Phyton(Austria) Spccial issuct "APGC 2004Ⅲ Vol.4S Fasc 4 (525)―(528) Accurate Estimation of Tree Positions inん 1 102005 αF放 均件 ど θ″ η l i s Forest using Portable lmaging Lidar Data by Hough Transform By Y URAN01)&K OMASA2) K c y w o r d st Hough transfonll、 Japancsc larch(と cr′ た′ ο′ マ β′ qρ,s)ttnrcst、 ponabic ir11 aging lidar、 rcmotc scnsing、 trcc positioll S ti nllll a r v U R A N O Y & O M A S A K 2 0 0 5 A c c u r a t c c slⅢ llation oftrcc positions ill larix lcptolcpis fbr― cst using portablc imaging lldar data by hotigh transftannPhyton(Hom、 ― Austria)45(4):(525)― (528) It、 vas rcpoltcd in IPCC 2000 that forcst is ollc of thC lllost impOItant carbon sinks oll Ea“h Forcsts allcviatc cttccts of globalvarming 、 and absorb air pollutants as vcll 、 Esscntially、 ftlrcst in3alnagcilacnt plays an irnpoltant rolc in tackhng cnvironillcntal problcills Mapping cxact position of trccs in forcst.、 vhich is onc colllpOncnt of forcst illanagcmcnt、 is not casy Duc to ttrcstざ complcx strtictttrc and tilldcrgrowths^mcasuring works arc difictllt and takc longcr tilllc by mantial lllcthod in this stildy、 a ground―bascd rclllotc scnsing techniquc to cstilllatc trcc pOsitions in Sorcst has bccn dcvclopcd ttsing a poltablc illlaging lidar Thc tidar is a portablc lllcastェ ring illstRlmcllt 、 vhich uscs ncar inara― rcd lascr Thc lidar systclll is Only cOntrollcd by onc lllan and lllcasurcillcnts oftrcc positions only takc a shortcr tilllc(l c abOllt 8 11lintltcs) Nonllally.natural trccs do not gro、 v straight and lcan to Othcr dircctions ln this casc、 thc lldar can not dircctly tncastirc thc foot of thc trccs bccatisc of thc tindcrgro、 vths Hcncc thc trcc olll thc ccntcr of thc stcill at nlcastirablc licight(in l130St Cascs thc POSitiOns that wcrc cstilllatcd針 lllcastlrablc hcight is 2 111 or illorc)arc qtlitc dittcrcnt fronl thc lal actモ trcc position and thc cl■ or iS aboll1 0 8i nl lll RMSE(Root Mcan Sqtiarc Error)A Hollgh trallsbrm illcthod Was thcrcttrc tiscd to cnablc thc lllcaStircincnt of icallingvcll trccs as as、 thosc covcrcd、 VthS ith thick tindcrど ro｀ Ⅲ Dcpartlllcnt of Agro― Environnicntal Scicncc. TakushoktェUllivcrsity Hokkaido Jullior va、Hokkaido 074-85ぷ S. Japan Fax i +81-164-23-4411、 c i13all: Collcgc、4558 Mcmti、 Ftlkaga、 uran o“ takuShOku― hc acJp : Dcpanlllcllt Of Biological and EllvirolllllCntal Enginccrinど 、Thc Uni、rcrsitv of Tokvo l― 1_1 rayoi. ヽ Btlllkyo.Tokyo l13-8657、 Japan Fax: +81-3-5思 41-817S c illailt aolllasa“ で11lail ccc tト 的k y o a c J p (526) By cornparing thc cstimatcd trcc positions using Hough transforlll and thc mantial mcas― urements of 24 trees in about 1200 m= Japanese larch(と `′ ′ ll qβ ′′ θ′ qβな)10rcst, statistical icsults vcd an accuracy of0 16 m in RMSE sho、 Introduction An establishcd practicc of mcasuring trcc position is by thc use of somc traditlonal incasuring instltlments such as mcasuring tapcn Clinometcrs‐ and pOckct