Mobile machine vision development for urine patch detection

Abstract

This paper discusses the research of integrating and implementing current technologies to detect urine patches. Currently, cattle urine is damaging the agricultural industry and poses an increasing threat to the environment. Farmers invest a lot of money treating urine patches and nitrate leaches into waterways, which pollutes drinking water and causes fish to die. Current methods are slow and inefficient in detecting urine patches across large areas of land. The proposed method comprises of mounting a smartphone onto a quadcopter and using the smartphone's camera to run OpenCV image processing libraries to visually detect urine patches. The quadcopter serves as a fast method of moving the smartphone, at a fixed height from the ground, to survey and detect urine patches. When a urine patch is detected, its GPS coordinates are obtained, from the smartphone's GPS sensor, and they are sent to the farmer so the patches can be located for treatment.