Work Data Management System Using Image Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing work site traceability systems struggle to provide detailed insights into worker activities, requiring cumbersome sensor wear and lacking efficient data visualization for improving production efficiency and product quality.
Innovation Solution
A work data management system that classifies worker tasks into detailed time-series steps using image data from network cameras, eliminating the need for sensor wear by employing a system with a detailed step classification unit and data visualization unit for intuitive display of work processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If IC tag reading devices are arranged at each input and output port to improve traceability, then product traceability is improved, but detailed work content by worker cannot be grasped
Solution Approach 1:
The work process is segmented into multiple detailed steps (e.g., taking out a component, assembling, tightening, cleaning) that are sequentially identified and recorded. This segmentation allows comprehensive tracking of both product movement and detailed work content without requiring additional sensors on workers.
Solution Approach 2:
Instead of directly measuring worker actions with sensors, the system uses image data copying from cameras to recognize and record work steps. The image data serves as a copy of the actual work scene, enabling detailed work content analysis without physical intervention.
2Loss of information
If acceleration sensors and position sensors are worn by workers to capture work data, then work content can be identified, but time and effort are significantly increased
Solution Approach 1:
The system extracts work step information from image data captured by fixed cameras, removing the need to attach sensors to workers. This extraction approach eliminates the time-consuming process of sensor installation while still capturing detailed work content through image analysis.
Solution Approach 2:
The patent replaces the mechanical sensor-wearing system with an optical measurement system using image data from cameras. This substitution eliminates physical contact with workers, reducing installation time and effort while maintaining accurate work content detection.
3Ease of operation
If fixed cameras are used to capture work site images, then no worker intervention is needed, but detailed work steps cannot be identified from general images
Solution Approach 1:
The system dynamically adjusts image analysis parameters and processing methods based on the captured images to accurately identify specific work steps. This dynamic processing enables precise work step recognition from general camera images without requiring workers to wear sensors or change their behavior.
Data Source
AI summary
There is provided a work data management system which includes a detailed step classification unit that classifies work performed by a worker at a work site into a plurality of time-series detailed steps based on work data including image data of the work site and a data visualization unit that causes a result display unit to display at least the image data and the plurality of time-series detailed steps.