Target integrity assessment based on image and track information

Abstract

The rapid and accurate integrity assessment of targets can provide important guarantee and reference for the subsequent decision of the implementer. The current research on target integrity assessment mainly adopts single data source or complex probability model, which leads to inability to balance the needs of accuracy and real-time. In order to solve this problem, a new target integrity assessment method based on image and track information is proposed in this paper. Image texture, corner points and track parameters before and after the execution of the target are used to transform the integrity assessment issue into a classification issue, and a comprehensive assessment result is made by combining various classification results. The experimental results show that the assessment results based on both image and track information reached 97.5%, higher than the evaluation results from a single data source, and the evaluation time was controlled in milliseconds, which not only improved the accuracy rate but also ensured the real-time assessment.

Keywords

• targets integrity assessment
• knn
• texture feature
• image classification
• track information