Construction Progress Recognition via Temporal Image Difference Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for recognizing progress at construction sites, such as Japanese Laid-Open Patent Publication No. 2017-107443, may fail to accurately capture the latest progress status, especially when rework occurs or the image-capturing spot changes.
Innovation Solution
A difference recognition method and system that acquires first and second captured-image data at different time points, recognizes differences based on these images, and uses a machine learning model to identify changes, including rework, even if the image-capturing spot differs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If progress status is determined based on a single latest captured image, then the determination process is simple and fast, but the accuracy of progress recognition deteriorates when rework occurs or image-capturing spot changes
Solution Approach 1:
The patent segments the progress determination process into multiple steps: acquiring multiple captured images at different time points, determining differences between images, and using a learning model to recognize progress status based on these differences. This segmentation allows the system to handle rework and location changes by comparing temporal changes rather than relying on a single image.
Solution Approach 2:
The patent performs preliminary actions by acquiring and storing multiple captured images at different time points before the actual progress determination. These pre-acquired images serve as a database for later comparison, enabling the system to accurately detect changes even when rework occurs or capturing locations differ.
2Measurement precision
If multiple captured images at different time points are used to recognize progress, then the accuracy of progress recognition improves, but the data processing time and computational load increase
Solution Approach 1:
The system pre-acquires and stores multiple captured images at different time points, preparing them in advance for processing. This preliminary data collection enables faster processing during actual progress determination, as the images are already ready for comparison without needing to acquire them in real-time during the analysis phase.
Solution Approach 2:
The patent uses captured images as data copies to represent the construction site state at different times. By working with these image copies rather than directly processing the physical site, the system can efficiently compare temporal changes through image processing algorithms, reducing the time needed for actual progress analysis.
3Adaptability or versatility
If a learning model is used to recognize differences between images, then the adaptability to different construction sites and conditions improves, but the model training complexity and data requirements increase
Solution Approach 1:
The learning model is designed to be universal, capable of recognizing progress status across different construction sites, work types, and conditions. The model learns from diverse training data representing various construction scenarios, enabling it to adapt to new sites without requiring site-specific retraining, thus achieving multi-functionality.
Solution Approach 2:
The patent utilizes parameter changes in the learning model to adapt to different construction sites. The model adjusts its recognition parameters based on the specific characteristics of each construction site, allowing it to maintain high accuracy across varied conditions while managing training complexity through parameter optimization rather than complete model redesign.
Data Source
AI summary
A difference recognition method and a difference recognition system that correctly recognizes the latest progress status. A difference recognition system recognizes the latest progress status of a construction site from a difference, which indicates progress of a construction site. The difference recognition system includes a data management unit that acquires first captured-image data captured in a first state at a first time point and second captured-image data captured in a second state at a second time point that is subsequent to the first time point, and a work determination unit serving as a difference recognition unit that recognizes a newly installed building element from the first captured-image data and second captured-image data as a difference between the first state and the second state. The task determination unit determines the latest progress status from the recognized difference.