Research on Estimating Rice Canopy Height and LAI Based on LiDAR Data.

Abstract

Rice canopy height and density are directly usable crop phenotypic traits for the direct estimation of crop biomass. Therefore, it is crucial to rapidly and accurately estimate these phenotypic parameters. To achieve the non-destructive detection and estimation of these essential parameters in rice, a platform based on LiDAR (Light Detection and Ranging) point cloud data for rice phenotypic parameter detection was established. Data collection of rice canopy layers was performed across multiple plots. The LiDAR-detected canopy-top point clouds were selected using a method based on the highest percentile, and a surface model of the canopy was calculated. The canopy height estimation was the difference between the ground elevation and the percentile value. To determine the optimal percentile that would define the rice canopy top, testing was conducted incrementally at percentile values from 0.8 to 1, with increments of 0.005. The optimal percentile value was found to be 0.975. The root mean square error (RMSE) between the LiDAR-detected and manually measured canopy heights for each case was calculated. The prediction model based on canopy height (R 2 = 0.941, RMSE = 0.019) exhibited a strong correlation with the actual canopy height. Linear regression analysis was conducted between the gap fractions of different plots, and the average rice canopy Leaf Area Index (LAI) was manually detected. Prediction models of canopy LAIs based on ground return counts (R 2 = 0.24, RMSE = 0.1) and ground return intensity (R 2 = 0.28, RMSE = 0.09) showed strong correlations but had lower correlations with rice canopy LAIs. Regression analysis was performed between LiDAR-detected canopy heights and manually measured rice canopy LAIs. The results thereof indicated that the prediction model based on canopy height (R 2 = 0.77, RMSE = 0.03) was more accurate.

Plant phenotyping relevance

LiDARを用いてイネのキャノピー高とLAIを非破壊推定するプラットフォームを構築し、手測定との精度検証も行っており、表現型取得手法が研究の中心です。

Code and data availability

The supplied article blocks describe LiDAR-based rice canopy height and LAI estimation, but contain no public phenotype dataset, point cloud data, images, author code repository, or trained model deposit. No data or code availability statement with an authors' public URL appears anywhere in the supplied text; only the

No evidence-backed public reproduction asset is currently recorded.