飞行器仿真测试中合作目标投影光斑的跟踪识别

针对飞行器仿真测试中合作目标投影光斑的跟踪识别问题,提出了一种投影光斑跟踪识别新方法,该方法主要由预测、识别及修正3个阶段组成。预测阶段主要结合投影光斑运动特点对传统卡尔曼滤波进行改进,提高投影光斑的位置预测精度。识别阶段则根据投影光斑位置的预测值,分两种情况对光斑进行处理:如果下一时刻投影光斑在视场内,则根据相应的判别准则和匹配策略在下一时刻图像中快速搜索投影光斑的最优匹配光斑;如果下一时刻投影光斑在视场外,则根据测量系统相关信息,对视场外投影光斑在图像平面上的位置进行求解,并将求解结果加入相应的运动轨迹,实现对视场外投影光斑的跟踪识别。在完成投影光斑的跟踪识别后,根据跟踪识别结果对投影光斑相关参数进行修正。仿真实验和实际实验结果表明,本文方法能够有效跟踪识别飞行器仿真测试中合作目标的投影光斑,其最大跟踪识别误差不超过2.5pixel,即使跟踪识别过程中存在投影光斑进出视场的情况也不受影响。

Tracking and recognition of projective spots for cooperation targets in vehicle simulation test

To track and recognize the projective spots of a cooperation target in the vehicle simulation test, a tracking and recognition method is proposed by combining the advantages of statistical methods and heuristic methods. This method tracks and recognizes the projective spots in a "prediction-recognition-modification" loop. In the prediction phase, the traditional Kalman filter is improved based on the movement characteristics of projective spots so as to precisely predict the projective spot positions. According to the predicted positions, the recognition of the projective spots in the recognition phase is divided into two parts. If a projective spot is in the field-of-view at next time instance, its optimal matching spot in the image will be rapidly searched with the gain function and matching strategy. If the projective spot is out of the field-of-view at next time instance, its position on the image plane will be calculated with the measuring system information, and the calculated position will be added into the corresponding trajectory. Once the recognition of the projective spots is completed, the parameters related to projective spots are modified with the recognition results. Simulation and real experimental results indicate that the proposed method can effectively track and recognize the projective spots, whose maximum error is no more than 2.5 pixel, even if the projective spots exit and enter the scene during the measurement.