An edge extraction method based on gradient with ACS

A new edge extraction method based on ACS and gradient is proposed. Due to information feedback, the complicated problem can be resolved by the simple ants. The new food definition and movement mechanism to the ant is applied in this study. Multi-scale gradient is taken as the food objective of the ant. Besides, two directional functions included strong and weak one for prior directional searching are used, and the pheromone is graded. Furthermore, the thresholding is automatic for the final edge extraction. According to the compare with other methods in the experiment, the efficiency of the proposed method is proved, and the method is not sensitive to the noise.