Using cameras to assess diet uniformity of dairy cow diets

Why this research is being done

We aim to feed dairy cattle consistent diets so that every mouthful of feed has a similar nutrient composition. We are exploring the use of image analysis of feed during the mixing process to evaluate if we can differentiate between well- and poorly-mixed diets. By utilizing a camera system, we record the mixing process of total mixed rations above the mixer wagon. Dozens of regions of interest within the mixer wagon are selected and compared to assess uniformity. The objective of this study is to use image analysis to determine if a diet is uniform during the feed mixing process.

Results

This work uses image analysis to distinguish well-mixed diets from poorly-mixed diets within a mixer wagon. Diets contained the same proportion of ingredients but differed in mixing time and load size. Ground truth data on diet uniformity was collected by analyzing 10 samples per diet along the feedbunk, after feed delivery, for particle size distribution and mineral analysis. Several challenges exist for analyzing video on farm including, changes in light, dust, shadow. Results showed that it is possible to determine if a diet is uniform based on image analysis.

Conclusions

Our ground truth data, showed that diet uniformity can be greatly impacted by load size and mixing time. We will continue to explore the use of video to evaluate feed uniformity, moving towards solutions that can be utilized on farm. Initial results indicate that video may be a viable solution to evaluate diet uniformity as we were able to differentiate between well- and poorly-mixed diets when comparing multiple regions within the mixer wagon.

Contact information

Dr. Jackie Boerman jboerma@purdue.edu | 765-496-6290
Dr. Amy Reibman reibman@purdue.edu