The work aims to propose an image edge extraction method based on Hough transform (HT) coupling ant colony optimization (ACO), in order to solve the reduced edge positioning accuracy, edge information loss and false edge, etc. in the image edge extraction method caused by the noise. Firstly, HT was carried out for input image to eliminate the effect of noise and line segment on image edge. Secondly, the difference between the pixel gradient of the image and the statistical average of the circular neighborhood was calculated, and the weight function between the two was constructed and used as the pheromone and the heuristic information of ant colony. Finally, the ant colony optimization algorithm was used to guide the ant colony to search image, so as to finish the image edge extraction. The experiment results showed that, compared with the current edge detection technique, the proposed algorithm had higher extraction accuracy and efficiency, and could obtain complete edges with abundant details, which effectively reduced the noise effect. In conclusion, the proposed algorithm has a stronger anti-noise performance and can further improve the edge extraction accuracy. Because of that, it can be better applied in the field of package barcode recognition and image processing.