Probabilistic terrain classification in unstructured environments

Häselich, Marcel; Arends, Marc; Wojke, Nicolai; Neuhaus, Frank; Paulus, Dietrich

doi:10.1016/j.robot.2012.08.002

Cited by 35 publications

(15 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Three‐dimensional LiDARs are widely applied on ALVs for obstacle detection. A type of velodyne HDL‐64E LiDAR is widely adopted in mainstream ALVs, such as in Google self‐driving car, KIT's AnnieWAY (Kammel et al., ), Stanford's Junior (Montemerlo et al., ), the vehicle in paper (Häselich et al., ), and the vehicle in our own team (Chen et al., ) (shown in Figure ).…”

Section: A Novel Setup Methods Of 3‐d Lidarsmentioning

confidence: 99%

LiDAR Based Negative Obstacle Detection for Field Autonomous Land Vehicles

Shang

et al. 2015

J. Field Robotics

View full text Add to dashboard Cite

Negative obstacles for field autonomous land vehicles (ALVs) refer to ditches, pits, or terrain with a negative slope, which will bring risks to vehicles in travel. This paper presents a feature fusion based algorithm (FFA) for negative obstacle detection with LiDAR sensors. The main contributions of this paper are fourfold: (1) A novel three‐dimensional (3‐D) LiDAR setup is presented. With this setup, the blind area around the vehicle is greatly reduced, and the density of LiDAR data is greatly improved, which are critical for ALVs. (2) On the basis of the proposed setup, a mathematical model of the point distribution of a single scan line is deduced, which is used to generate ideal scan lines. (3) With the mathematical model, an adaptive matching filter based algorithm (AMFA) is presented to implement negative obstacle detection. Features of simulated obstacles in each scan line are employed to detect the real negative obstacles. They are supposed to match with features of the potential real obstacles. (4) Grounded on AMFA algorithm, a feature fusion based algorithm is proposed. FFA algorithm fuses all the features generated by different LiDARs or captured at different frames. Bayesian rule is adopted to estimate the weight of each feature. Experimental results show that the performance of the proposed algorithm is robust and stable. Compared with the state‐of‐the‐art techniques, the detection range is improved by 20%, and the computing time is reduced by an order of two magnitudes. The proposed algorithm had been successfully applied on two ALVs, which won the champion and the runner‐up in the “Overcome Danger 2014” ground unmanned vehicle challenge of China.

show abstract

Section: A Novel Setup Methods Of 3‐d Lidarsmentioning

confidence: 99%

LiDAR Based Negative Obstacle Detection for Field Autonomous Land Vehicles

Shang

et al. 2015

J. Field Robotics

View full text Add to dashboard Cite

show abstract

“…An effective approach to object segmentation from 2D images and 3D point clouds is the Markov random field (MRF) algorithm [24–30]. …”

Section: Related Workmentioning

confidence: 99%

Intuitive Terrain Reconstruction Using Height Observation-Based Ground Segmentation and 3D Object Boundary Estimation

Song

Cho

et al. 2012

Sensors

View full text Add to dashboard Cite

Mobile robot operators must make rapid decisions based on information about the robot’s surrounding environment. This means that terrain modeling and photorealistic visualization are required for the remote operation of mobile robots. We have produced a voxel map and textured mesh from the 2D and 3D datasets collected by a robot’s array of sensors, but some upper parts of objects are beyond the sensors’ measurements and these parts are missing in the terrain reconstruction result. This result is an incomplete terrain model. To solve this problem, we present a new ground segmentation method to detect non-ground data in the reconstructed voxel map. Our method uses height histograms to estimate the ground height range, and a Gibbs-Markov random field model to refine the segmentation results. To reconstruct a complete terrain model of the 3D environment, we develop a 3D boundary estimation method for non-ground objects. We apply a boundary detection technique to the 2D image, before estimating and refining the actual height values of the non-ground vertices in the reconstructed textured mesh. Our proposed methods were tested in an outdoor environment in which trees and buildings were not completely sensed. Our results show that the time required for ground segmentation is faster than that for data sensing, which is necessary for a real-time approach. In addition, those parts of objects that were not sensed are accurately recovered to retrieve their real-world appearances.

show abstract

“…At low level, lidar has been fused with other range‐based sensors (lidar and radar) using a joint calibration procedure (Underwood, Hill, Peynot, & Scheding, ). Additionally, lidar has been fused with cameras (monocular, stereo, and thermal) by projecting 3D lidar points onto corresponding images and concatenating either their raw outputs (Dima, Vandapel, & Hebert, ; Wellington, Courville, & Stentz, ) or precalculated features (Häselich, Arends, Wojke, Neuhaus, & Paulus, ). This approach potentially leverages the full potential of all sensors, but suffers from the fact that only regions covered by all modalities are defined.…”

Section: Introductionmentioning

confidence: 99%

Multimodal obstacle detection in unstructured environments with conditional random fields

Kragh

Underwood

2019

Journal of Field Robotics

View full text Add to dashboard Cite

Reliable obstacle detection and classification in rough and unstructured terrain such as agricultural fields or orchards remains a challenging problem. These environments involve large variations in both geometry and appearance, challenging perception systems that rely on only a single sensor modality. Geometrically, tall grass, fallen leaves, or terrain roughness can mistakenly be perceived as nontraversable or might even obscure actual obstacles. Likewise, traversable grass or dirt roads and obstacles such as trees and bushes might be visually ambiguous. In this paper, we combine appearance‐ and geometry‐based detection methods by probabilistically fusing lidar and camera sensing with semantic segmentation using a conditional random field. We apply a state‐of‐the‐art multimodal fusion algorithm from the scene analysis domain and adjust it for obstacle detection in agriculture with moving ground vehicles. This involves explicitly handling sparse point cloud data and exploiting both spatial, temporal, and multimodal links between corresponding 2D and 3D regions. The proposed method was evaluated on a diverse data set, comprising a dairy paddock and different orchards gathered with a perception research robot in Australia. Results showed that for a two‐class classification problem (ground and nonground), only the camera leveraged from information provided by the other modality with an increase in the mean classification score of 0.5%. However, as more classes were introduced (ground, sky, vegetation, and object), both modalities complemented each other with improvements of 1.4% in 2D and 7.9% in 3D. Finally, introducing temporal links between successive frames resulted in improvements of 0.2% in 2D and 1.5% in 3D.

show abstract

Probabilistic terrain classification in unstructured environments

Cited by 35 publications

References 39 publications

LiDAR Based Negative Obstacle Detection for Field Autonomous Land Vehicles

LiDAR Based Negative Obstacle Detection for Field Autonomous Land Vehicles

Intuitive Terrain Reconstruction Using Height Observation-Based Ground Segmentation and 3D Object Boundary Estimation

Multimodal obstacle detection in unstructured environments with conditional random fields

Contact Info

Product

Resources

About