3D Object Rotation-Based Mechanical Parts Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate identification and classification of components using a 2D camera under insufficient lighting is challenging, especially for rotating objects, as edges may be undetectable, leading to misidentification due to lack of orientation information.
Innovation Solution
The method involves obtaining and storing 2D images of objects at different white balance values and rotation states, using these images to recognize objects through comparison with a captured image, even under insufficient lighting, by providing images to an image recognition module for matching and weight-based prioritization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If 2D camera is used under insufficient lighting, then device complexity is reduced, but image recognition accuracy deteriorates due to undetectable edges
Solution Approach 1:
The system performs preliminary white balance adjustment on images before recognition. By pre-processing images to correct color temperature and lighting conditions, the system enables accurate edge detection and recognition even under insufficient or variable lighting conditions, maintaining measurement precision without increasing device complexity
Solution Approach 2:
The system changes the white balance parameter of images to adapt to different lighting conditions. By adjusting the white balance value, the system optimizes image quality for recognition purposes, allowing accurate identification of mechanical parts despite variations in lighting intensity and color temperature
2Ease of operation
If orientation information is not captured during image acquisition, then ease of operation is improved, but object identification accuracy deteriorates due to misidentification
Solution Approach 1:
The system transitions from 2D image analysis to 3D object recognition by generating multiple rotated versions of the input image. This dimensional approach allows the system to identify objects regardless of their orientation in the original image, maintaining ease of operation while improving identification accuracy through multi-angle comparison
Solution Approach 2:
The system creates a universal recognition database that includes images of mechanical parts in multiple rotation states. This multi-functional approach enables a single image capture to be used for identifying parts in any orientation, eliminating the need for precise orientation control during image acquisition
3Measurement precision
If multiple rotated images are generated for recognition, then object identification accuracy is improved, but processing time increases
Solution Approach 1:
The system applies partial action by generating and comparing only the necessary number of rotated images based on the specific recognition task. Rather than exhaustively processing all possible rotations, the system generates a sufficient set of rotated images to achieve accurate identification, optimizing the balance between recognition accuracy and processing time
Solution Approach 2:
The system performs preliminary generation and storage of rotated images in a database before actual recognition tasks. This pre-processing allows rapid retrieval and comparison during operation, reducing real-time processing time while maintaining high recognition accuracy through pre-computed rotated versions
Data Source
AI summary
Technologies are generally described for 3D object recognition through 2D image processing based on white balancing and object-rotation in machine vision systems. According to some examples, image recognition of an object captured with a camera under insufficient lighting may be achieved through white balancing. Processing cost reduction may be achieved in the learning process for image recognition through automatic generation of rotated 2D images of target objects to be detected, such as machine parts, from a small number of 2D images of a target object and generation of a 3D image of the target object from the rotated 2D images. Image recognition may thus be ensured even under insufficient lighting through execution of the image recognition process for multiple images and learning the successful recognition results. Some examples may be implemented in mechanical parts selection, where 2D images of the parts may be available beforehand.