基于帧差的灰度投影的快速运动物体检测

Moving Object Detection Based on the Gray Projection of Frame Subtraction

基于传统的背景差分法和帧差法,提出将两者结合同时利用灰度投影法进行快速的运动物体检测.该算法利用相邻3帧图像序列进行垂直灰度投影,然后把相邻帧的投影相减,通过前后帧差确定中间帧的运动物体的垂直范围.然后在缩小的范围内进行水平投影,和背景图的水平灰度投影相减得到物体的水平范围.实验结果证明,这种方法能够提高检测的效率和精度.

Based on the traditional methods of background subtraction and frame subtraction, a method of gray projection combining background subtraction and frame subtraction to implement the fast detection of moving objects is presented. The algorithm utilizes three adjacent frames of image sequences to do the gray projection in the vertical direction, and then the two adjacent frames of gray projection are subtracted. The moving object of middle frame can be detected in the vertical direction through the two-frame gray projective subtractions. Then horizontal projection is implemented, based on the vertical projection, the horizontal projection of background subtrated to gain the moving object in the horizontal direction. The experimental results prove that this method can improve the efficiency and precision of detection.