Satellite Component Recognition Method Based on Improved YOLOv10

Recognizing satellite components is very important for on-orbit service tasks such as space debris removal and orbital refueling, but it is also a challenging task. This paper proposes a satellite component recognition method based on improved YOLOv10 to address the issues of complex lighting variations and dynamic Earth backgrounds in space environments. This method improves the feature selection and representation ability of the model by introducing the ConvLSTM module to improve the recognition accuracy. It also simplifies the model size and computational complexity by using lightweight strategies to improve recognition efficiency. A public dataset and a self-made dataset are used to test the proposed method, and the results demonstrate that the proposed method can effectively recognize satellite typical components such as the satellite body, solar panels, antennas, tripods, nozzles, and payloads, and has good recognition performance.