A study on the three-dimensional coordinate calibration using fuzzy system

In Surveillance system usually look for the color or edge of an object in order to identify its key features, and this leads us to the key topic of geometric detection algorithms. For spatial calibration method in this thesis, the ellipse is used as the calibration target object. The method proposed by this paper applies fuzzy logic to the image pre-processing, and then links adjacent objects within the image to designate as region of interest. Sobel edge detection is passed onto the region of interest, and then the ellipse is approximated by the least square method. Spatial calibration can subsequently be performed using data from the detected information of ellipse. By adding fuzzy logic to the image pre-processing, not only can we reduce the area of interest for detection, eliminate some of the noise, and accelerate the detection efficiency, but also improve the level of intelligence of the original surveillance system.