| Summary: | 碩士 === 大同大學 === 資訊工程研究所 === 90 === Moving object counting can be applied in quite a lot of fields. Examples are vehicle counting and entrance customer in/out counting. In this thesis, we proposed an object counting system using motion estimation technique. In this paper, we aimed at the vehicle objects. Current techniques for road-traffic monitoring rely on sensors which have limited capabilities, both costly and disruptive to install. The use of video cameras, coupled with computer vision techniques are an attractive option other than current sensors. In this thesis, a moving object counting system based on motion estimation will be presented.
In this paper we rely on the extraction of motion from image sequences so as to segment vehicles (i.e. moving objects) from the scene. Our approach is based on the block matching algorithm (abbreviated as BMA), which is used for motion estimation in the MPEG image sequences compression standard. Before BMA processing, we use image subtraction and complex block detection to select the candidate blocks to reduce block-matching processing time.
The tracking step is based on the following idea. The BMA yields block displacements between two successive frames. We model the vehicle objects using the blocks having similar motion vectors. And we try to find the center point of every vehicle object as our “tracking point”. After finding the tracking points in two frames, we use the least distance to track the same object with these tracking points so that we know the entering and leaving the scence of the objects.
|
|---|
