Method for straight edge detection based on ant colony algorithm

Straight lines are important intermediate level symbol in image analysis. In order to extract precise straight edges form images, a method based on ant colony algorithm is proposed. Firstly, a weighting method is designed to choose a start point for each ant. Phase information of edge points is used as heuristic information to guide iteratively searches. In each cycle, terminated probability is introduced to end the current search and pheromones on the traversed route of each ant are updated adaptively to the length of the route. When iteration is completed, straight edges can be extracted according to the intensity of pheromones. Compared with traditional Hough algorithm, the proposed method can extract straight edges precisely, and keep the detail of straight edges at the same time.