Anti-interference aerial target tracking for infrared seeker via spatiotemporal correlation of topological graph networks

The bottleneck of artificial interference in the background of complex air combat is the major obstacle in the precise detection and guidance of infrared air-to-air missiles. Occlusion and similarity of infrared point-source interference to aerial targets pose a serious challenge to the anti-interference tracking performance of an infrared imaging seeker. Based on the analysis of relative motion of the target and interference and the spatial topological structure of multi-object in the image plane, we propose an anti-interference tracking algorithm based on the spatiotemporal correlation of an image plane topological graph network. Through infrared image preprocessing and connected-region labeling, the regions of the image plane are automatically and accurately extracted to achieve accurate detection of suspected targets. The topological graph network of the image plane multi-objects is established by using the geometric topological relation of the connected regions, spatial direction relation, and spatial distance relation. Using the anti-interference tracking strategy of the spatiotemporal correlation of the topological graph network, the topological graph of the current frame is registered according to its spatial topological relationship; further, automatic acquisition of the target under a strong interference situation is realized. The optimization of the parameters effectively solves the problem of searching and matching suspected target nodes in the graph alignment. One-pass evaluation experiments show high tracking accuracy and speed of 90.0% and 248.6 FPS, respectively.