This paper proposes a new scheme of image technology to extract the color feature of objects and reduce the impact of environmental interference caused by the shadows and ambient illumination changes. This preliminary process is important for the further tracking by the particle filter. This paper designs a series of procedures for the moving objects tracking. The first step is the foreground segmentation to separate the moving objects from the background after color system transformation. The next part uses the closing and opening algorithm of morphology processing to deal properly with the noise and intensify the object shape. The final part applies the particle filter which is used properly in non-linear and no noises restriction system to tracking multiple crossing objects. According to the experimental analysis, the proposed method can track multiple crossing targets accurately.
All Science Journal Classification (ASJC) codes
- Computer Science(all)