English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Individual tree detection and crown delineation (ITDD) are critical in forest inventory management and remote sensing based forest surveys are largely carried out through satellite images. However, most of these surveys only use 2D spectral information which normally has not enough clues for ITDD. To fully explore the satellite images, we propose a ITDD method using the orthophoto and digital surface model (DSM) derived from the multi-view satellite data. Our algorithm utilizes the top-hat morphological operation to efficiently extract the local maxima from DSM as treetops, and then feed them to a modified superpixel segmentation that combines both 2D and 3D information for tree crown delineation. In subsequent steps, our method incorporates the biological characteristics of the crowns through plant allometric equation to falsify potential outliers. Experiments against manually marked tree plots on three representative regions have demonstrated promising results – the best overall detection accuracy can be 89%. Introduction Forest is one of the most important land surfaces and plays an important role in the global ecosystem. Timely and accurate measurements of the forest parameters at the individual tree level such as tree count, tree height, and crown size are essential for quantitative analysis of forest structure, ecological modeling, biomass estimation, and evaluation of deforestations (Mohan et al., 2017; Weng et al., 2015). Over the past several decades, remote sensing techniques have greatly improved the capability of extracting forest metrics with high spatial resolution imagery (Gonçalves et al., 2017; Sousa et al., 2015). However, the majority of these methods use spectral or texture information and are limited to the radiometric quality which makes them vulnerable to erroneous detections, either over-/undersegmenting tree crowns at the individual plot level. The addition of 3D information to these forest metric estimations can greatly enhance the measurement accuracy. Recently, much attention has been given to lidar data, which provides an accurate 3D representation of the surface objects. A number of algorithms have been proposed to analyze the forest structure at individual tree level with this data (Ferraz et al., 2016; Kathuria et al., 2016; Liu et al., 2015). Many of the methods that are based on either lidar or photogrammetric 3D points use the normalized digital surface model (nDSM) or the canopy height model (CHM, for the forest applications), which can naturally highlight the treetops and directly offer the tree heights (Liu et al., 2015; Lu et al., 2014). The CHM can be generated by subtracting the digital terrain model (DTM) from the digital surface model (DSM). Similar to the 2D image-based methods, the CHM-based methods use procedures such as image smoothing, local maximum localization, and template matching to detect the individual trees and their boundaries (Koch et al., 2006; Popescu et al., 2002). In Strîmbu and Strîmbu (2015), they used graph theory to model the forest topological structure and correct two potentially over-identified treetops. Also, multi-scale segmentations have been proposed to dynamically select the best set of apices and generate the final segmentation (Véga et al., 2014). For tree crown delineation, image segmentation methods such as valley following, region growing, and watershed segmentation can be directly used on CHM (Ferraz et al., 2016; Kathuria et al., 2016; Strîmbu and Strîmbu, 2015). Among these methods, the watershed segmentation is the most popular as it can naturally and efficiently model the treetops and crowns, for example, Liu et al. (2015) proposed the Fishing Net Dragging (FND) method which uses the watershed segmentation with the Gaussian filtering to find the boundaries of trees. Even though these methods have demonstrated great successes in their applications, they are limited by the cost of lidar surveying making them impractical for repetitive acquisition of at large scales (Wulder et al., 2013). Also, lidar systems offering compensatory spectrometers are not widely available to provide additional spectrum data for more advanced analysis, such as LAI (leaf area index), NDVI (normalized difference vegetation index) for vegetation classification. A demand for such data normally requires an additional light for multi-/hyperspectral data acquisition while which subsequently brings in registration issues. Changlin Xiao is with the Future Cities Laboratory, SingaporeETH Centre, ETH Zurich, 1 Create Way, CREATE Tower, #06-01, 138602, Singapore; and the Department of Civil, Environmental and Geodetic Engineering, The Ohio State University, 218B Bolz Hall, 2036 Neil Avenue, Columbus, OH 43210. Rongjun Qin is with the Department of Civil, Environmental and Geodetic Engineering, The Ohio State University, 218B Bolz Hall, 2036 Neil Avenue, Columbus, OH 43210; and the Department of Electrical and Computer Engineering, The Ohio State University, 205 Dreese Labs, 2015 Neil Avenue, Columbus, OH, 43210 (qin.324@osu.edu). Xiao Xie is with the Research Center for Industrial Ecology & Sustainability, Institute of Applied Ecology, Chinese Academy of Sciences. No.72, Wenhua Road, Shenhe District, Shenyang City 110016, China; and the Key Laboratory for Environment Computation & Sustainability of Liaoning Province. No.72, Wenhua Road, Shenhe District, Shenyang City 110016, China. Xu Huang is with the Department of Civil, Environmental and Geodetic Engineering, The Ohio State University, 218B Bolz Hall, 2036 Neil Avenue, Columbus, OH 43210. Photogrammetric Engineering & Remote Sensing Vol. 85, No. 1, January 2019, pp. 51–xxx. 0099-1112/18/51–xxx © 2018 American Society for Photogrammetry and Remote Sensing doi: 10.14358/PERS.85.1.xxx PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING J anuar y 2019 51 Pre-print, the final version is subject to minor edits

Individual Tree Detection and Crown Delineation with 3D Information from Multi-view Satellite Images