研究目的
Investigating the navigation of an autonomous underwater robot (AUR) using video stream for local maneuvering and 3D reconstruction of environmental objects.
研究成果
The developed visual navigation method and 3D reconstruction technique for AURs showed effectiveness in increasing localization accuracy and reducing computational time. Future work includes continuation of natural experiments and improvement of program-algorithmic tools.
研究不足
The efficiency of the virtual network of coordinate referencing depends on the geometry of the AUV trajectory. The method requires further improvement in terms of navigation accuracy and computational costs.
1：Experimental Design and Method Selection:
The approach combines visual odometry with stereo images and a virtual network for coordinate referencing. An original algorithm for tracing features on images is proposed to increase localization accuracy.
2：Sample Selection and Data Sources:
Experiments were conducted on model scenes prepared using a simulator and with real stereo images from a camera (GC1380).
3：0). List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Stereo camera GC1380 with a shooting frequency of 20 fps, resolution of 1369x1024, and a basic distance of 30 cm.
4：Experimental Procedures and Operational Workflow:
The method involves forming a network of basis coordinate systems (BCS) for referencing, tracking features across images, and performing 3D reconstruction from sequences of images.
5：Data Analysis Methods:
The effectiveness of the navigation method and 3D reconstruction was evaluated through computational experiments with model and real data.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
