Rice Plant Height Estimation Using a Non-Survey Grade Laser Scanner: A Hokuyo UTM 30LX Case Study

Abstract

This study evaluated the estimation of rice plant height without ground surface detection using a HOKUYO UTM 30LX laser scanner. The selected laser scanner is small and lightweight (approximate 0.37 kg) with a rectangular footprint. Field observations were conducted in 2016 with a line scanner installed above the rice canopy. Based on results of our investigation, the laser pulse of the HOKUYO UTM 30LX finds it more difficult to reach the ground surface because of the 4-cm long rectangular footprint. Although the laser scanner is small and lightweight, and the selected target areas in this study were larger than in the previous study, relatively accurate results were still obtained. More specifically, the top of the rice plant was found to correspond to p^t = 1, and the bottom of the rice plant was chosen at four percentile ranks ( p^b = 70, p^b = 80, p^b = 95, and p^b = 99). The relative vertical distance ( ) was identified as the distance between the bottom and top of the rice plant. According to the results,  strongly correlated with H with r² ≥ 0.86 in all cases, and the root mean squared error was 1.0 cm. The slope of the regression line was only close to 1.0 (r² ≥ 0.96) for p^b = 99. The plant height exceeded the relative vertical distance by 22.0 cm. The accuracy of the results largely depends on the laser footprint size of the specific laser scanner used, and larger laser footprints correspond to larger bias between rD and H. Results of this study, which confirmed the viability and feasibility of our proposed method for estimating rice plant height using the HOKUYO UTM 30 LX, are presented and discussed.