Pork Backfat Thickness Detection Based on Complex Background

Abstract:
Intelligence is the inevitable trend and inevitable result of the transformation of various smart factory industries. By designing an intelligent detection algorithm, it is necessary to use computer vision to measure the size of items and discriminate inferior products to reduce accidental errors caused by labor and the cost of machine use in the factory. Based on computer vision and image processing technology, and according to the requirements of online measurement of pig carcass backfat thickness, this paper studies the pig carcass backfat thickness detection algorithm. Tests have proved that this method has a detection accuracy of 94% when the backfat thickness detection error is ±1mm. The detection time of a single sample is 0.36-0.52s, which can meet the needs of online detection. The algorithm in this paper adds image processing methods based on complex backgrounds, including conventional denoising, clustering and segmentation, and finally edge detection algorithms. In horizontal image correction, it uses improved vanishing point correction and vertical contour line correction. The design of the straight line fitting algorithm, the final extraction of the backfat thickness adopts the algorithm design based on straight line pixel detection, which can effectively improve the accuracy of the algorithm.