Research on the quality detection method of infiltration and drip irrigation belt holes based on machine vision and deep learning edge detection algorithm DexiNed

Abstract

The drip irrigation belt needs to punch the inner dripper during the production process. If the drip hole is missed or misaligned, it will affect the growth of crops when it is used. The existing hole position detection methods are not efficient online drip irrigation. The method of hole quality detection. This algorithm uses the OpenCV module to convolve the image with filters such as Gaussian functions to obtain the low-pass filtering result of an image for edge detection, and then uses the deep learning edge detection algorithm DexiNed (Dense Extreme Inception Network: Towards a Robust CNN Model for Edge Detection) performs deep edge detection on the blurred image, and generates the corresponding edge detection graphics, and then uses the Hough transform algorithm to locate the drip groove and drip hole, and compare with the preset value, Judging whether the drip hole is qualified. Through the algorithm experiment, the overall detection effect is very good. This paper proposes a quality detection algorithm for drip irrigation tape based on deep learning and machine vision, which can achieve fast and accurate hole quality detection.