Adaptive thresholding in dynamic scene analysis for extraction of fine line

This paper presents an adaptive threshold method whereby a fine thin line of one-pixel width lines could be detected in a gray level images. The proposed method uses the percentage difference between the mean of the pixels within a window and the center pixel. The minimum threshold value however is...

Full description

Bibliographic Details
Main Authors: Syed Abu Bakar, Syed Abdul Rahman, Green, R. J.
Format: Article
Language:English
Published: 2000
Subjects:
Online Access:http://eprints.utm.my/1983/
http://eprints.utm.my/1983/1/SyedAbdulRahman2000_AdaptiveThresholdingInDynamicScene.pdf
Description
Summary:This paper presents an adaptive threshold method whereby a fine thin line of one-pixel width lines could be detected in a gray level images. The proposed method uses the percentage difference between the mean of the pixels within a window and the center pixel. The minimum threshold value however is heuristically set to 32. If the percentage difference is greater than 40% then the threshold value will be set to the difference value. This method has been applied in detecting moving objects with fine lines and the results showed that the method was able to pickup straight thin edges that belong to the moving object