LTP与光照突变补偿结合的运动目标检测算法

针对现有运动目标检测算法在光照突变条件下鲁棒性不强、易发生误检等问题,提出了将LTP算子与光照突变补偿模型结合的运动目标检测算法。首先利用LTP算子获取当前帧与其背景图像的纹理特征图,然后计算当前帧图像像素点作为前景点的概率并依据概率自适应地更新背景,再判断图像是否发生光照突变补偿并采用线性模型补偿突变图像序列,最后利用背景减除法获取运动目标。实验结果表明本文算法在光照突变条件下对运动目标检测效果理想,检测精度指标PR高于同类其他算法,且具有良好的鲁棒性。

Moving Objects Detecting Algorithm Based on LTP Combining with Illumination Compensation

Existing algorithms of moving objects detecting under the condition of illumination mutation provide lower accuracy and higher false results, so an algorithm combining the LTP operator with illumination compensation is proposed in this paper. Firstly, the LTP operator is taken advantage of to obtain texture feature of the current frame and its background. Secondly, calculate the probability that the points of current frame are the foreground and update the background frame according to the probability. Thirdly, judge whether the illumination mutation happens or not, and the images of illumination mutation are dealt with linear compensation model. Lastly, background subtraction is employed to get the moving objects. Experiment result shows that the algorithm of moving objects detecting has strong robustness under the condition of illumination mutation, and the PR is higher than other algorithms.