Linear estimation of cows by the point cloud

The article analyzes the importance and relevance of implementing 3D camera technology in the agribusiness sector, especially in the field of farming and animal husbandry. The focus is on the complexity and labor-intensive nature of the existing on-farm linear estimation process for cows, which is currently not automated and requires considerable effort. As a potential solution to this problem, the use of an Intel RealSense D435 3D camera in combination with developed algorithms for efficient extraction and processing of information from the cow point cloud is proposed. The introduction of the article analyzes the existing research and development on the problem at hand, which emphasizes the importance and timeliness of the topic. Furthermore, the authors describe the methodology of collecting 3D data of the cow croup using the mentioned 3D camera and justify the choice of this equipment for solving the task at hand. In the course of the work, algorithms were developed and adapted to perform filtering, preprocessing of the point cloud obtained from cows, followed by segmentation and measurement of linear parameters of the animals. These algorithms were subjected to laboratory tests on a specially designed cow croup model. The purpose of these tests was to compare the results obtained from both manual measurements and the automated process. Based on the results of the laboratory tests, it was found that the average error of the measurements made by the algorithm was 3.5%, while the maximum error did not exceed 9.2%. The algorithm was also tested directly on the farm. This stage allowed to verify the performance and efficiency of the proposed solution in real conditions. The test results confirmed the high applicability and implementation potential of the developed system. Thus, an innovative solution is proposed that can improve the current approach to measuring the linear parameters of the cow.

1. Ivanov Yu. A. Digital animal husbandry: development prospects // Equipment and technologies in animal husbandry. - 2019. - No. 1 (33). - P. 4-7.

2. Ivanov Yu. A., Zarikeev A. R. USE OF ARTIFICIAL INTELLIGENCE IN ASSESSING THE EXTERIOR OF CATTLE // Equipment and technologies in animal husbandry. - 2021. - No. 4 (44). - P. 6-10.

3. Kirsanov V. V. et al. Comparative analysis and selection of technical vision systems in dairy farming // Bulletin of NGIEI. - 2019. - No. 1 (92). - P. 69-79;

4. Pavkin D. Yu. et al. Contactless assessment of dairy cow fatness using ToF technology // Agroengineering. - 2021. - No. 2 (102). - P. 39-44.

5. Tsoi Yu. A., Baisheva R. A. Technological aspects of creating a “smart” dairy farm // Agrarian Science of the Euro-North-East. - 2019. - Vol. 20. - No. 2. - P. 192-199.

6. Han X. F. et al. A review of algorithms for filtering the 3D point cloud // Signal Processing: Image Communication. - 2017. - Vol. 57. - P. 103-112; 8. He Y., Chen S. Recent advances in 3D data acquisition and processing by time-of-flight camera //IEEE Access. – 2019. – T. 7. – P. 12495-12510;

7. He Y., Chen S. Recent advances in 3D data acquisition and processing by time-of-flight camera //IEEE Access. – 2019. – Т. 7. – С. 12495-12510;

8. Keselman L. et al. Intel realsense stereoscopic depth cameras //Proceedings of the IEEE conference on computer vision and pattern recognition workshops. – 2017. – С. 1-10;

9. Lazaros N., Sirakoulis G. C., Gasteratos A. Review of stereo vision algorithms: from software to hardware //International Journal of Optomechatronics. – 2008. – Т. 2. – №. 4. – С. 435-462;

10. Nagy S. Á. et al. Impact evaluation of score classes and annotation regions in deep learning-based dairy cow body condition prediction //Animals. – 2023. – Т. 13. – №. 2. – С. 194.

11. O’Mahony N. et al. Two-and three-dimensional computer vision techniques for more reliable body condition scoring //Dairy. – 2022. – Т. 4. – №. 1. – С. 1-25.

12. Shilin D. et al. IDENTIFICATION OF THE COW’S NIPPLES USING A 3D CAMERA WITH TOF TECHNOLOGY //Annals of DAAAM & Proceedings. – 2022. – С. 77-85.

13. Tadic V. et al. Perspectives of realsense and zed depth sensors for robotic vision applications //Machines. – 2022. – Т. 10. – №. 3. – С. 183.

14. Zhou Q. Y., Park J., Koltun V. Open3D: A modern library for 3D data processing //arXiv preprint arXiv:1801.09847. – 2018.