Application of technical vision systems for diagnosing the quality of cattle feed

Russian and foreign literature on the development of diagnostic systems and scanning of objects using a vision system with deep machine learning programs has been analyzed during the study. The features of the technological process of feeding cattle have been studied. A system of non-contact assessment of the dry matter content/humidity of the components of the feed mixture of natural cultivation on the example of a corn silo using technical vision systems was proposed. A database of images of corn silage was collected and the dependences on the intensity of the reflecting light flux of the silage were revealed taking into account changes in humidity. The research was conducted in 2020 on the basis of the Federal Scientific Agroengineering Center VIM (FNAC VIM), using experimental equipment of the Institute of General Physics of the Russian Academy of Sciences named after A. M. Prokhorov and FNAC VIM. A stand with a technical vision system has been developed that allows to classify the components of a cattle feed mixture by color characteristics. The obtained dependences of the reflecting intensity of corn silage allow us to assert the prospect of using a vision system for express-evaluation of the quality indicators of feed mixture components. Taking into account the level of robotization of technological processes of feeding cattle, the problem of assessing the quality indicators (in particular, the dry matter/moisture content) of the components of a feed mixture is relevant.