engleski

Wavelet-Based Edge Detection for Robot Vision Applications

Edge detection is a very important step in low level image processing. It can be used in measurement and prediction. If separate lines are connected in contours, next step can be object recognition, tracking or similar applications. Edge detection can be used as a step in measurement of distance from the obstacles. In this paper, a new wavelet edge detector is developed. It is compared with well known detectors. Depending on which standard algorithm is used the execution time of edge detector is improved by 343% (in comparison to the Perwitt edge detector) to 658% (in comparison to the Canny edge detector). Results show that quality of edge detection is increased as well. In comparison to multi wavelet approach improvement is 57%.

Image processing; Edge detection; Computer vision; Real-time computation

nije evidentirano