Assessing the utility of NAIP digital aerial photogrammetric point clouds for estimating canopy height of managed loblolly pine plantations in the southeastern United States

被引：5

作者：

Ritz, Alison L. ^{[1
]}

Thomas, Valerie A. ^{[1
]}

Wynne, Randolph H. ^{[1
]}

Green, P. Corey ^{[1
]}

Schroeder, Todd A. ^{[2
]}

Albaugh, Timothy J. ^{[1
]}

Burkhart, Harold E. ^{[1
]}

Carter, David R. ^{[1
]}

Cook, Rachel L. ^{[3
]}

Campoe, Otavio C. ^{[4
]}

Rubilar, Rafael A. ^{[5
]}

Rakestraw, Jim ^{[6
]}

机构：

[1] Virginia Polytech Inst & State Univ, Forest Resources & Environm Conservat, Blacksburg, VA 24061 USA

[2] US Forest Serv, USDA, Southern Res Stn, Knoxville, TN 37919 USA

[3] North Carolina State Univ, Dept Forestry & Environm Resources, Raleigh, NC 27695 USA

[4] Univ Fed Lavras, Lavras, MG, Brazil

[5] Univ Concepcion, Dept Silvicultura, Cooperat Prod Forestal, Concepcion, Chile

[6] Int Paper, Statesboro, GA 30458 USA

来源：

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION | 2022年 / 113卷

关键词：

Remote sensing; Height estimation; Photogrammetry; Canopy height model; SURFACE MODELS; FOREST STANDS; TREE HEIGHT; IMAGES; ATTRIBUTES; LIDAR; ALS;

D O I：

10.1016/j.jag.2022.103012

中图分类号：

TP7 [遥感技术];

学科分类号：

081102 ; 0816 ; 081602 ; 083002 ; 1404 ;

摘要：

Remote sensing offers many advantages to supplement traditional, ground-based forest measurements, such as limiting time in the field and fast spatial coverage. Data from airborne laser scanning (lidar) have provided accurate estimates of forest height, where, and when available. However, lidar is expensive to collect, and wall-to-wall coverage in the United States is lacking. Recent studies have investigated whether point clouds derived from digital aerial photogrammetry (DAP) can supplement lidar data for estimating forest height due to DAP's lower costs and more frequent acquisitions. We estimated forest heights using point clouds derived from the National Agricultural Imagery Program (NAIP) DAP program in the United States to create a predicted height map for managed loblolly pine stands. For 534 plots in Virginia and North Carolina, with stand age ranging from 1 year to 42 years old, field-collected canopy heights were regressed against the 90th percentile of heights derived from NAIP point clouds. Model performance was good, with an R2 of 0.93 and an RMSE of 1.44 m. However, heights in recent heavily thinned stands were consistently underestimated, likely due to between-row shadowing leading to a poor photogrammetric solution. The model was applied to non-thinned evergreen areas in Virginia, North Carolina, and Tennessee to produce a multi-state 5 m x 5 m canopy height map. NAIP-derived point clouds are a viable means of predicting canopy height in southern pine stands that have not been thinned recently.

引用

页数：13