Multi-view stereo vision can perceive three-dimensional information of three-dimensional scenes under various conditions, and pose detection technology based on multi-view stereo vision plays an important role in more and more fields with its unique advantages.
In the aerospace field, as more and more man-made aircraft are continuously launched into space, a large number of space junk and abandoned satellites are produced. If effective measures are not taken, the probability of effective flying objects being hit will become higher and higher. In the on-orbit missions, such as cleaning up the space junk and repairing the fault of the target object, it is necessary to ensure the relative position between the end of the manipulator and the target satellite, and an effective posture measurement method is particularly important. Dr. Jiang Cuicui, from the Institute of International Interdisciplinary Science, Beijing University of Aeronautics and Astronautics, studied this problem and made two improvements in the traditional visual positioning method.
One is that in terms of accuracy, the parameters on which the traditional visual positioning algorithm depends have certain errors, and these errors will be amplified in the calculation process. The improved algorithm selects five independent parameters to transform the spatial coordinates, and reduces the number of matrix transformations in the operation, effectively reduces the calculation error and improves the positioning accuracy. On the other hand, in terms of operation speed, the traditional visual positioning algorithm is very complex and time-consuming in the calculation of coordinate transformation. After the algorithm is improved, the complexity of the calculation process is reduced, and the parallel calculation method based on FPGA is used, which greatly reduces the time required for positioning calculation.
After improving the algorithm, it is necessary to verify the effectiveness of the improved algorithm. For this reason, NOKOV optical 3D motion capture system was selected in the laboratory. NOKOV measures the positioning accuracy of motion capture to sub-millimeter, and uses FPGA technology to calculate the edge, and the output data delay is controlled at millisecond level.
A Marker is pasted on the experimental object-the target satellite model and the mechanical arm, and three or more motion capture cameras can see the Marker and calculate its three-dimensional coordinates, thus calculating the relative positional relationship between the satellite model and the robotic arm. Taking the relative position obtained by the motion capture system as the true value can effectively verify the accuracy of the visual positioning algorithm.
Contact Us
-
We're always happy to answer your questions and provide more information.
Tell us about your issue and we'll guide you to the best solution.
- Application *
-
- Capture Volume * m m m
- Objectives to be Tracked *
- Number of Objectives (optional)
- Camera Type (optional)
- Camera Count (optional)
- Submit