Machine Vision Color Classification Using Masked Region Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine vision systems struggle to distinguish between colors of illuminated objects, particularly in applications like car dashboards where colored LEDs appear similar due to shape and material effects, leading to inconsistent color determination.
Innovation Solution
A method involving color-balanced image processing, where a color-balanced image is received, a region of interest is identified, and a mask image is generated to isolate the lit area, allowing for analysis of image attributes and color classification using a trained classifier.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a machine vision system directly analyzes illuminated areas without preprocessing, then the analysis process is simple, but color distinction accuracy deteriorates due to shape and material effects
Solution Approach 1:
The patent applies segmentation by dividing the illuminated area into multiple regions based on intensity thresholds. The lit area is segmented from the background and further divided into regions with different intensity characteristics, allowing the system to analyze color properties in each segment separately rather than treating the entire illuminated area as a single unit. This resolves the contradiction by enabling accurate color distinction through region-based analysis while keeping the processing methodology systematic and manageable.
Solution Approach 2:
The patent extracts the essential color information from the illuminated areas by isolating the lit regions through thresholding and masking operations. By extracting only the relevant illuminated pixels and their color properties, the system eliminates interference from shape variations and material effects. This extraction approach improves color distinction accuracy by focusing solely on the chromatic characteristics of the light sources.
2Measurement precision
If histogram analysis is applied to illuminated areas with different shapes, then the analysis covers the entire lit region, but color determination becomes inconsistent due to shape variations
Solution Approach 1:
The patent applies local quality by analyzing color properties in specific regions of the illuminated area rather than treating the entire lit region uniformly. By dividing the lit area into multiple regions and analyzing each region's color characteristics separately, the system achieves consistent color determination that is independent of the overall shape. Each local region's color properties are evaluated on their own merits, making the analysis adaptable to various shapes while maintaining consistency.
Solution Approach 2:
The patent changes the analysis parameters by using intensity-based region division instead of shape-based analysis. By transforming the problem from shape-dependent histogram analysis to intensity-based regional color analysis, the system achieves parameter invariance to shape changes. The color determination relies on chromatic parameters of separated regions rather than on the geometric parameters of the overall illuminated area.
3Measurement precision
If the camera angle changes relative to angled LEDs, then the lit area position and shape change in the image, but color analysis becomes unreliable
Solution Approach 1:
The patent applies dynamics by making the region of interest definition adaptive to changes in camera angle and lighting conditions. Instead of using fixed geometric regions, the system dynamically determines the lit area boundaries through intensity thresholding that automatically adjusts to the current viewing angle. This dynamic approach ensures that the color analysis always focuses on the actual illuminated regions regardless of their position or shape in the image, achieving angle independence.
Data Source
AI summary
System and method for distinguishing colors of illuminated objects using machine vision. A color-balanced image that includes at least one lit area is received, as well as an indication of a region of interest that includes one of the one or more lit areas. A mask image is generated based on the region of interest. A color-balanced image of the region of interest is generated by masking the color-balanced image with the mask image, and a plurality of image attributes for the region of interest is determined by analyzing the color-balanced image of the region of interest. A color is determined based on the plurality of image attributes using a trained classifier, and the determined color stored, e.g., in a memory medium.