Abstract
Simple ways of health management such as wearable devices are invented for daily life. However, way of analyzing urine that indicates one's condition is not established. Color, transparency, volume and number of urination inform human health condition. Our study propose a method of estimating volume of liquid imitating male urination by a camera. This urine analysis would be the first step of logging data of urine for health management. We imaged liquid imitating male urination discharged in air to estimate liquid volume with multiple cylindrical model. The model is on the idea of adding up total volume of cylinder calculated by initial velocity and diameter of liquid extracted from binary image. Initial velocity of liquid is calculated from the approximated curve of white pixels in binary images representing liquid. Liquid diameter is calculated from the normal vector of the approximated curve line. In conditions of different discharge angle, speed, and caliber, average estimation accuracy was within 5% of error. In addition, we extracted liquid color by taking advantage of using camera. We showed possibility that our method could be applied for urine volume management in future.
