Image Edge Detection Using Logarithmic Grayscale Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional edge detection methods are ineffective in images with non-uniform and uneven illumination, such as underwater glimmer images, as they are overly sensitive to background noise and struggle to distinguish between noise and target pixels.
Innovation Solution
An image edge detection method that involves convolving pixels with logarithm-transformed grayscale values using Prewitt operators to calculate brightness differential approximations, followed by gradient magnitude calculation and threshold comparison to identify edge pixels, while also employing piecewise linear and logarithmic conversions to enhance grayscale values and reduce noise interference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional edge detection methods are used on images with non-uniform illumination, then the detection process is simple, but the method is overly sensitive to background noise and cannot effectively distinguish target pixels from noise
Solution Approach 1:
The patent applies preliminary grayscale enhancement through piecewise linear conversion and logarithmic conversion before edge detection. This preprocessing step adjusts the grayscale distribution to suppress background noise and enhance target features, making the subsequent edge detection more reliable in non-uniform illumination conditions
Solution Approach 2:
The patent transforms the grayscale values using logarithmic conversion (g'(x,y) = log(g(x,y) + 1)) to change the parameter distribution of pixel intensities. This parameter transformation compresses the dynamic range and reduces the impact of non-uniform illumination, improving noise resistance while maintaining detection accuracy
2Measurement precision
If grayscale values are enhanced through piecewise linear and logarithmic conversions, then anti-interference capability is improved, but the processing time and computational complexity increase
Solution Approach 1:
The grayscale enhancement is divided into two separate conversion stages: piecewise linear conversion followed by logarithmic conversion. This segmentation allows each conversion to be optimized independently and applied only where needed, improving overall processing efficiency while maintaining detection precision
Solution Approach 2:
The piecewise linear conversion applies different enhancement factors to different grayscale ranges, applying stronger enhancement only to specific ranges where target features are located. This partial action approach avoids unnecessary computation on already well-illuminated regions, reducing overall processing time
Data Source
AI summary
The edge detection method according to an embodiment of the present disclosure may include: convolving pixels to be processed, based on templates in respective predetermined directions for edge detection operators and a logarithm of grayscale values for respective pixels adjacent to the pixels to be processed, so as to obtain brightness differential approximations in the respective predetermined directions for the pixels to be processed; calculating gradient magnitudes of the pixels to be processed, based on the brightness differential approximations in the respective predetermined directions for the pixels to be processed; comparing the gradient magnitudes of the pixels to be processed with a gradient magnitude threshold, and selecting the pixels to be processed whose gradient magnitudes are larger than or equal to the gradient magnitude threshold as edge pixels; and extracting a contour of a target within an image, based on the resultant edge pixels.