3D Imaging for Robotic Arm Entry-Window Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic milking systems face challenges in providing a reliable decision basis for controlling robotic arms due to insufficient data quality from image capture, which can lead to potential animal injuries.
Innovation Solution
A system using a 3D imaging camera and control unit to define a well-defined obstacle-free space for robotic arm operation, utilizing reference objects like kick rails and poles to enhance navigation and udder detection, ensuring safe and efficient robotic arm control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If image capture apparatus is used to determine animal location, then automation is improved, but data quality is insufficient for reliable decision-making
Solution Approach 1:
A reference object (kick rail) is introduced as an intermediary element between the camera and the animal. This reference object serves as a mediator that provides stable, distinguishable features for the image processing system to reliably determine spatial relationships and define safe operational zones, thereby improving decision-making reliability without reducing automation.
2Reliability
If 3D image data is processed to define safe space, then reliability is improved, but data processing complexity increases
Solution Approach 1:
The system extracts and processes only a specific subset of the 3D image data - namely, the portion containing the reference object and the defined safe operational space. By taking out only the relevant data subset rather than processing all 3D image data, the system achieves reliable robotic arm control while reducing computational complexity and processing requirements.
3Measurement precision
If reference object is used for navigation, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The reference object (kick rail) serves as a simple, fixed structural element that provides navigation references without requiring complex or expensive sensing mechanisms. This straightforward reference structure improves measurement precision for robotic arm navigation while avoiding the need for sophisticated navigation equipment, thereby reducing overall system complexity.
Data Source
Figure 1~2
Figure 3~4
AI summary
A camera (110) registers three-dimensional image data (Djmg3D) of a milking location comprising a reference object (Rl) and a rotating platform (130) upon which an animal (100) is standing with its hind legs (LH, RH) facing the camera (110). A control unit (120) checks if an entry window for a robotic arm can be found in the image data (Dimg3D) by searching for the reference object (R1) in the image data (Dimg3D) - If the reference object (R1) is found, the control unit (120) searches for an acceptable obstacle-free volume (V) in the image data (Dimg3D), said volume (V) being located within an allowed space (S) relative to the reference object (R1), and containing a continuous area (A) facing the camera (110) which continuous area (A) exceeds a threshold size. Said volume (V) starts at a predefined position (P) and extends at least a primary distance (doK) in a depth direction away from the camera (110). A decision basis (DB) for controlling the robotic arm to perform at least one action relating to the animal (100) includes information about said volume (V).