Development of Visible and Near Infrared Camera System for Mounting on Small UAV

Abstract

We developed a camera system that can acquire image data simultaneously in both visible and infrared bands. Two small-sized single-board computers (Raspberry Pi A+) were used to control the camera modules, obtain position data, and communicate with a PC control system on the ground. The control software on this system was developed and written in the Python programming language. Since the weight of this camera system is less than 300g including the battery, it can be mounted on a small-sized Unmanned Aerial Vehicle (UAV) platform. In order to use this camera system in various research fields, it is useful to convert the output digital number (DN) to represent the reflectance in each band. The reflectance of observed targets can be estimated from the measured relation between the input power and output DN of the camera system. The reflectance extracted from images taken by this camera system was compared with the reflectance measured by the spectral radiometer. Although the correlation of the two reflectances was high, the slope and offset of the regression line in some bands should be improved. This was probably caused by the estimated error of the spectral characteristics in each band, since the spectral characteristics of the camera modules were not available to the public. As an example of the application for this camera system, we introduced a Normalized Difference Vegetation Index (NDVI) measurement of a paddy field during the growth stage from May to October in 2016. It was found that the time series of NDVI change observed in the paddy field was similar to that observed in a past study.