Automated Extraction of Crater Rims on 3D Meshes Combining Artificial Neural Network and Discrete Curvature Labeling

Christoff, Nicole; Jordá, L.; Viseur, Sophie; Bouley, Sylvain; Manolova, Agata; Mari, Jean-Luc

doi:10.1007/s11038-020-09535-7

Cited by 6 publications

(1 citation statement)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many researchers [39,40] proposed different parameters to evaluate the performance of the CDA. These methods work by counting craters detected in different states, i.e., true positives (TPs) corresponding to real detected craters and false positives (FPs) corresponding to detected craters that do not exist.…”

Section: Simulation Verification On Sandmentioning

confidence: 99%

Robust Crater Detection Algorithm Based on Maximum Entropy Threshold Segmentation

Ping

Deng

2022

International Journal of Aerospace Engineering

View full text Add to dashboard Cite

For future lunar exploration and planetary missions, the digital elevation model (DEM) of the target object may not be well prepared before the mission, so developing a new robust crater detection algorithm (CDA) without prepared high-precision DEM is needed to meet the requirements of a high-reliability and high-precision detection and navigation system. In this paper, we presented a new robust lunar CDA method based on maximum entropy threshold segmentation. By calculating the entropy distribution of the ternary image, the threshold for retaining the maximum amount of image information is selected adaptively, a variety of evaluation indicators are proposed, and a multiple-indicator constraint matrix is constructed to realize the extraction and fitting of the craters. The proposed method has the following advantages: (1) it has strong robustness and is capable of extracting complete craters under multiple illumination conditions, which makes it suitable for the extraction of large-scale planetary and lunar images; (2) the extracted crater edges are clear and complete and do not merge with the surrounding environment edge; and (3) it avoids the problem of parameter sensitivity that is present in a traditional CDA algorithm. The proposed method was verified using an image taken by the Chang’e-2 lunar probe, and a comparison with the traditional method based on morphology and adaptive Canny edge detection shows that the number of craters detected increases by more than 35%, while the computational efficiency is improved by more than 40%.

show abstract

Section: Simulation Verification On Sandmentioning

confidence: 99%