Defect Detection Method Using Image Segmentation for Parallel Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deep learning network architectures for product defect detection face challenges in balancing detection precision and speed, as high-resolution images result in reduced detection speed and lower resolution images compromise precision, failing to meet industrial speed requirements.
Innovation Solution
A product defect detection method involving a computer device with a storage device and processor that divides input images into preset blocks, using a defect recognition model trained on corresponding block positions, allowing for simultaneous detection and improving detection rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the resolution of the input image is increased, then detection precision is improved, but detection speed is reduced
Solution Approach 1:
The patent divides the input image into multiple blocks (e.g., 4 blocks) and processes each block independently through the defect recognition model. This segmentation allows the system to maintain high detection precision by analyzing detailed regions while improving overall detection speed by parallel processing multiple regions simultaneously, thus resolving the contradiction between precision and speed.
Solution Approach 2:
The patent introduces a spatial dimension by dividing the image into blocks and assigning different blocks to different processing threads or models. This dimensional transformation from single-image processing to multi-block parallel processing enables simultaneous achievement of high precision and high speed by leveraging computational parallelism.
2Speed
If the size of the input image is reduced, then detection speed is improved, but detection precision is reduced
Solution Approach 1:
Instead of reducing the entire image size, the patent segments the image into multiple smaller blocks that are processed in parallel. Each block maintains sufficient resolution for precise defect detection, while the parallel processing of multiple blocks achieves the speed benefits of smaller input sizes. This resolves the contradiction by distributing the processing workload across multiple independent units.
3Measurement precision
If high-resolution images are used, then detection precision is improved, but industrial speed requirements are not met
Solution Approach 1:
The patent divides the image into multiple blocks and processes them concurrently using multiple processing threads or models. This segmentation strategy enables the system to maintain high detection precision through detailed analysis of each block while achieving industrial-speed throughput by parallel processing multiple blocks simultaneously, thus resolving the contradiction between precision and productivity.
Solution Approach 2:
The patent implements continuous processing by feeding image blocks into the defect recognition model in a streamlined manner, with results aggregated to produce the final detection output. This continuous pipeline approach ensures that high-resolution analysis is performed without interrupting the flow of processing, maintaining both precision and industrial speed requirements.
Data Source
AI summary
A product defect detection method which includes acquiring a detection image of a product to be detected is provided. The method further includes dividing the detection image into a first preset number of detection blocks. Once a detection result of each detection block is obtained by inputting each detection block into a preset defect recognition model, according to a position of each detection block in the detection image, a detection result of the product is determined according to the detection result of each detection block.