In this paper, we propose an adaptive technique for the automatic extraction and tracking of moving objects in video sequences that works robustly under the influence of image-specific disturbances (e.g. brightness variations, shadow and partial occlusion). For this technique, we apply the colour information, a neural recognition system and a recursive filtering algorithm to the improvement of the matching quality when disturbances occur. This suggested intensity-based technique is adaptive and robust compared to the conventional intensity-based methods.
展开▼