Adaptive Edge Extraction Using Local Threshold Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing devices face challenges in reliably extracting edges from images due to varying luminance levels caused by environmental factors, leading to inconsistent edge detection across different regions of an image.
Innovation Solution
An image processing device divides the image into multiple partial regions, each with unique threshold values for edge extraction, allowing for adaptive edge detection based on local pixel value distributions and updating these values dynamically based on the number of edges extracted in each region.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the same threshold value is used for edge extraction in the entire image region, then the processing is simple and fast, but edges cannot be reliably extracted in regions with varying luminance levels
Solution Approach 1:
The image region targeted for edge extraction is divided into multiple partial regions (e.g., first through fourth partial regions). Different threshold values are assigned to each partial region based on its specific luminance characteristics. This segmentation allows the system to adapt to varying luminance conditions in different parts of the image, thereby improving edge extraction reliability without requiring a single complex adaptive threshold for the entire image.
Solution Approach 2:
Each partial region is assigned a threshold value tailored to its local luminance characteristics. For example, regions with lower luminance (such as shadowed areas) use lower threshold values, while regions with higher luminance use higher threshold values. This local quality approach ensures that edge extraction is optimized for each specific region's conditions, resolving the contradiction between reliability and complexity by avoiding a globally complex adaptive system.
2Reliability
If a low threshold value is used for edge extraction, then more edges can be detected in dark regions, but unnecessary edges are extracted in bright regions
Solution Approach 1:
The image is segmented into multiple partial regions with different luminance characteristics. By applying different threshold values to each segment, the system can use lower thresholds in dark regions to ensure complete edge detection while using higher thresholds in bright regions to filter out false positives. This spatial segmentation resolves the contradiction by allowing region-specific threshold optimization.
Solution Approach 2:
The threshold value is adjusted according to the local luminance characteristics of each partial region. In dark regions where edge detection might be incomplete, lower thresholds ensure all relevant edges are captured. In bright regions where noise and false edges are more prevalent, higher thresholds eliminate unnecessary edge detections. This local adaptation eliminates false positives while maintaining detection completeness.
3Object-generated harmful factors
If a high threshold value is used for edge extraction, then false positives are reduced in bright regions, but required edges are missed in dark regions
Solution Approach 1:
The image region is divided into multiple partial regions, allowing different threshold values to be applied to dark and bright regions separately. This segmentation enables the system to maintain high thresholds in bright regions to reduce false positives while using lower thresholds in dark regions to ensure complete edge detection, thus resolving the contradiction between reducing false positives and maintaining detection completeness.
Solution Approach 2:
The threshold value is locally adapted to match the luminance characteristics of each partial region. In dark regions, lower thresholds ensure that all required edges are detected despite the lower signal strength. In bright regions, higher thresholds filter out noise and false positives. This local quality adjustment resolves the contradiction by optimizing threshold selection for each region's specific conditions.
Data Source
AI summary
Provided is an image processing device in which a region targeted for edge extraction of a taken image is divided into a plurality of partial regions 11(1) to 11(7) for which threshold values for edge extraction different from each other are set, respectively. An edge extraction unit 5, which extracts edges from the region targeted for edge extraction, performs processing of extracting the edges in the partial regions 11(1) to 11(7) by using the threshold values set so as to correspond to the each of the partial regions 11(1) to 11(7).