| Summary: | 碩士 === 國立交通大學 === 電控工程研究所 === 100 === Human activity recognition system is now a very popular subject for research and application. Using a fixed camera to track a person and recognize his (her) activity is widely seen in home surveillance. For real-time surveillance, the embedded algorithms must be efficient and fast to meet the real-time constraint.
In the thesis, a new person tracking and continuous activity recognition is proposed. We build two background models, in grayscale and HSV color space as well to extract the human correctly, and we could also reduce the shadowing effect well. For better efficiency and separability, the binary image is firstly transformed to a new space by eigenspace and then canonical space transformation, and the recognition is finally done in canonical space. A three image frame sequence, 5:1 down sampling from the video, is converted to a posture sequence by template matching. The posture sequence is classified to an action by fuzzy rules inference. Fuzzy rule approach can not only combine temporal sequence information for recognition but also be tolerant to variation of action done by different people and time.
Moreover, we make use of the hue component to recognize the medical pouch’s color when one is taking medicine. By combining with the hue-based pouch’s color model and human activity recognition system, we can know someone is taking medicine and its medical pouch’s color as well. Finally, we also employ the activity recognition system to record a student’s activity in the daily living.
|
|---|
