A Method for Tree Image Segmentation Combined Adaptive Mean Shifting with Image Abstraction

• Published in: Journal of information processing systems 2020, 16(6), 66, pp.1424-1436
• Main Authors: Ting-ting Yang, Su-yin Zhou, Ai-jun Xu, Jian-xin Yin
• Format: Article
• Language: English
• Subjects: 컴퓨터학
• ISSN: 1976-913X; 2092-805X
• DOI: 10.3745/JIPS.02.0151

Although huge progress has been made in current image segmentation work, there are still no efficientsegmentation strategies for tree image which is taken from natural environment and contains complexbackground. To improve those problems, we propose a method for tree image segmentation combining adaptivemean shifting with image abstraction. Our approach perform better than others because it focuses mainly onthe background of image and characteristics of the tree itself. First, we abstract the original tree image usingbilateral filtering and image pyramid from multiple perspectives, which can reduce the influence of thebackground and tree canopy gaps on clustering. Spatial location and gray scale features are obtained by stepdetection and the insertion rule method, respectively. Bandwidths calculated by spatial location and gray scalefeatures are then used to determine the size of the Gaussian kernel function and in the mean shift clustering. Furthermore, the flood fill method is employed to fill the results of clustering and highlight the region ofinterest. To prove the effectiveness of tree image abstractions on image clustering, we compared differentabstraction levels and achieved the optimal clustering results. For our algorithm, the average segmentationaccuracy (SA), over-segmentation rate (OR), and under-segmentation rate (UR) of the crown are 91.21%,3.54%, and 9.85%, respectively. The average values of the trunk are 92.78%, 8.16%, and 7.93%, respectively. Comparing the results of our method experimentally with other popular tree image segmentation methods, oursegmentation method get rid of human interaction and shows higher SA. Meanwhile, this work shows apromising application prospect on visual reconstruction and factors measurement of tree. KCI Citation Count: 0