Key points are not available for this paper at this time.
In order to achieve rapid and efficient moving target detection, an improved moving target detection algorithm for background modeling is proposed. Firstly, the first frame is selected as the initial background images from the video images' sequence. Then, we read the continuous N+1 frames of images and use the frame different methods to generate N different images and different background images. Next, the rectangular radial basis function is introduced to count the number of differences within the rectangular width in each different image, and select the pixel value of the original image corresponding to the differential image with the maximum frequency as the reference background. Finally, accurate and complete moving targets are extracted by the different background method. Experimental results show that the proposed algorithm can reconstruct highly reliable background images and exhibit excellent performance and robustness in moving target detection. This not only improves the detection efficiency of moving targets, but also provides a new powerful tool for research and application in related fields.
Wang et al. (Fri,) studied this question.