other state-of-the-art methods.

Target detection is a critical task in interpreting aerial images. Small target detection, such as vehicles, is challenging. Different lighting conditions affect the accuracy of vehicle detection. For example, vehicles are difficult to distinguish from the background in RGB images under low illumination conditions. In contrast, under high illumination conditions, the color and texture of vehicles are not significantly different in thermal infrared (TIR) images. To improve the accuracy of vehicle detection under various illumination conditions, we propose an adaptive multi-modal feature fusion and cross-modal vehicle index (AFFCM) model for vehicle detection. Based on the single-stage object detection model, AFFCM uses red, green, blue, and thermal infrared (RGB-T) images. It comprises three parts: 1) the softpooling channel attention (SCA) mechanism calculates the cross-modal feature weights of the RGB and TIR features using a fully connected layer during global weighted pooling. 2) We design a multi-modal adaptive feature fusion (MAFF) module based on the cross-modal feature weights derived from the SCA mechanism. The MAFF selects features with high weight, compresses redundant features with low weight, and performs adaptive fusion using a multi-scale feature pyramid. 3) A cross-modal vehicle index is established to extract the target area, suppress complex background information, and minimize false alarms in vehicle detection. The mean average precision (mAP) on the Drone Vehicle dataset are 14.44% and 5.02% higher than those obtained using only RGB or TIR images. The mAP is 2.63% higher than that of state-of-the-art (SOTA) methods that utilize RGB and TIR images.

Object detection based on remote sensing imagery has become increasingly popular over the past few years. Unlike natural images taken by humans or surveillance cameras, the scale of remote sensing images is large, which requires the training and inference procedure to be on a cutting image. However, objects appearing in remote sensing imagery are often sparsely distributed and the labels for each class are imbalanced. This results in unstable training and inference. In this paper, we analyze the training characteristics of the remote sensing images and propose the fusion of the aggregated-mosaic training method, with the assigned-stitch augmentation and auto-target-duplication. In particular, based on the ground truth and mosaic image size, the assigned-stitch augmentation enhances each training sample with an appropriate account of objects, facilitating the smooth training procedure. Hard to detect objects, or those in classes with rare samples, are randomly selected and duplicated by the auto-target-duplication, which solves the sample imbalance or classes with insufficient results. Thus, the training process is able to focus on weak classes. We employ VEDAI and NWPU VHR-10, remote sensing datasets with sparse objects, to verify the proposed method. The YOLOv5 adopts the Mosaic as the augmentation method and is one of state-of-the-art detectors, so we choose Mosaic (YOLOv5) as the baseline. Results demonstrate that our method outperforms Mosaic (YOLOv5) by 2.72% and 5.44% on 512 × 512 and 1024 × 1024 resolution imagery, respectively. Moreover, the proposed method outperforms Mosaic (YOLOv5) by 5.48% under the NWPU VHR-10 dataset.

Target detection is a critical task in interpreting hyperspectral remote sensing images. Small target (such as airplanes) detection is challenging, especially in large-scale complex scenes with high spectral variability of different land cover types. In this article, we propose a transfer learning-based, large-scale, robust hyperspectral target detector (TLH2TD) to improve the accuracy of hyperspectral target detection (HTD) in large-scale complex scenes. TLH2TD learns the spatial features of hyperspectral targets from high-resolution remote sensing images and achieves high-precision HTD with fused spatial–spectral features. It comprises three parts. First, the coupled target-background sample expansion (CTBSE) module is designed to expand the labeled hyperspectral target and background samples with sufficient high-resolution, labeled RGB images, and a few labeled hyperspectral samples. Second, the hard positive and negative example mining (HPNEM) module trains the hard positive and negative samples to enhance the discriminative ability of the network, addressing the problem of inadequate sample training in large-scale hyperspectral images (HSIs). Third, the spatial–spectral weighted subspace (SSWS) module is designed to fuse the spatial features extracted from the target detection network and the spectral features based on the Mahalanobis distance. The results show that: 1) the TLH2TD achieves average area under the curve (AUC) values of 0.96, 0.96, and 0.93 on small-, medium-, and large-sized HSIs, respectively, achieving the highest accuracy compared with other HTD algorithms; 2) the TLH2TD exhibits the highest detection time efficiency for medium- and large-sized HSIs; and 3) for large-sized HSIs, when most HTD algorithms fail, TLH2TD exhibits significantly higher accuracy and time efficiency than other methods. This is an important achievement in meeting the robust target detection tasks of large-scale hyperspectral remote sensing images.