Image Analysis Based Pixel Area Calculation Of Tree Leaf

Tree leaf area reflects the growth status of trees and it's an important parameter in forestry applications, so fast and accurate measurement of tree leaf area is of great significance. With the development of image processing, computer-based image analysis is becoming the new method of tree leaf area calculation.In this paper, how to calculate the pixel area of tree leaf with simple background and with natural background is studied. For simple background images, extracting tree leaf image from the background using automatic threshold algorithm, then labeling the connected region of the image using 4-neighborhood pixels connected region labeling method, and calculating the pixel area of leaf. For natural background image, it can be found that there exist significant texture feature differences between the target leaf and natural background. Due to the focus of the camera, the target leaf appears clear and the background becomes blurred. Therefore, the fractal feature is used as an important texture feature in the proposed algorithm. For leaf images with trunk background, the differences are smaller in texture but bigger in color between leaf and the background. In this case, color based fuzzy clustering segmentation algorithm is utilized for image segmentation, and then get the leaf pixel area through the post-processing. Experimental results demonstrate that the proposed algorithms for leaf pixel area calculating are available and accuracy.