Multi-resolution point cloud generation based on heterogeneous sensor fusion system

Various three-dimensional environment representation techniques exist depending on the type of sensor, and the level of expression varies depending on the characteristics of the sensor. Our sensor system is composed of a three-dimensional lasers scanner and a stereo camera. A rigid transformation matrix representing a three-dimensional relative pose is obtained by calibration between a stereo camera and laser scanner and creates a multi-resolution point cloud. By combining a stereo camera with a high resolution and a laser scanner with wide range and low-resolution, it is able to complement each other and achieve better results.