双目视觉-激光测距传感器目标跟踪系统

Target Tracking System of Binocular Vision and Laser Range Sensor

针对视觉传感器在目标跟踪定位过程中,沿光轴方向量测精度较低的问题,提出了使用一维点激光测距传感器进行修正的方法。同时,为了充分发挥双目视觉传感器和激光测距传感器在目标跟踪方面的优势,利用信息融合技术来提高信息的利用率和测量的精度。在系统量测过程中,由于激光测距传感器受到二维转台机械性能影响,导致测距值成为时滞信息,提出了时滞信息直接更新算法将常时滞信息的估计值作为当前时刻量的量测值带入后续算法中,从而解决时滞问题;随后利用联邦卡尔曼滤波算法解决了双目视觉传感器和激光测距传感器之间的估计相关问题。仿真结果表明,经过直接更新算法后时滞信息的实时性得到提高,实验同时验证了该算法的有效性,提高了目标跟踪系统的量测精度,改善了双目视觉传感器及激光测距传感器本身带来的精度和实时性方面的不足。

Aiming at improving the low accuracy of the binocular vision sensor along the optical axis in target tracking, the correcting method by one-dimensional laser range sensor is proposed. Meanwhile, in order to take full advantage of the binocular vision sensor and the laser range sensor in target tracking, the information fusion is used to improve the measurement accuracy and data utilization. In the process of measurement, the frequency of the laser range sensor is affected by the mechanical property of two-dimensional turntable, which causes the data of laser range sensor time-delay, Utilizing the direct updated algorithm to estimate the one-step prediction data and make it into real-time data in follow-up algorithm. Then using federated Kalman filter to solve the dependence of the multi- sensor data. Simulation results show that the real-time performance of the time delay information is enhanced by the direct updated algorithm. Meanwhile, experimental results show that the final information fusion algorithm improves the accuracy of target tracking and improves the sensor shortages of the accuracy and timing.