無人空中載具數位影像應用於林分蓄積量之推估

Estimating the Stand Stocks by Unmanned Aerial Vehicle Imagery

朱宗威(Tsung-Wei Chu)

指導教授 ： 陳朝圳 ； 陳建璋

國立屏東科技大學/農學院/森林系所/碩士(2014年)

https://doi.org/10.6346/NPUST.2014.00223

摘要

過去森林資源調查消耗許多人力、金錢及時間。遙感探測技術的提升，開始利用航照影像進行森林資源調查，節省成本的消耗。無人空中載具(Unmanned Aerial Vehicle, UAV)具有快速、即時性及低成本特性，可快速拍攝雲下高解析力航空照片。本研究地區位於屏東縣臺糖人工林，以UAV航照影像及樣區資料為研究材料，比較不同航照重疊度使用Pixd4D軟體正射接合差異，再以UAV航空照片進行林分性態值萃取，並建立立木材積式及空中材積式，最後以本研究試驗流程設計UAV技術進行人工林調查流程。在不同航照重疊度比較，發現以80%航照重疊度取得之影像位移較小，使用UAV航照進行林分性態值萃取，將樹種所測得林分性態值，與地面樣區量測資料進行迴歸分析，桃花心木樹冠幅之決定係數(R2)為0.9、樹高之決定係數(R2)為0.97；光臘樹樹冠幅之(R2)為0.79、樹高之(R2)為0.85，臺灣櫸樹冠幅之決定係數(R2)為0.63、樹高之決定係數(R2)為0.79。以UAV影像所萃取之樹冠鬱閉度及樹高為變數，所建立之各樹種空中材積式其決定係數R2分別為，桃花心木為0.70、光臘樹為0.71，臺灣櫸為0.50。

關鍵字

無人空中載具 ； 航照測量 ； 林分性態值萃取

並列摘要

Past research has indicated that forest resources inventory consumes lots of manpower, money and time. With the enhancement of remote sensing technology, using aerial images for forest resources measure that could let cost savings. In this study, we use Pingtung County TaiSugar plantation as a study area. Use UAV image and field sampling data as the research materials. We use orthophotos with Pix4D software to compare the difference of overlapping. This study will use UAV aerial photographs to extract the stand characteristics and establish the volume formula and aerial volume formula of standing timber. Finally, we use research trial to design the process workflow with UAV technology for plantation survey. Results showed that the 80% overlapping will have the smallest distorting. The study is based on the data collected by UAV aerial image and field sampling plot. We use regression analysis to establish the relationship between the stand characteristics extracting from UAV aerial image and field sampling plot surveying. Results showed that crown and tree height can directly extract from UAV aerial image, the coefficient of crown determination (R2) were Swietenia macropnylla 0.90, Fraxinus griffithii 0.79 and Zelkova serrata 0.72 and coefficient of tree height Swietenia macropnylla 0.97, Fraxinus griffithii 0.85 and Zelkova serrata 0.63 respectively. Meanwhile, the aerial stand volume table had been established by independent variables of crown closure and tree height extracting from UAV image that the coefficient of volume (R2) was Swietenia macropnylla 0.70, Fraxinus griffithii 0.71 and Zelkova serrata 0.50.

並列關鍵字

UAV ； photogrammetry ； stand characteristics extract

參考文獻

林金樹、黃家培、柯建全、張建禕 (2004) 應用高解析力衛星影像偵測與描繪開闊地立木樹冠之研究。中華林學季刊 38(2): 163-178。

邱志明、呂明倫、彭炳勳 (2012) 森林資源調查技術演進。林業研究專訊 19(5): 19-24。

陳正豐、洪富文、張添榮 (1996) 三種桉樹在臺灣東部地區生長之研究。臺灣林業科學11(3): 233-238。

彭炳勳 (2009) 應用空載光達資料推測阿里山地區單木樹高與林分高度之研究。中華林學季刊 42(1): 167-180。

謝依達、吳守從、鄭祖安、魏浚紘、鍾玉龍、陳建璋 (2013) 氣象因子對臺東地區柳杉與樟樹人工林生長之影響。作物、環境與生物資訊 11(3): 209-216。