TY - GEN
T1 - TV Guidance Simulation Platform Based on Deep Learning
AU - Yu, Zhaowei
AU - Chen, Wanchun
AU - Chen, Zhongyuan
AU - Liu, Xiaoming
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/11
Y1 - 2019/11
N2 - Hardware-in-the-Loop TV Guidance System Simulation Platform (HIL-GSS) is an experimental platform designed by the School of Astronautics, Beihang University, which is used for the hardware-in-the-loop simulation of TV guided missile-target engagement. The platform has been applied to research and experiment for many years and highly praised. However, the platform currently has the following problems: 1) Fidelity of the fixed-size target is not high enough; 2) The target recognition algorithm is only based on color feature; 3) The contour recognition algorithm based on Hu moments detects the target with slow speed and poor accuracy. In order to solve above problems, we tried to improve the contour recognition algorithm firstly. Secondly, we used 3D model to build the scene of target simulation with Unity 3D. Thirdly, we applied YOLO, a deep-learning-based target recognition algorithm, to achieve accurate recognition of targets in different attitudes. In the end, several experiments were carried out by using YOLO and color-based target recognition algorithm respectively to verify the feasibility of using YOLO for TV guidance simulation experiment.
AB - Hardware-in-the-Loop TV Guidance System Simulation Platform (HIL-GSS) is an experimental platform designed by the School of Astronautics, Beihang University, which is used for the hardware-in-the-loop simulation of TV guided missile-target engagement. The platform has been applied to research and experiment for many years and highly praised. However, the platform currently has the following problems: 1) Fidelity of the fixed-size target is not high enough; 2) The target recognition algorithm is only based on color feature; 3) The contour recognition algorithm based on Hu moments detects the target with slow speed and poor accuracy. In order to solve above problems, we tried to improve the contour recognition algorithm firstly. Secondly, we used 3D model to build the scene of target simulation with Unity 3D. Thirdly, we applied YOLO, a deep-learning-based target recognition algorithm, to achieve accurate recognition of targets in different attitudes. In the end, several experiments were carried out by using YOLO and color-based target recognition algorithm respectively to verify the feasibility of using YOLO for TV guidance simulation experiment.
KW - Hu moments
KW - TV guidance
KW - Unity3D
KW - YOLO
KW - target simulation
UR - https://www.scopus.com/pages/publications/85085865096
U2 - 10.1109/CIS-RAM47153.2019.9095814
DO - 10.1109/CIS-RAM47153.2019.9095814
M3 - 会议稿件
AN - SCOPUS:85085865096
T3 - Proceedings of the IEEE 2019 9th International Conference on Cybernetics and Intelligent Systems and Robotics, Automation and Mechatronics, CIS and RAM 2019
SP - 89
EP - 94
BT - Proceedings of the IEEE 2019 9th International Conference on Cybernetics and Intelligent Systems and Robotics, Automation and Mechatronics, CIS and RAM 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 9th IEEE International Conference on Cybernetics and Intelligent Systems and Robotics, Automation and Mechatronics, CIS and RAM 2019
Y2 - 18 November 2019 through 20 November 2019
ER -