An approach for the calibration of a combined RGB-sensor and 3D-camera device

Abstract

The elds of application for 3d cameras are very dierent, because high image frequency and determination of 3d data. Often, 3d cameras are used for mobile robotic. They are used for obstacle detection or object recognition. So they also are interesting for applications in agriculture, in combination with mobile robots. Here, in addition to 3d data, there is often a necessity to get color information for each 3d point. Unfortunately, 3d cameras do not capture any color information. Therefore, an additional sensor is necessary, such as RGB plus possibly NIR. To combine data of two dierent sensors a reference to each other, via calibration, is important. This paper presents several calibration methods and discuss their accuracy potential. Based on a spatial resection, the algorithm determines the translation and rotation between the two sensors and the inner orientation of the used sensor.