Research Of Fish Motion Tracing Based On Video Monitoring

The biological monitoring based on fish as the main subject is increasingly becoming the main development direction of the environmental protection, which could timely forecast and prevent the aquatic environment pollution mainly through real-time monitor, behaviours analysis, moving conditions of the organisms individual or groups. With the fastest growing of the computer technology and the popularity of the computer vision technology nowadays, a fish motion tracking study based on video monitoring is conducted in the paper. Fish motion tracking, as the basis of the water environment protection study, capture, trace and monitor the fish motion target accurately and rapidly, is a important method in the research. Video object segmentation and tracking is a key step of the video monitoring technology in this research, is also the basis of the fish motionvideo to the monitoring, object capturing and behavioural analysing. The paper separately optimize the algorithm from the video object segmentation and video object tracking, can improve operation efficiency, enhance the applicability, better the detecting, extracting, segmenting and tracking of the video object.To the video object segmentation, the paper presents the multi-scale morphology watershed object segment based on label, which inherits the merits as simply operation, rapidly segmentation of the traditional Hausdorff, with the multi-scale morphology optimize the object outline and marks the over-segment using the label, can extract and segment the object precisely.There are so many motion object tracing method based on video theory, but still no general one. The paper proposes the weighted Hausdorff tracking algorithm based on multi-value edge, transfers the object edge into multi-value and weights it, overcomes the demerits of the traditional Hausdorff that easy to be interfered with the noise, stress the object and defines the object model update to achieve the object tracking accurately and quickly.To the video tracking experiment with fish, the paper verifies the proposed video segmenting and tracing algorithm through the fish video sequence, analyzes and reflects the validity and suitability of the given algorithm through the time performance and tracking precision.