Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

High-Throughput and Accurate 3D Scanning of Cattle Using Time-of-Flight Sensors and Deep Learning

Gbenga Omotara, Seyed Mohamad Ali Tousi, Jared Decker, Derek Brake, Guilherme N. DeSouza

TL;DR

This work addresses the need for rapid, precise cattle phenotyping by introducing a multi-view 3D scanning system built around eight time-of-flight sensors and embedded controllers. It combines RGB-D data with a fine-tuned Mask R-CNN for cattle segmentation, followed by pairwise and multi-view registration using colored ICP and Poisson surface reconstruction to produce accurate cattle meshes and derive volume and surface-area measurements. The approach is validated on known objects and live cattle, with emphasis on synchronization, segmentation accuracy, and robust reconstruction in real-world environments, including untamed animals. The platform offers high throughput, mobility, and reliable performance that can support nutrition management, welfare monitoring, and breeding studies in livestock research and farming operations.

Abstract

We introduce a high throughput 3D scanning solution specifically designed to precisely measure cattle phenotypes. This scanner leverages an array of depth sensors, i.e. time-of-flight (Tof) sensors, each governed by dedicated embedded devices. The system excels at generating high-fidelity 3D point clouds, thus facilitating an accurate mesh that faithfully reconstructs the cattle geometry on the fly. In order to evaluate the performance of our system, we have implemented a two-fold validation process. Initially, we test the scanner's competency in determining volume and surface area measurements within a controlled environment featuring known objects. Secondly, we explore the impact and necessity of multi-device synchronization when operating a series of time-of-flight sensors. Based on the experimental results, the proposed system is capable of producing high-quality meshes of untamed cattle for livestock studies.

High-Throughput and Accurate 3D Scanning of Cattle Using Time-of-Flight Sensors and Deep Learning

TL;DR

This work addresses the need for rapid, precise cattle phenotyping by introducing a multi-view 3D scanning system built around eight time-of-flight sensors and embedded controllers. It combines RGB-D data with a fine-tuned Mask R-CNN for cattle segmentation, followed by pairwise and multi-view registration using colored ICP and Poisson surface reconstruction to produce accurate cattle meshes and derive volume and surface-area measurements. The approach is validated on known objects and live cattle, with emphasis on synchronization, segmentation accuracy, and robust reconstruction in real-world environments, including untamed animals. The platform offers high throughput, mobility, and reliable performance that can support nutrition management, welfare monitoring, and breeding studies in livestock research and farming operations.

Abstract

We introduce a high throughput 3D scanning solution specifically designed to precisely measure cattle phenotypes. This scanner leverages an array of depth sensors, i.e. time-of-flight (Tof) sensors, each governed by dedicated embedded devices. The system excels at generating high-fidelity 3D point clouds, thus facilitating an accurate mesh that faithfully reconstructs the cattle geometry on the fly. In order to evaluate the performance of our system, we have implemented a two-fold validation process. Initially, we test the scanner's competency in determining volume and surface area measurements within a controlled environment featuring known objects. Secondly, we explore the impact and necessity of multi-device synchronization when operating a series of time-of-flight sensors. Based on the experimental results, the proposed system is capable of producing high-quality meshes of untamed cattle for livestock studies.
Paper Structure (16 sections, 7 figures, 2 tables)

This paper contains 16 sections, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Scanners for Phenotyping large Animals
  4. Proposed Scanning System
  5. Platform
  6. Hardware
  7. Software and Pipeline
  8. Advantages over Other Systems
  9. Experimental Results and Discussions
  10. Implementation Considerations: Sensor Synchronization
  11. Experiments on Known Objects
  12. Experiments on Real Cattle
  13. Fine-Tuning Mask R-CNN
  14. Registering All the Views
  15. Extracting Surface and Volume measurements on Cattle
  16. ...and 1 more sections

Figures (7)

  • Figure 1: The proposed platform for 3D cattle scanner system.
  • Figure 2: The block diagram of the proposed pipeline.
  • Figure 3: The camera setup for experimenting with the known objects and laser interference.
  • Figure 4: A qualitative comparison between the scanner outputs to demonstrate the effect of using and not using the proper synchronization delay.
  • Figure 5: The statistical results of scanning one known object (a cylinder) in different orientations.
  • ...and 2 more figures