The Development, Design, and Testing of a Remote Sensing System to Measure Cattle Hip Heigh

The development of a cattle hip height sensor arose from the need to increase the efficiency, accuracy, and precision with which hip height measurements are taken. Eight infrared sensors were used to measure the hip height of cattle in a commercially available chute system. Stationary testing was conducted to evaluate the accuracy of the technology without the variability animals naturally impose. A repetition test was conducted as a part of the stationary tests that indicated there was no significant difference found in measured height with movement of the target forward, backward, and center. However, there was a 0.16 inch average absolute error between measured height with the left and right shifts of the target, which was a significant difference. A color test using four colors of felt fabric was conducted to suggest the effects of common cattle hair coat colors on the sensors' measurements. The color test demonstrated that the measured distances to a black target were significantly different than those to all other colors, with the exception of grey. There were no significant differences in measured target distances when using brown, white, and the bare back of the decoy. Because black colors absorb infrared light, talcum powder was added to the black fabric as a reflector, which improved the sensing ability causing an 18.97 in. decrease in absolute error at a 35 15/16 in. distance from sensor to target. This was the only significant difference in the talcum powder test. Data collection on the live animals involved the following tests: visual vs. sensor, effects of position on sensor measurements, and effects of position on visual measurements. Talcum powder was also added to black cattle, which improved the repeatability and accuracy of the live animal tests. During live animal testing, there was no significant difference between the visual and sensor measurements. It was also found that there were significant differences in repeatability between the methods of restraining cattle to gather hip height data. Data from the visual vs. sensor comparison test was used to show the correlation between the visual and IR sensor measurements by averaging all of the measurements per animal together to represent one averaged measurement per animal this correlation provided an R2 of 0.76. The frame score plot of the averaged data provided an R2 of 0.75.