基于回波能量空时分布的微弱目标智能TBD技术OA

Track Before Detect(TBD)of maneuvering faint targets remains a challenge in the field of target detection.The motion states of maneuvering targets are often difficult to accurately describe using models,leading to ineffective energy accumulation in model-based TBD methods and significant degradation in detection performance.To address this issue,the spatio-temporal distribution characteristics of target echo energy are explored thoroughly and an intelli-gent TBD technique specifically for maneuvering faint targets is proposed.This method achieves effective accumulation of echo energy without requiring prior information on motion models.Improved YOLO instance segmentation networks are employed to detect targets using multi-frame echo data accumulated based on the maximum value principle,and the algorithm's effectiveness and feasibility are verified through simulation experiments.Under the signal-to-noise ratio of 8 dB,the detection probability for single targets exceeds 86%,while for multiple targets it exceeds 80%.