Automated Construction Progress Monitoring via 2D Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current construction progress monitoring systems lack automation, relying heavily on user intervention and often fail to accurately compare 3D point clouds from construction sites to building information models due to insufficiently detailed or accurate 3D models, particularly for elements like pipes and ducts.
Innovation Solution
A method utilizing a deep neural network to produce a 3D progress model by assigning probabilities of element presence and state to 3D points, aggregating confidences from multiple images, and refining estimates using primitive templates, enabling automated visualization and quantification of construction element placement and progress.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If geometric comparison of 3D points to 3D BIM is used, then automated progress monitoring is achieved, but measurement precision deteriorates due to insufficiently detailed or accurate 3D models
Solution Approach 1:
The patent introduces 2D images as an intermediary medium between the construction site and the progress monitoring system. Instead of directly comparing 3D point clouds to 3D BIM models, the system uses 2D images captured from the site, processes them through computer vision algorithms, and compares the extracted 2D element information with 2D drawings or BIM projections. This intermediary approach allows for more accurate detection of construction elements like pipes and ducts, as the 2D images provide direct visual evidence of element presence, location, and state, overcoming the limitations of insufficiently detailed 3D models.
Solution Approach 2:
The patent replaces the traditional mechanical/geometric comparison system with a computer vision-based system. Instead of relying on geometric algorithms to compare 3D point clouds directly, the system uses image processing, machine learning classifiers, and optical recognition techniques to identify construction elements in 2D images. This substitution enables more robust and accurate detection of construction progress, as computer vision algorithms can better handle variations in element appearance, lighting conditions, and perspective that plague direct 3D geometric comparison.
2Extent of automation
If 3D point cloud comparison is used, then automated monitoring is enabled, but device complexity increases due to need for detailed 3D BIM models
Solution Approach 1:
The patent extracts the essential information needed for progress monitoring from the complex 3D BIM model and represents it in a simplified 2D format. By extracting element locations, dimensions, and states from the 3D model and projecting them onto 2D drawings or plans, the system reduces the complexity requirement. This allows automated monitoring to function with simpler, more manageable 2D representations rather than requiring fully detailed and accurate 3D BIM models, thus reducing device complexity while maintaining automation capability.
Solution Approach 2:
The patent transitions the progress monitoring problem from a three-dimensional 3D space to a two-dimensional 2D representation. Instead of working with complex 3D point clouds and 3D BIM models, the system uses 2D images captured from the construction site and compares them with 2D drawings or plans. This dimensional reduction simplifies the data structures and processing requirements, making the automated monitoring system more manageable and less complex, while still providing accurate progress detection through the 2D/2D comparison approach.
3Measurement precision
If manual verification is used, then measurement precision is maintained, but productivity decreases due to user intervention requirements
Solution Approach 1:
The patent implements a self-service automated monitoring system that performs progress assessment without requiring manual verification. The system autonomously captures 2D images from the construction site, processes them through computer vision algorithms, identifies construction elements and their states, and compares them with project documentation. This self-service capability maintains measurement precision by using robust image processing and machine learning techniques, while dramatically improving productivity by eliminating the need for manual inspection and verification, allowing continuous automated monitoring throughout the construction process.
Data Source
AI summary
A method of automatically producing maps and measures that visualize and quantify placement and progress of construction elements, such as walls, ducts, etc. in images. From a set of images depicting a scene, element confidences per pixel in each of the images are produced using a classification model that assigns such confidences. Thereafter, element confidences for each respective one of a set of 3D points represented in the scene are determined by aggregating the per-pixel element confidences from corresponding pixels of each of the images that is known to observe the respective 3D points. These element confidences are then updated based on primitive templates representing element geometry to produce a 3D progress model of the scene.