A New Individual Tree Crown Delineation Method for High Resolution Multispectral Imagery

Abstract

In current individual tree crown (ITC) delineation methods for high-resolution multispectral imagery, either a spectral band or a brightness component of the multispectral image is employed in delineation with reference to edges or shapes of crowns, whereas spectra of tree crowns are seldom taken into account. Such methods normally perform well in coniferous forests with obvious between-crown shadows, but fail in dense deciduous or mixed forests, in which tree crowns are close to each other, between-crown shadows and boundaries are unobvious, whereas adjacent tree crowns may be of distinguishable spectra. In order to effectively delineate crowns in dense deciduous or mixed forests, a new ITC delineation method using both brightness and spectra of the image is proposed in this study. In this method, a morphological gradient map of the image is first generated, treetops of multi-scale crowns are extracted from the gradient map and refined regarding the spectral differences between neighboring crowns, the gradient map is segmented using a watershed approach with treetops as markers, and the resulting segmentation map is refined to yield a crown map. Evaluated on images of a rainforest and a deciduous forest, the proposed method more accurately delineated adjacent broad-leaved tree crowns with similar brightness but different spectra than the other two typical ITC delineation algorithms, achieving a delineation accuracy of up to 76% in the rainforest and 63% in the deciduous forest.

A New Individual Tree Species Recognition Method Based on a Convolutional Neural Network and High-Spatial Resolution Remote Sensing Imagery

“…Our study area was located in a park with flat terrain, where trees were manually planted and taken care of by gardeners. In addition, compared with natural forests with abundant and The CSI tree crown delineation produced inevitable errors due to under-and oversegmentation, which are common problems in tree crown delineations [35], especially when only two-dimensional information is used. Some non-crown segments produced by the under-and over-segmentation affected the classification map and reduced classification accuracies.…”

“…In this study, the method of visual evaluation was adopted. Adjusting the parameters in the algorithm, the match between the delineation lines and the real tree crowns was visually assessed to obtain the most accurate delineation map [35].…”

Monitoring spring phenology of individual tree crowns using drone‐acquired NDVI data

“…In order for the presented methodology to be scalable to larger study areas, some logistical concerns must be addressed. The manual digitizing of individual crowns should ideally be replaced either by object‐based segmentation of crowns and automated species classification making use of the fine spatial resolution textural and multi‐temporal dataset (Xu et al, 2020, example of workflows which could be modified for drone datasets: Guirado et al, 2019; Qiu et al, 2020; Sheeren et al, 2016; Wagner et al, 2018), or auxiliary datasets such as detailed forest inventories, for example based on laser scanning data (Burt et al, 2019; Duncanson et al, 2014). A further limiting factor for the deployment of drones to improve upon phenocam‐based monitoring of phenology is their timely deployment and operational costs considering the fieldwork effort involved for reaching remote sites and the susceptibility to weather conditions.…”

Assessment of Tree Detection Methods in Multispectral Aerial Images

“…Their results are promising and will undoubtedly boost the exploration of fully automatic approaches. Employing multispectral imagery, Qiu et al [51] introduced an individual tree delineation map on multispectral images from cameras on a UAS overflying a forest stand. Using the gradient map, they extracted treetops and refined the delineation employing spectral differences.…”