Electronic Counter Scale Verification Using Deep Learning and 3D Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The verification of electronic counter scales is labor-intensive due to the need for annual re-inspection of properties like weighing performance, repeatability, bias load, and discriminatory power, which is a significant workload for commodity exchange venues.
Innovation Solution
An intelligent verification method using deep learning for electronic counter scales, involving scale pan and indicating value detection, enables unmanned verification by collecting images and depth data to determine spatial positions and indicating values, utilizing a three-axis mechanical arm for loading weights.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional manual verification methods are used for electronic counter scales, then verification can be performed, but the workload is large and labor-intensive
Solution Approach 1:
The patent replaces manual mechanical verification operations with an automated system that uses deep learning models (YOLOv5 for scale pan detection, CRNN for indicating value recognition) to automatically detect, identify, and verify scale parameters. The system uses cameras to capture images of the scale pan and display, processes them through neural networks, and automatically records verification data, eliminating the need for manual measurement and recording operations.
Solution Approach 2:
The verification system performs self-service by automatically detecting its own working state and verifying scale parameters without human intervention. The system autonomously captures images, processes them through detection models, extracts scale pan positions and indicating values, and records verification results, making the verification process self-contained and independent of manual operations.
2Reliability
If annual re-inspection is conducted for weighing performance and other properties, then verification requirements are met, but it creates a significant workload for commodity exchange venues
Solution Approach 1:
The patent creates a multi-functional verification system that can simultaneously measure multiple scale properties (weighing performance, repeatability, bias load, discriminatory power) using a single integrated apparatus. The system uses one camera to capture both scale pan position and indicating value, processes both types of data through neural networks, and records all verification parameters in a unified manner, eliminating the need for multiple separate verification devices.
Solution Approach 2:
The patent introduces deep learning models as intermediary components that bridge the gap between raw image data and verification results. The YOLOv5 model intermediates to detect scale pan positions, the CRNN model intermediates to recognize indicating values, and these processed results then feed into the verification calculation, simplifying the overall system architecture by using AI algorithms to handle complex data processing tasks.
3Ease of operation
If manual detection of scale pan position and indicating value is performed, then verification can be completed, but manual labor is required
Solution Approach 1:
The patent replaces manual visual detection and measurement operations with an automated image processing system. Cameras capture images of the scale pan and display, deep learning models automatically detect positions and recognize values, and the system automatically calculates verification results, eliminating all manual detection and measurement operations while maintaining high accuracy.
Data Source
AI summary
The present disclosure discloses an electronic counter scale intelligent verification method based on deep learning detection and identification, which includes: collecting a top image, top depth data, and a front image of the electronic counter scale, and selecting a reading display apparatus on the front image; inputting the top image into a scale pan detection and identification model to obtain a scale pan target detection result, verifying the result with the top depth data, obtaining a spatial position of a scale pan, and outputting the spatial coordinates of the four vertices of the scale pan; inputting the top image into an indicating value detection and identification model to obtain indicating value characters and a position of the display apparatus, correcting similar characters and low-confidence characters, and obtaining an indicating value of the electronic counter scale; and performing unmanned verification of weighing performance, repeatability, bias load, and discrimination.