An Image Binarization Segmentation Method Combining Global and Local Threshold for Uneven Illumination Image

• Published in: Intelligent Computing Theories and Application pp. 379 - 390
• Main Authors: Wang, Jin-Wu, Xie, Daiwei, Dai, Zhenmin
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Binarization; Canny operator; Combination of global and local thresholds; Uneven illumination
• ISBN: 9783031138690; 3031138694
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-031-13870-6_31

For images with uneven illumination, the segmentation method based on global threshold will be affected by illumination, and the effect to low contrast and uneven illumination images is poor. In the segmentation method based on local threshold, the image will have discontinuous gray distribution at the boundary of different sub images, resulting in artifacts. In this paper, a binary image segmentation method is proposed, which uses the minimum filter to eliminate the uneven illumination and combines the global threshold with the local threshold according to the edge information of Canny operator. Experiments show that this method can achieve ideal segmentation effect for images with uneven illumination.