Chaohua Ma
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chaohua Ma has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Wuming Zhang
School of Geospatial Engineering and Science, Sun Yat-Sen University, Zhuhai, China
3 shared publications - Yuan Li
School of Geospatial Engineering and Science, Sun Yat-Sen University, Zhuhai, China
3 shared publications - Zhouxuan Xiao
School of Geospatial Engineering and Science, Sun Yat-Sen University, Zhuhai, China
3 shared publications A. Li
School of Geospatial Engineering and Science Sun Yat‐sen University Zhuhai P.R. China
1 shared publication- Jie Shao
Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China
1 shared publication Junying Lv
School of Geospatial Engineering and Science, Sun Yat-Sen University, Zhuhai, China
1 shared publication
Latest Publications
- 29 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-18
Automated Rock Detection From Mars Rover Image via Y-Shaped Dual-Task Network With Depth-Aware Spatial Attention Mechanism
Extracting rocks from Mars rover images using convolutional neural networks is a semantic segmentation problem gaining increasing attention in planetary science and artificial intelligence. However, this task still faces the challenges of inaccurate extraction of rock boundaries and small rocks. An important reason is that the textural features of the Mars rover images are not discriminative enough between the target and the background. In order to obtain 3-D information to enhance the rock extraction without introducing additional data, we designed a dual-task branch network with a Y-shaped encoder-decoder structure. In our network, the primary semantic segmentation task branch is used to decode textural features into semantic features and output rock extraction results; the auxiliary task depth estimation branch decodes textural features into depth features in 3-D, and transmits them to the primary branch through the spatial attention module to enhance the identification ability of semantic features for boundaries and small targets. In addition, another reason impedes accurate Mars rock extraction is the lack of high-quality annotated training datasets. Therefore, we created a dataset containing 6,325 images pairs with corresponding annotations and depth information, SimMars6K. Ablation and comparison experiments based on this dataset and two actual datasets show that our method achieves 83.1% IoU and 90.8 F 1 -score on simulated dataset, and outperforms other methods on actual datasets with 2% improvement in recall. Transfer learning experiment shows that the pre-training on our simulated data can bring up to additional 5% gain on mIoU and 6% gain on mPA for general model.
- 29 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-18
Automated Rock Detection From Mars Rover Image via Y-Shaped Dual-Task Network With Depth-Aware Spatial Attention Mechanism
Extracting rocks from Mars rover images using convolutional neural networks is a semantic segmentation problem gaining increasing attention in planetary science and artificial intelligence. However, this task still faces the challenges of inaccurate extraction of rock boundaries and small rocks. An important reason is that the textural features of the Mars rover images are not discriminative enough between the target and the background. In order to obtain 3-D information to enhance the rock extraction without introducing additional data, we designed a dual-task branch network with a Y-shaped encoder-decoder structure. In our network, the primary semantic segmentation task branch is used to decode textural features into semantic features and output rock extraction results; the auxiliary task depth estimation branch decodes textural features into depth features in 3-D, and transmits them to the primary branch through the spatial attention module to enhance the identification ability of semantic features for boundaries and small targets. In addition, another reason impedes accurate Mars rock extraction is the lack of high-quality annotated training datasets. Therefore, we created a dataset containing 6,325 images pairs with corresponding annotations and depth information, SimMars6K. Ablation and comparison experiments based on this dataset and two actual datasets show that our method achieves 83.1% IoU and 90.8 F 1 -score on simulated dataset, and outperforms other methods on actual datasets with 2% improvement in recall. Transfer learning experiment shows that the pre-training on our simulated data can bring up to additional 5% gain on mIoU and 6% gain on mPA for general model.
