원문정보
초록
영어
In order to further improve motion video target tracking and detection accuracy, a motion video tracking algorithm based on memory watershed disc gradient expansion template is proposed in this article. Firstly, heterodromous diffusion is executed to preprocess the motion video image in order to reduce the noise interference in the motion target tracking and detection process, and meanwhile the differential operation and the morphological operation are executed to extract and operate the outer profile of the motion object. Secondly, in allusion to the excessive division in the watershed algorithm, target marking and division of the watershed algorithm are realized on the basis of characteristic memory, and the accurate detection of the motion target is realized on the basis of the disc gradient expansion template. Finally, according to the experimental comparison, the algorithm proposed in this article can be applied in the complex background detection of the motion target, and such algorithm can not only improve the motion target detection accuracy, but also significantly improve computation speed.
목차
1. Introduction
2. Motion Video Sequence Detection
2.1. Algorithm Procedure
2.2. Adoption of Heterodromous Diffusion for Motion Sequence Denoising
2.3. Profile Extraction and Morphological Operation
2.4. Accurate Division of Background and Target
3. Experiment and Analysis
3.1. Setting of Experimental Environment
3.2. Experiment Result and Comparison Analysis (Qualitative)
3.3. Experiment Result and Comparison Analysis (Quantitative)
3.4. Algorithm Operating Speed Comparison
4. Conclusion
References
