Method for screening ideal frames of pig images based on machine vision technology

Abstract

The invention discloses a method for screening ideal frames of pig images based on a machine vision technology, and relates to the field of breeding equipment, and the technical scheme is that the method comprises the steps: collecting pig images through a RGBD camera, preprocessing the collected pig images, converting the collected pig images into image frames, and storing the image frames; carrying out contour box selection on a pig body in the obtained image frame to obtain a contour box; according to the body sizes of the pigs in different growth stages, presetting standard positioning frame parameters corresponding to the growth stages; comparing a contour box in the image frame with a standard positioning frame, and performing primary screening; and performing secondary screening according to the gradient of the contour box in the image subjected to primary screening, and taking the image as an ideal frame. The method has the beneficial effects that the screening method provided by the scheme is beneficial to the efficiency improvement and accurate screening of a machine vision weight estimation technology in an image acquisition link, and the accuracy and reliability of an output result of the vision weight estimation technology are ensured by reducing the image redundancy and improving the quality of an image library.

Description

technical field [0001] The invention relates to the field of farming equipment, in particular to a method for screening ideal frames of pig images based on machine vision technology. Background technique [0002] In the case of strict prevention and control of African swine fever, non-contact measurement technology is particularly important. Growth information such as pig body size and weight is an important indicator to describe its quality and nutritional status. In terms of quality, pig live body size can be used to estimate its lean meat percentage, backfat thickness, eye muscle area, etc. In terms of nutritional status, the body size characteristics of live pigs and the parameters of daily weight gain are important indicators for judging their growth trend. The traditional process of measuring body size and weight is mainly by driving pigs manually, measuring them on weighbridges or weighing them subjectively. The measurement method prevails. This traditional method ...

AI Technical Summary

Problems solved by technology

This traditional method poses a huge challenge to the stability of the equipment in the long-term breeding environment, because due to the complex environment of agricultural production, high temperature and high humidity will affect the measurement accuracy of the sensor, and the animals are living and cannot cooperate with the detection according to human will , and at the same time, if it is forcibly restrained into manual weighing equipment, it will undoubtedly cause huge stress

[0003] The non-contact detection of pig body size and weight estimation technology based on machine vision technology is also the direction that scholars are keen to explore in recent years. However, machine vision weight estimation technology will encounter various unqualified data in the image collection process. For example, the postures of pigs such as head down, head up, bending, and tilting do not meet the requirements of image data, that is, there are a lot of redundant and invalid images. Therefore, there is an urgent need for a method of screening ideal image frames of pigs to eliminate a large number of unqualified Image, improve the quality of the image library to ensure the accuracy and reliability of the output results of the visual weight estimation technology

Images