3D Tool Positioning for Rotary Milking Platform Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic milking systems, particularly rotary milking platforms, face challenges in accurately determining tool positions due to semi-manual programming methods, which are tedious and error-prone, especially in large setups where numerous milking stalls require precise tool placement.
Innovation Solution
A system utilizing a camera and control unit to register three-dimensional image data of tools, applying algorithms to identify and calculate precise positions based on spatial distribution, eliminating the need for human intervention and ensuring accurate storage for automated operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If semi-manual programming is used to teach tool positions, then the milking robot can be programmed with tool positions, but the process is tedious and error-prone
Solution Approach 1:
The patent replaces the manual mechanical control system (joystick) with an automated optical detection system (camera) to determine tool positions. The camera captures images of tools in the storage location, and the control unit automatically calculates positions based on image analysis, eliminating the need for manual operation and significantly reducing programming time while improving accuracy.
Solution Approach 2:
The system enables self-service by allowing the milking robot to automatically determine its own tool positions through camera-based detection. The robot's control unit processes images and calculates positions without human intervention, making the system self-programming capable and eliminating the tedious manual teaching process.
2Productivity
If reference position translation is used for multiple stalls, then programming work is reduced, but considerable errors occur in individual tool positions
Solution Approach 1:
The patent applies segmentation by determining tool positions independently for each stall rather than using a single reference translation for all stalls. The camera captures and analyzes tool positions separately in each storage location, allowing individual position optimization for each stall while maintaining overall system efficiency.
Solution Approach 2:
The patent replaces the mathematical translation method with direct optical measurement using a camera. Instead of calculating positions through coordinate transformation from a reference stall, the system directly images and measures tool positions in each stall, eliminating cumulative errors and improving individual position accuracy.
3Reliability
If operator control is used for safety, then safety is maintained, but the operator cannot accurately see the grip device position
Solution Approach 1:
The patent introduces an intermediary system (camera) that acts as a mediator between the operator and the tool positions. The camera captures precise images of tool locations, and the control unit processes this visual information to determine exact positions, providing both safety through automated control and precision through optical measurement without requiring direct operator observation.
Data Source
Figure 1~2
Figure 3~6
AI summary
The positions of the tools (141, 142, 143, 144) in an automatic milking arrangement are determined by registering, via a camera (130) at an origin location (PC), three-dimensional image data (Dimg3D) representing the tools (141, 142, 143, 144) whose positions are to be determined. Using an algorithm involving matching the image data (Dimg3D) against reference data, tool candidates (TC1, TC2, TC3, TC4) are identified in the three-dimensional image data (Dimg3D). A respective position (PT1, PT2, PT3, PT4) is calculated for the tools (141, 142, 143, 144) based on the origin location (PC) and data expressing respective distances from the origin location (PC) to each of the identified tool candidates (TC1, TC2, TC3, TC4). It is presumed that the tools (141, 142, 143, 144) are arranged according to a spatially even distribution relative to one another. Therefore, any tool candidate (TC2) is disregarded, which is detected at such a position that the position for the candidate (TC2) deviates from the spatially even distribution.