Scrap Image Segmentation for Accurate Automated Grade Discrimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scrap metal discrimination technologies face challenges in efficiently and accurately determining the grade of scrap metal, particularly iron scrap, due to varying shapes and sizes, leading to inconsistent human-based visual discrimination and difficulties in automation, and existing methods are limited to specific scenarios like using magnetic cranes, making them impractical for diverse scrap handling.
Innovation Solution
A scrap discrimination system and method using machine learning models to extract and discriminate scrap parts from camera images, employing semantic segmentation to identify central scrap parts and neural networks to determine grades, with models trained on teacher data to improve accuracy and automate the process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual inspection by workers is used to determine scrap metal grade, then the process can handle various sizes and shapes of scrap pieces, but the discrimination results are inconsistent due to worker skill levels and the process is time-consuming
Solution Approach 1:
The patent replaces the mechanical visual inspection process performed by workers with an automated image processing system using machine learning models. The system captures images of scrap metal and uses trained models to automatically determine grades, eliminating human subjectivity and significantly reducing inspection time while maintaining or improving accuracy.
Solution Approach 2:
The system enables scrap metal grading to be performed autonomously without human intervention. The machine learning models self-evaluate the scrap pieces based on captured images, making the system self-sufficient in performing the discrimination task that previously required skilled workers.
2Extent of automation
If automated image processing is used to discriminate scrap metal grades, then the process can be automated and time reduced, but it is difficult to recognize the overall shape of scrap from captured images
Solution Approach 1:
The patent divides the image processing task into multiple stages using different specialized models. First, a scrap part extraction model identifies and extracts the scrap region from the captured image. Then, a scrap discrimination model analyzes the extracted scrap to determine its grade. This segmentation allows each model to focus on specific aspects, improving overall accuracy.
Solution Approach 2:
The system performs preliminary extraction of the scrap part from the captured image before conducting the actual grade discrimination. This preliminary action prepares the data in an optimal format for the discrimination model, ensuring that the model receives clean, focused input data that improves recognition accuracy.
3Measurement precision
If the scrap discrimination system uses multiple models for extraction and discrimination, then the accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent employs machine learning models that can handle multiple functions within a unified framework. The scrap part extraction model and scrap discrimination model work together in an integrated system that processes different aspects of scrap analysis, achieving multi-functionality without requiring entirely separate systems for each task.
Data Source
AI summary
A scrap discrimination system and a scrap discrimination method that can improve scrap discrimination technology are provided. A scrap discrimination system includes a scrap part extraction model (221) that extracts, based on a camera image, a scrap part located in a central portion included in the camera image with reference to a window (107) defined in advance in an image, a scrap discrimination model (222), generated by teacher data including training images, that sorts grades of scrap and a ratio of each grade from a scrap image extracted by the scrap part extraction model (221), and an output interface (24) that outputs information on the grades of scrap and the ratio of each grade as discriminated based on the scrap image using the scrap discrimination model (222).