—In this paper, an adaptive thresholding techniquebased on gray level co-occurrence matrix (GLCM) ispresented to handle images with fuzzy boundaries. AsGLCM contains information on the distribution of graylevel transition frequency and edge information, it is veryuseful for the computation of threshold value. Here thealgorithm is designed to have flexibility on the edgedefinition so that it can handle the object’s fuzzyboundaries. By manipulating information in the GLCM, astatistical feature is derived to act as the threshold value forthe image segmentation process. The proposed method istested with the starfruit defect images. To demonstrate theability of the proposed method, experimental results arecompared with three other thresholding techniques.
展开▼