Liu Xin, Jie Song, Mengmeng Sun, Dazhao Fan, Lv jia yang, Suo ming ze, Rongrong Zhang, Zhen Yan, Li yong jian
The exponential growth of remote sensing satellite deployments globally has exacerbated the imbalance between limited satellite-to-ground downlink capacity and the surging requirement for low-latency, mission-critical data transmission. This pressing issue is driving a transformative shift in remote sensing paradigms, transitioning from traditional “passive data collection with ground-based post-processing” to a novel model of “active sensing and real-time on-orbit processing” facilitated by intelligent satellites. However, there remains a significant deficiency in comprehensive surveys that systematically address on-orbit image processing technologies for intelligent remote sensing satellites, particularly those that provide integrative analyses of system architectures, cutting-edge advancements, and illustrative application scenarios. To address this shortfall, this paper systematically reviews research progress in on-orbit image data optimization and enhancement, as well as intelligent interpretation and thematic product generation technologies, from the perspective of the Layered Collaborative On-orbit Image Processing (LCOIP) framework. It elucidates the supporting role of these technologies in disaster response, national defense security, environmental protection, and agricultural remote sensing applications. Key technical challenges are identified. Furthermore, promising future development directions are explored, such as autonomous intelligent on-orbit processing by single satellites and collaborative on-orbit processing by functionally heterogeneous constellations. This aims to provide theoretical references and technical guidance for the development and application of next-generation intelligent remote sensing satellite systems.