An Experimental Comparison of Range Image Segmentation Algorithms
IEEE Transactions on Pattern Analysis and Machine Intelligence
Learning Occupancy Grid Maps with Forward Sensor Models
Autonomous Robots
A Complete U-V-Disparity Study for Stereovision Based 3D Driving Environment Analysis
3DIM '05 Proceedings of the Fifth International Conference on 3-D Digital Imaging and Modeling
Obstacle detection with a Photonic Mixing Device-camera in autonomous vehicles
International Journal of Intelligent Systems Technologies and Applications
Real-Time Approach for Adaptive Object Segmentation in Time-of-Flight Sensors
ICTAI '08 Proceedings of the 2008 20th IEEE International Conference on Tools with Artificial Intelligence - Volume 01
Hi-index | 0.00 |
For modern assistance systems, time of flight (tof) cameras play a key role in the perception of the environmental situation. Due to benefits of this sensor type, especially in combination with video sensors, it outperforms typical sensor systems, like stereo cameras. However, tof depth image segmentation is yet not solved satisfactorily. Common approaches use homogeneous constraints and therefore assume objects to be parallel to the camera plane. This consequently leads to segmentation errors. This paper proposes a fast segmentation algorithm for detecting distinct planes using statistical knowledge. By projecting the depth image data along the image axis, u-v-histogram images can be constructed. It is shown, that in the histogram images, 3D planes correspond to line segments. A fast approach for line segment extraction in the histogram images is used to find the relevant planes. The algorithm was successfully tested with real data under varying conditions and can be applied for different tof camera sensors.