Vegetation Horizontal Occlusion Index (VHOI) from TLS and UAV Image to Better Measure Mangrove LAI

Xianxian Guo,Jinyan Tian,Chen Shi,Le Wang,Dameng Yin,Sheng Nie

doi:10.3390/rs10111739

Abstract

Accurate measurement of the field leaf area index (LAI) is crucial for assessing forest growth and health status. Three-dimensional (3-D) structural information of trees from terrestrial laser scanning (TLS) have information loss to various extents because of the occlusion by canopy parts. The data with higher loss, regarded as poor-quality data, heavily hampers the estimation accuracy of LAI. Multi-location scanning, which proved effective in reducing the occlusion effects in other forests, is hard to carry out in the mangrove forest due to the difficulty of moving between mangrove trees. As a result, the quality of point cloud data (PCD) varies among plots in mangrove forests. To improve retrieval accuracy of mangrove LAI, it is essential to select only the high-quality data. Several previous studies have evaluated the regions of occlusion through the consideration of laser pulses trajectories. However, the model is highly susceptible to the indeterminate profile of complete vegetation object and computationally intensive. Therefore, this study developed a new index (vegetation horizontal occlusion index, VHOI) by combining unmanned aerial vehicle (UAV) imagery and TLS data to quantify TLS data quality. VHOI is asymptotic to 0.0 with increasing data quality. In order to test our new index, the VHOI values of 102 plots with a radius of 5 m were calculated with TLS data and UAV image. The results showed that VHOI had a strong linear relationship with estimation accuracy of LAI (R2 = 0.72, RMSE = 0.137). In addition, as TLS data were selected by VHOI less than different thresholds (1.0, 0.9, …, 0.1), the number of remaining plots decreased while the agreement between LAI derived from TLS and field-measured LAI was improved. When the VHOI threshold is 0.3, the optimal trade-off is reached between the number of plots and LAI measurement accuracy (R2 = 0.67). To sum up, VHOI can be used as an index to select high-quality data for accurately measuring mangrove LAI and the suggested threshold is 0.30.

Highlights

Leaf area index (LAI), defined as half of the total leaf area per unit ground surface area [1], serves as a key indicator of carbon and nutrient cycling, rates of energy exchange between plants and the atmosphere, and ecological processes such as photosynthesis and transpiration [2,3,4]
This research presents an exploration of the ability of vegetation horizontal occlusion index (VHOI) based on terrestrial laser scanning (TLS) dataset and high-resolution unmanned aerial vehicle (UAV) image to quantitatively represent data quality and screen out the poor-quality data in terms of plot-level LAI retrieval
By analyzing the relationship of VHOI and estimation accuracy of TLS-derived LAI (R2 = 0.72, RMSE = 0.137), the study demonstrated that VHOI can quantitatively characterize TLS data quality

Summary

Introduction

Leaf area index (LAI), defined as half of the total leaf area per unit ground surface area [1], serves as a key indicator of carbon and nutrient cycling, rates of energy exchange between plants and the atmosphere, and ecological processes such as photosynthesis and transpiration [2,3,4]. The indirect methods in situ involve specially designed optical instruments, such as hemispherical photography [14,15] and LI-COR’s Plant Canopy Analyzer [16,17,18], which retrieve LAI based on gap fraction that is defined as the probability of light passing through the forest canopy without being blocked by foliage [19]. These indirect in situ methods provide accurate LAI measurement with easy and quick operation. They are susceptible to sky illumination and weather conditions, limiting the time flexibility [14,16,20]

Objectives

Results

Conclusion