Pig backfat thickness measuring method and system

Abstract

The invention relates to the technical field of animal detection, and discloses a pig backfat thickness measurement method, which comprises the following steps: acquiring an RGB-D video of the hip of a pig, determining a key frame from each frame of image in the RGB-D video by a pre-trained Mask R-CNN model, and obtaining a segmented image with a pig target, wherein a feature extraction network of the Mask R-CNN model adopts a lightweight network MobilenetV3, using a depth separable convolution and inverse residual structure, and introducing a lightweight attention module; according to the physiological part and hip width of the pig, segmenting he segmented image with the pig target to obtain a key region image of the hip of the pig; subjecting the hip key area image of the pig to four-channel input to a trained deep learning model network for fat thickness prediction. The deep learning model network uses a resenet50 backbone network to be combined with an FPN structure, and backfat thickness value prediction is more accurate. In addition, the invention also provides a system for realizing the method.

Description

technical field [0001] The invention relates to the technical field of animal detection, in particular to a method and system for measuring pig backfat thickness. Background technique [0002] The fat thickness of pigs is one of the important indicators reflecting the body condition of pigs. Regulating the feeding amount according to the fat thickness of pigs is an efficient breeding method, which affects the tenderness and flavor of the meat. The traditional and common method of measuring fat thickness is mainly by manual measurement. In the actual production, the breeder strictly follows the general regulations of the international pig industry, and uses the backfat tester to manually measure the backfat of the sow P2 point (the last rib). The measurement result of this method is relatively accurate, but it consumes manpower and material resources, and it is easy to cause the stress response of the pigs, which is not conducive to the welfare of the pigs. [0003] At pre...

AI Technical Summary

Problems solved by technology

The image collected by this patent is a color image, which is easily disturbed by light, shadow and its own surface texture, so it is difficult to extract the backfat area, and the extraction result is not accurate enough, and the patent is based on the actual length represented by each pixel Calculate the thickness of backfat, therefore, it will also lead to inaccurate measurement results of backfat thickness

Images