Research on Remote Sensing Image Target Detection Algorithm Based on Optical Measurement Information

In recent years, remote sensing information analysis of imaging reconnaissance or earth observation has become an important technical means, and optical RSIs are widely used for target detection (TD) and identification in military intelligence reconnaissance, intelligent traffic monitoring and other fields because of their characteristics of obtaining detailed information of target features. For the acquired massive RSIs, how to obtain the target area of interest and extract the target features from the large scene images, so as to support the application of TD and recognition is still an important research topic. The main objective of this paper is to investigate the TD algorithm of RSIs based on optical measurement information. In this paper, a saliency feature analysis method combining foreground features and background a priori knowledge is investigated, which calculates the boundary probability measure similarity and clusters to generate the background saliency map, introduces the feature weight constraint clustering to generate the foreground saliency map, and guides the weighted fusion to obtain the ship target saliency map after the background saliency map. The experimental demonstration shows that the accuracy of this method is ≥94% and the recall rate is ≥91%, which can effectively eliminate the difficulty of acquiring the ship area when the cloud interference.