應用於樹木修剪之影像處理技術

Image Processing Technology Applied to Tree Pruning

李承鴻(CHANG-HUNG LEE)

指導教授 ： 朱元三

國立中正大學/工學院/電機工程學系/碩士(2016年)

摘要

台灣每年平均遭受3至4個颱風侵襲，每當颱風來臨前政府負責行道樹的管理單位即會發包將樹木修剪防颱，但常修剪過度，形成只剩主幹的情形。修剪的方式應以外實內虛為主，保持樹木外觀形體的自然式修剪，將內部不良的樹枝做修剪，以減少風的阻力，也使得樹木依然能維持生命機能。

希望利用現有的影像處理的技術，將色彩空間轉換為LAB色彩空間，且選擇切割效果較好的色彩通道，並運用色彩特徵、影像濾波(模糊)、OTSU自動二值化處理、灰階影像進行型態學中的侵蝕、擴張，盡量保持圖像細節特徵的條件下，將圖像中的雜訊除去，使得樹木影像從背景中，單純切割出樹木的枝幹分布位置，再透過主樹幹定位、體積估算、平衡係數估算，估算出當樹木在受到風吹襲下的平衡度，且讓我們知道必須優先對哪部分的樹枝做修剪，最後使用不良枝的判別方法，在影像中找尋適合且符合自然式修剪的修剪處，並計算修剪後的平衡係數，兼顧平衡和自然式修剪的方式，一階段一階段的修剪樹枝，最終呈現修剪完畢的影像。

關鍵字

LAB色彩空間 ； 影像切割 ； 樹木修剪

並列摘要

Taiwan suffered an annual average of 3-4 typhoons. Before the typhoon come, the department of street trees management will be pruning trees against the typhoon. But often excessive pruning, trees only left main part of trunk. We need to use a natural pruning way to prune trees, it good for maintaining the appearance of trees. The internal bad tree branches be pruned to reduce wind resistance.
We want to use the existing image processing techniques, the color space conversion to LAB color space, and better channel of color features, image filters (blur), OTSU automatic binary processing and Morphological operations not only to keep the image detail, but also to remove the noise. So we can segment the trees form the image. Then through the main trunk positioning, volume estimates, balance coefficient estimates, we can estimate the balance of the trees when wind is blowing, and let us know which part of the branches must be pruning first.
Finally, we use the method of determining bad branches for finding suitable pruning position. The way combined the balance of the trees and the natural method, to prune trees step by step. In the end, we will render pruning finished image.

並列關鍵字

LAB color space ； image segmentation ； tree pruning

參考文獻

[2] Colour Management Consultancy，http://www.colourphil.co.uk/index.shtml

[4] C. Tomasi and R. Manduchi, “Bilateral Filtering for Gray and Color Images”, IEEE International Conference on Computer Vision, 1998, pp. 839-846.

[5] NOBUYUKI OTSU. “A threshold selection method from gray-level histogram”,

[7] Chen Zhongshan, and Yan Tu. “Improved Image Segmentation Algorithm based on OTSU Algorithm”, International Journal of Advancements in Computing Technology, 2012.

[10] Feng Chen, Yun-Feng li, “A Multiresolution Segmentation method for Tree Crown Image using Wavelets”, Proceedings of the 2007 International Conference on Wavelet Analysis and Pattern Recognition, 2007, pp. 1556-1559.