3D Robot Vision Object Detection Across Changing Camera Positions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems struggle to accurately detect objects when the positional relationship between the imaging device and the object changes, leading to potential failures or increased detection time.
Innovation Solution
An image processing system that detects objects by acquiring the positional relationship between the imaging device and the object using robot positional information, and storing a model pattern as three-dimensional positional information based on this relationship. The system then executes a detection process by matching feature points from the detected image with the stored model pattern.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the imaging device is moved by a robot to change its position relative to the object, then the detection coverage and flexibility are improved, but the detection accuracy deteriorates when the positional relationship differs from the taught position
Solution Approach 1:
The patent transforms the two-dimensional image coordinate system into a three-dimensional robot coordinate system by introducing depth information through the robot's positional data. This dimensional transformation allows the system to accurately locate objects in 3D space regardless of the imaging device's position, resolving the contradiction between detection coverage and accuracy.
Solution Approach 2:
The patent introduces an intermediary coordinate transformation process that acts as a bridge between the image coordinate system and the robot coordinate system. By using the robot's positional information as an intermediary, the system can accurately map detected objects to the robot's working space, maintaining detection accuracy across different positions.
2Adaptability or versatility
If the positional relationship between the imaging device and the object changes, then the system's flexibility is improved, but the detection reliability deteriorates due to mismatch with taught model patterns
Solution Approach 1:
By incorporating the third dimension (depth) from robot positional information into the coordinate system, the system can maintain reliable object detection across varying positions. The 3D transformation ensures that the model pattern matching remains accurate regardless of changes in the imaging device's position relative to the object.
3Device complexity
If traditional 2D image coordinate system transformation is used, then the system complexity is kept low, but the detection precision deteriorates when positional relationships change
Solution Approach 1:
The patent enhances the traditional 2D coordinate transformation by adding the third dimension using robot positional information. This relatively simple extension from 2D to 3D transformation significantly improves detection precision when positional relationships change, without substantially increasing system complexity.
Data Source
AI summary
The present invention addresses the problem of providing an image processing system that can accurately and efficiently detect a target object even when the positional relationship between an image capturing device and the target object differs between the time of teaching and the time of detection. An image processing system 1 includes: a control unit 52 that obtains, on the basis of position information of a robot 2 for specifying the position of a visual sensor 4 in a robot coordinate system and on the basis of position information showing the position of a target object W in an image coordinate system, the positional relationship between the visual sensor 4 and the target object W; and a storage unit 51 that stores, on the basis of a model pattern consisting of feature points extracted from a teaching image and on the basis of the positional relationship between the visual sensor 4 and the target object W when capturing the teaching image, the model pattern in the form of three-dimensional position information. The control unit 52 performs detection processing, in which the target object W is detected from a detection image on the basis of a result obtained by matching the model pattern with the feature points extracted from the detection image including the target object W.