compasscs(AvERY&BURKHART 1983,HuscH&a1 1982,NAGUMO&MINOWA 1990,NIsHIo 1998)HowcVCr,thc usc of thcse inethods is difflcult and cntalls too much timc and manpowcr(JAPAN ASSOCIATION OF FORESTRY TECHNIQUE 1998)Hencen measurements of each tree by airbome lidar systcm were studied(OMASA&a12000,2003)and nlCasurcmcnts of cach trcc by ground― bascd portablc imaging lidar havc bccn cxamined(OMASA&a1 2002.URANO& OMASA 2003),whiCh rcvcalcd thatthc mcasurcments were much casicr and did not nccd much manpo、 vcr Bascd on thc rcsults‐lllcasurcmcnt clrors comparing to actual trcc posi― tions mcasurcd by cxisting mcthods using traditional instrtlments、vere 0 81 m in RMSE(Root Mcan Squarc Error)(URAN0 2004)An attcmpt was thcrcforc madc to fllrther improve this acctlracy Bccausc tree position uFaS regarded as thc ccntcr oftree stem at a height,if a trcc lcans to othcr dircctions.thc trcc position、 vill bc dittcrent fronl thc actual dcpcnding on lcaning degrec and measurement hcight v c dcvclopcd a method using Hough translom to con― Consequently,in this study、 sider leaning trees and so improvc thc cstimation oftree positlons in Japancsc larch “ ′ (とα′ホ ′ lr'rο 響 is)foreSt With thick undergrOwths using a poHablc imaging lidar in order to detect trec positions of cach Japanese larch by the lidar on the ground. cach stcnl、vas mcasurcd at rnuch highcr level frolll thc bottonl duc to forcsts'com― plex strtlcture and undcrgroM/ths Matcrial alld Mcthods A Japancsc larch forcst、 vith thick undcrgro、 vths、 vas sclcctcd fljr thc study Thc forcst has bccll lllanagcd ftar ccological sttidics for ycars Trccs arc about 40 ycars old lt and 13 to abo、 15 11l in hcight in this torcst、 24 Japancsc larchcs in an arca of about 1200 111=、 vcrc lllcaSurCd usillg 、 LP卜 425HA a poltablc llllaging lidar systcnl produccd by RIEGL Its lllcast11■ blc rangc is bct、 c cn 2 111 and 60 m 、vith rangc accuracy of about 8 nllll. and horizontal and vcnical anglllar acctiracics arc 0 009 dcgrccs Thc lidar、 vas sct at a point、 vhcrc it can rccognisc as illany Japancsc larchcs as pos― siblc lll thc arca Thc iand gradicnt、 vas conil■ llcd by incas、 lrinどa point on thcど round frolll pltiral dircctions by thc lidar Firstl、 vc dctcctcd thc stclll shapc of cach trcc tising 3-dilllcnsional datti ob― t a i n c d b y t h c l i d a r T h c s t c l l l c c n t c r aA s にa w h idc hb、 y t h c s t c m s h a p c a t a l l l c a s u r a b i c ahsc i g h t . 、 cstilllatcd by a 11lcthod dcscribcd by OMASA&a1 2002 1n ordcr to cstilllatc acctiratc trcc positions、 which mcant thc ccntcr of thc trcc bottolll、wc uscd Hough transゴ onll(H()UGH 1962、 NAKAGAヽ IA 1999)ヽ lsinどpoints of stcm ccntcrs in 3-dimcnsional coOrdinatc(Fig l )ThC intcrscction point vccn thc linc calctilatcd by Hoに bct、 lgh tratlsfor】 vas rcどardcd as thc trcc n alld thc grotind surfacc 、 positlon (527) T群 1μ 群 シ Stelll centers Ｅ ︶ 〓 智 宮 一ゥ宮 ト ︵ Trcc p蕊品n/―…'HOugh iine ()2 ()4 ()6 08 1() Rcla市 、c 、一 a 、i s ( m ) tirc Fig i Estilllation of a trcc PoSitiOn inaxis x― by Hotigh transfol■ ll Althotigh this tlど ht) vs x― is y一 alld hciど ollly sho、 axis and hctgh〔 and キlidal data includc 3-dilllclls10nal coordinatcs(x a、 Rcsults and Discusslon vs thc distriblltion of trccs at thc study sitc As a rcsult,cstima, Fig 2 sho、 tion by Hough transform 、vas morc accurate than cstimation by cxisting mcthod vith comparing to thc actual Fig 3 sho、 v s thc crlor variations of both cstilllatlons、 distancc from thc lidar Thc cFOrS Ofthe estinlation by cxisting lllcthod arc incrcas― ing as thc distancc命 olll thc lidar bccomcs lollgcr This is cxpcctcd sincc tllc accu― racy of cxisting illetllod bccomcsvorsc 、 as tllc distancc ionl thc lidar bccollacs ionger On tlle othcr hand、 cstilllation by Hollgh transforttl illlprove the accuracy ドcd an sho、 racy rcgardlcss ofthc distance flonl thc lidar Both statistical rcsilits accl〕 of0 81 m and 0 161n in RMSE.rcspcctively iS(lll) Rciati、lc、― ttヽ 1() n ・× ・ ． × ● 一 × ． ｘ 氏 ■ ● ● × ● ｘ ● Ｘ 文 ｏ 一 ン ︻ ． 洋 ” 滓 狩 一 ユ 一 ● 一 宮︶ ″︼ ′電︲ん。″ ͡ 〓〓●▼一 一 1 使c w 村ャ k■ 1 ●‐ヽc h 臨 l l l n dt l Pl oせ 車 と ● n 0 , ど、 十い uど i trtl応 1い IIn 、 X ,stillatcJ 1(1● Ps、 11 い trFg nlctilod 、Eも、 Fig 2 Trcc positlon lllap OfJapancsc larch lorcst X axis and v― illctan rciati、 c cooiditlatcs Vn Ⅲ ■Ⅲill thc tlgttrc on thc grotilld surfacc as thc lidar sct point is 、 thc origill(、 y)=(00)ShOヽ (528) 15 ︵Ｅ ︶ ﹄９ 一ロ X I卜 せヽ「 ヽ lI〕 1ぷHttCtil(屯 (Rヽ4ヽi=いヽi nl) O l チ 、1 いは「l r a n (村. n l ( R I N l ヽ 1 ‐( ) I lⅢ l 10 ∬艦 枇猟ご謎押子 !:n総増縦lr縛掛18n↑ 碑吐 試単乱温 &碓提品‖ :押1梢i approxllnation Conscqucntly,this study concllldcd that nlcastlrcmcnts by thc lidar and cs― tilllations by Hotlgh transfornl、 vcrc rc8ardcd as an accuratc mcthod to dctcct trcc vith thick undcrgrowths This nlcthOd also givcs Othcr advan― posttions in a forcst、 tagcs stich as fc、 v、vOrkcrs arc llccdcdn takcs lcsser tiine and no spccial skilis and trainings arc rcquircd R cfcrcnccs ア Aヽ ERヽ TE &BuRKHART H E 1983 Forcstillcasurcnicnts ― ヽlcGraw―Hil! Ncwヽ Ork HOtiGH P V C 1962 Mcthod and lllcalls lorrccOgnizing cOmplc、 pattcrns ―U S Patcnt3、 069、654 H u s c H B , M I L L E R C I & B E E R S T ｀ V 1 9 8 2 F o r c s t m c n s u r a t i o n _ VJ iヽl e v N、c 、 v York PCC 200培 R tthご hに 噂 。(mmm pallcl 結 撒 温 1告‖せざiC温 七;晶ぷ も 辞 縄 si:覇 幣 ,° JAPAN ASSOCIATION OF FORESTRY TECHNIQUE 199R SO Mcthod、 Of ftlicst cxamination‐ Tokyo―shosckll ToK)。 苅 抽町 Nlヤ 挫 楓 払ま 料 けヽ 結艦8濫革 艦眠iJ淵姑il撤 さ 艦牧 1紺 岳 ￨ぷ 骨 馬 艦 揮 試 縛齢陥柵温ぜ 縄鮮 肝 徹惣終 】 HaЫ Cimaど inさ 1週 ar JOumalof nc R的 で に S的 l sing So よ 8:i盤 呈 ! 注 営 島 ,° 群 1督 討 ―A持 S溜 1:ぶ i推 品 鮭群出品c2‖ :歓 ま 淵 品 描惜l品 蝉松猟 掛程 慧 需 l諾 ,! 輯 Cf性 ―_Gご :Ψ 持塩 賞淵ぜ鞘結詰 拭 意 tR札 10ギ と さ 始 蒜 cttmに cHhttbn に H Of的 _酌 破 弱 寸 Ю mn口 IS的 ぱ 舶 堤 蹴名 場 逆 i盟 F ttnttng Of hdttm潤 Uい° 樹 品ti品 L詔よ 古 「 樹ぶ 樹下 群札博 t、 艦 深紺:lを i品 1 桃鮒背:」 撫sPm調 掛 津 &骨 酎 恥 対 nh脚 ぃ ∝ 前 離 H by pttmにhttr hnung 鑑益 i樹 i借 1盟 料押 」 と
© Copyright 2024 ExpyDoc
