Radiometric Calibration of an Inexpensive LED-Based Lidar Sensor.

Jordan Laughlin,Craig Glennie,Preston Hartzell,Jan W Kovermann

doi:10.3390/s20185215

Jordan Laughlin, Craig Glennie + Show 2 more

Open Access

https://doi.org/10.3390/s20185215

Copy DOI

Abstract

Radiometric calibration of laser-based, topographic lidar sensors that measure range via time of flight or phase difference is well established. However, inexpensive, short-range lidar sensors that utilize non-traditional ranging techniques, such as indirect time of flight, may report radiometric quantities that are not appropriate for existing calibration methods. One such lidar sensor is the TeraRanger Evo 60 m by Terabee, whose reported amplitude measurements do not vary smoothly with the amount of return signal power. We investigate the performance of a new radiometric calibration model, one based on a neural network, applied to the Evo 60 m. The proposed model is found to perform similarly to those applied to traditional lidar sensors, with root mean square errors in predicted target reflectance of no more than ±6% for non-specular surfaces. The radiometric calibration model provides a generic approach that may be applicable to other low-cost lidar sensors that report return signal amplitudes that are not smoothly proportional to target range and reflectance.

Highlights

Radiometric calibration of lidar intensity to physical units such as relative reflectance is commonly practiced in the remote sensing community to extract target spectral information at the wavelength of the lidar light source [1,2,3,4,5,6,7]
We investigate a radiometric calibration model based on a simple neural network that empirically approximates the relationship between target reflectance and the measurements reported by the
A performance summary for each sensor’s trained radiometric calibration model is provided in Table 5, where calibration model predicted reflectance values are compared to those measured with a spectroradiometer

Summary

Introduction

Radiometric calibration of lidar intensity to physical units such as relative reflectance is commonly practiced in the remote sensing community to extract target spectral information at the wavelength of the lidar light source [1,2,3,4,5,6,7]. The sensor is economically priced under $150 and, among other observables, reports a return signal amplitude measurement These characteristics make the Evo 60 m a candidate for affordable, simultaneous acquisition of target range and reflectance if a suitable radiometric calibration model can be developed. Laser-based lidar sensor, the power received, Pr , by the sensor is related to the physical characteristics of the illuminated target, acquisition geometry, the atmospheric environment, and system characteristics. This relationship, often termed the lidar range equation, is given in [14] as: Pr =. Solving for target reflectance and inverting the value within Ccal leads to:

Methods

Results

Conclusion