Slow Change Detection System for Rust and Frost Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to detect slow changes in monitored objects, such as rust, frost, or stains, which develop gently over time, as they are often masked by sudden changes like human movement, making it difficult to isolate and identify these changes automatically.
Innovation Solution
A slow change detection system that acquires consecutive images of a monitored object, extracts a reference image area, and uses a change detection unit to identify slow changes by excluding sudden changes from the detected area, displaying the slow change areas on the images for visual recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image comparison methods are used to detect changes in monitored objects, then sudden changes like intruders and abnormal operations can be detected, but slow changes such as rust, frost, or stains cannot be effectively detected because they are masked by sudden changes
Solution Approach 1:
The patent segments the change detection process into two distinct pathways: sudden change detection and slow change detection. By dividing the monitoring function into separate processing streams, the system can apply different detection algorithms and time-window analyses to each type of change, thereby achieving high precision for both sudden and slow changes without mutual interference
Solution Approach 2:
The patent performs preliminary classification of changes by analyzing the temporal characteristics and magnitude of detected changes. By preliminarily identifying whether a change is sudden or slow based on initial detection data, the system can then apply appropriate follow-up analysis methods, ensuring that slow changes are not missed amidst sudden changes
2Reliability
If visual observation by observers is used to detect changes, then both sudden and slow changes can be detected, but the process requires significant labor and is prone to human error
Solution Approach 1:
The patent implements self-service detection through automated image processing algorithms that independently analyze consecutive images, extract change areas, classify change types, and generate detection results without human intervention. The system serves itself by automatically adjusting detection parameters and filtering false positives, thereby achieving high reliability while maintaining operational simplicity
3Extent of automation
If automatic image recognition processing is used to detect changed portions, then the detection process is automated, but slow changes are still difficult to isolate from sudden changes
Solution Approach 1:
The patent introduces dynamic adjustment mechanisms that automatically modify detection sensitivity and time-window parameters based on the detected change pattern. When slow changes are detected, the system dynamically extends the analysis time-window and adjusts sensitivity parameters to enhance detection precision, while maintaining high automation throughout the process
Data Source
AI summary
A slow change detection system includes: an image acquisition unit adapted to acquire a photographed image including consecutive images of a monitored object; a reference image acquisition unit adapted to acquire a reference image corresponding to the photographed image of the monitored object; a reference image area extraction unit adapted to extract, from the reference image, a reference image area that is an area corresponding to a photographed area of the photographed image; a change detection unit adapted to acquire a change area that is an area of the photographed image which is different from the reference image area and acquire a slow change area by excluding, from the acquired change area, a sudden change area derived from a history of the change area; and a display control unit adapted to superimpose and display information indicating the slow change area on the photographed image.