Research Article
Stability Tracking Detection of Moving Objects in Video Images Based on Computer Vision Technology
@INPROCEEDINGS{10.1007/978-3-031-50574-4_5, author={Ningning Wang and Qiangjun Liu}, title={Stability Tracking Detection of Moving Objects in Video Images Based on Computer Vision Technology}, proceedings={Multimedia Technology and Enhanced Learning. 5th EAI International Conference, ICMTEL 2023, Leicester, UK, April 28-29, 2023, Proceedings, Part II}, proceedings_a={ICMTEL PART 2}, year={2024}, month={2}, keywords={Computer Vision Technology Video image Edge Detection Moving Target Tracking Detection}, doi={10.1007/978-3-031-50574-4_5} }- Ningning Wang
Qiangjun Liu
Year: 2024
Stability Tracking Detection of Moving Objects in Video Images Based on Computer Vision Technology
ICMTEL PART 2
Springer
DOI: 10.1007/978-3-031-50574-4_5
Abstract
In order to accurately collect images of moving targets and improve the accuracy of target tracking and detection, this paper proposes a new gray-scale image moving target stability tracking and detection method based on computer vision technology. Set camera parameters and light source parameters, and accurately collect moving target images through computer vision technology to improve the accuracy of moving target tracking and detection. The video image is preprocessed, and the specific preprocessing steps include image enhancement and edge detection. The random forest is used as the classifier to eliminate the background, generate a rough target ROI map, and implement the corresponding scale recognition on the ROI area to realize the recognition of moving objects in video images. Combining the Camshift algorithm and the Kalman filter algorithm, the existing moving target tracking method is improved, and the stable tracking of the moving target is implemented. The stability detection of moving targets is implemented by the background difference method, and the selected background difference method is the ViBe algorithm. The test results show that the design method correctly handles more frames and has a higher accuracy rate. It processes more frames per second and has lower tracking and detection errors. The average tracking detection time of this method is less than 3000 s.
