Quantifying Residual Feed in a Fence-line Feedlot Bunk using Depth Camera Imaging Techniques

Authors

Jean Niwenshuti, University of Nebraska - LincolnFollow
Galen E. Erickson, University of Nebraska - LincolnFollow
Tami M. Brown-Bandl, University of Nebraska - LincolnFollow
Eric T. Psota, University of Nebraska-LincolnFollow
Yijie Xiong, University of Nebraska - LincolnFollow

Date of this Version

2023

Citation

2023 Nebraska Beef Cattle Report

UNL Beef, Institute of Agriculture and Natural Resources, University of Nebraska-Lincoln

Abstract

Feed bunk management requires intensive labor and relies on manual observation to estimate the amount of residual feed in the bunk. Alternative and innovative technologies were used to estimate the weight of residual feed in a concrete fence-line bunk using a depth camera. Depth cameras capture the distance between the camera and the object in their field of view. This study used a time-of-flight depth camera (Azure Kinect, Microsoft) to estimate the weight of residual feed in a partial fence-line concrete bunk using 11 common feed ingredients. The camera was fastened approximately 3.3 ft above the center of the bunk to collect images for individual ingredients added at a constant weight increment of 2.2 or 4.5 lb. The feed ingredients inside the bunk were stirred randomly after each picture collection to simulate the shape of residual feed after cattle's feeding event. Individual ingredients were then weighed using a scale for comparison with the image-estimated weights. Linear regression showed that the scale-measured weights and image-estimated weights were linearly related, with an R² ranging from 0.9833 to 0.9992. Results indicate that depth cameras are capable of accurately estimating the weight of residual feed in the bunk. Overall, this experiment demonstrates a first step in the development of feed bunk management tools using precision livestock management techniques.