Smartphone-Based Machine Vision for Grain Quality Assessment

Abstract

Improving agricultural production technologies using digital tools contributes to achieving the United Nations’ Sustainable Development Goals, particularly ensuring food security, responsible resource consumption, and innovative infrastructure. The article presents the results of a study of the potential of using a smartphone camera to determine the parameters of a grain using machine vision. In particular, the results of studies on determining the deviation from the actual size and average color of the grain depending on the camera’s characteristics and the environment are presented. As a result of the research, the regularity of changes in the rational values of grain illumination from the camera resolution was established. It was also found that when the camera resolution changes from 3 MP to 40 MP, the required illumination level changes from 2075 lx to 3059 lx. The dependence of the change in the rational values of grain illumination on the grain length at different camera resolutions was obtained. It was found that when determining the length of a grain from 5.5 mm to 7.0 mm using a camera with a resolution of 3 MP, the required illumination level varies from 2842 lx to 3661 lx, and when using a camera with a resolution of 31 MP, from 1311 lx to 2131 lx. A regular change in the rational values of grain illumination (1274–3437 lx) from the camera’s light sensitivity in a range of ISO 200 – 1300 made it possible to choose a rational value of grain illumination at the possible light sensitivity of the camera.