Can we use drones to estimate body dimensions of individual cows?

Scientists from Wageningen University & Research (WUR) investigated whether UAV technology can also be used to monitor traits of individual cows, in particular the cows’ weight and body dimensions.

Traditional monitoring of herds in extensive production systems is a time-consuming task where farmers need to locate cattle in large grazing areas. Especially in large remote areas which are difficult to access, detecting the herd with an UAV can support farmers in efficiently detecting and tracking the herd. If this technology could also be used to monitor individual traits, it could provide important and valuable insights in the health status of cattle without having to visit the cattle on a routine basis. If, for example, the UAV measured weight of a cow lowers significantly, the farmer could visit the individual cow for further health checks.

UAV technology for individual cow traits

To explore the use of UAV technology for individual cow traits, scientists from Wageningen Environmental Research and Wageningen Livestock Research used different types of UAVs and camera systems to estimate the body dimensions of individual Holstein Friesian cows. Over the course of 3 years, 25 different cows were monitored and 4,611 images, approximately 10 videos and a LiDAR (Light Detection And Ranging) dataset were analysed. By creating 3D models of the individual cows, the scientists were able to estimate the volume of a cow and, consequently, use a conversion formula to translate the cow’s volume into an estimated weight.

Exploring the use of UAVs in estimating height and weight of cattle has provided new insights in the possibilities and challenges occurring when monitoring cattle with different unique platforms and methods. Unfortunately, certain conditions during flight (e.g. the movement of cattle, rainy weather conditions, high wind speeds and hilly areas) provided a lot of uncertainties and difficulties, which, in turn, hampered the scientists in achieving high certainty regarding the accuracy of the 3D models and weight estimates of the cows. However, further research and developments in this field, for example simultaneous image gathering with multiple drones or automated processing of imagery in 3D models, can be explored in the future to improve results and apply the study in practice.

This study is part of the GenTORE project and has received funding from the European Union’s Horizon research and innovation program, under grant agreement No. 727213. The researchers also thank CARUS farm for their services and hospitality and making it possible to share their data of the cattle.