The goal of this project is to automate the inspection of dairy cow by means of body condition scoring (BCS) with a commodity RGB+D sensor and low cost embedded computer. The RGB+D sensor is mounted on the ceiling of an exit hallway from the UK diary parlor. All cows are milked twice a day and, upon being milked, walk down this corridor. As they walk under the sensor, the computer-on-module detects motion in the video stream, and then looks for the tail section of the animal. Once detected, the image is mated to an RFID tag reader, which reads the animal’s unique serial number. This serial number is embedded into the captured image and uploaded to Azure IOT storage, where it is downloaded from a remote site for further analysis.
Stacks Image 21