- 18 July 2023
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Planets
- Vol. 128 (7)
Extraction and Analysis of Three‐Dimensional Morphological Features of Centimeter‐Scale Rocks in Zhurong Landing Region
China's first Mars rover, Zhurong, has carried out an exploration mission in southern Utopian Planitia. As rocks and their three‐dimensional (3D) morphology are essential for studying the geological characteristics of the Martian surface, this paper presents an approach for automatically extracting 3D rocks from stereo rover images. With our approach, 6,185 cm‐scale rocks are extracted using the Zhurong NaTeCam images, with the F 1 ‐score achieving 91.86%. Approximate 3D morphological features are then calculated for each rock, and the statistical result shows that small rocks in centimeters make up the majority. Comparative morphological analyses of rocks in different areas show that landscape type affects the number of rocks but has no clear relationship to the proportions of rocks of different sizes. The outcrops of sedimentary rocks, the predominant type of rocks in the Zhurong landing area, and the float rocks are distributed differently regarding all morphological features, making the size‐frequency distribution of rocks around craters significantly different from other areas. Further comparative analysis with the Bonneville crater ejecta rocks extracted from the Spirit rover data is also conducted. Some similarities between the Zhurong outcrop rocks and the Bonneville crater ejecta rocks are found in 2D compactness . Meanwhile, the Bonneville crater ejecta rocks have a more scattered distribution in sphericity than both the float and outcrop rocks in the Zhurong landing area, indicating that the shapes of the Bonneville crater ejecta rocks may be more diverse. In addition, the sphericity is found to converge with height for rocks in both landing areas.
- 18 July 2023
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Planets
- Vol. 128 (7)
Extraction and Analysis of Three‐Dimensional Morphological Features of Centimeter‐Scale Rocks in Zhurong Landing Region
China's first Mars rover, Zhurong, has carried out an exploration mission in southern Utopian Planitia. As rocks and their three‐dimensional (3D) morphology are essential for studying the geological characteristics of the Martian surface, this paper presents an approach for automatically extracting 3D rocks from stereo rover images. With our approach, 6,185 cm‐scale rocks are extracted using the Zhurong NaTeCam images, with the F 1 ‐score achieving 91.86%. Approximate 3D morphological features are then calculated for each rock, and the statistical result shows that small rocks in centimeters make up the majority. Comparative morphological analyses of rocks in different areas show that landscape type affects the number of rocks but has no clear relationship to the proportions of rocks of different sizes. The outcrops of sedimentary rocks, the predominant type of rocks in the Zhurong landing area, and the float rocks are distributed differently regarding all morphological features, making the size‐frequency distribution of rocks around craters significantly different from other areas. Further comparative analysis with the Bonneville crater ejecta rocks extracted from the Spirit rover data is also conducted. Some similarities between the Zhurong outcrop rocks and the Bonneville crater ejecta rocks are found in 2D compactness . Meanwhile, the Bonneville crater ejecta rocks have a more scattered distribution in sphericity than both the float and outcrop rocks in the Zhurong landing area, indicating that the shapes of the Bonneville crater ejecta rocks may be more diverse. In addition, the sphericity is found to converge with height for rocks in both landing areas.
- 24 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
High-Precision Measurement of 3-D Rock Morphology on Mars Using Stereo Rover Imagery
Fine-scale 3-D morphological features of rocks on the Martian surface provide important clues to Mars exploration missions and scientific discoveries. To obtain such information, an automatic approach for high-precision measurement of 3-D morphological features of Martian rocks is proposed in this letter. The approach directly detects 3-D rocks from dense point cloud that is generated by interpolating triangulated irregular network (TIN) model, through a coarse-to-fine process combining cloth simulation filtering (CSF) and connected-component labeling algorithms. Multiple 3-D morphological features are then precisely extracted by modeling rock points and fitting local ground plane. Experimental results demonstrate that the proposed approach provides an effective way for the extraction of complete 3-D rocks and comprehensive morphological features with over 90% accuracy.
- 24 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
High-Precision Measurement of 3-D Rock Morphology on Mars Using Stereo Rover Imagery
Fine-scale 3-D morphological features of rocks on the Martian surface provide important clues to Mars exploration missions and scientific discoveries. To obtain such information, an automatic approach for high-precision measurement of 3-D morphological features of Martian rocks is proposed in this letter. The approach directly detects 3-D rocks from dense point cloud that is generated by interpolating triangulated irregular network (TIN) model, through a coarse-to-fine process combining cloth simulation filtering (CSF) and connected-component labeling algorithms. Multiple 3-D morphological features are then precisely extracted by modeling rock points and fitting local ground plane. Experimental results demonstrate that the proposed approach provides an effective way for the extraction of complete 3-D rocks and comprehensive morphological features with over 90% accuracy.