| Summary: | 碩士 === 義守大學 === 電機工程學系 === 104 === The main purpose of this thesis is to construct an autopilot system for an unmanned railcar based on computer vision technology. The system operation flow is that wireless camera on board the railcar to acquire continuous image frames and transfer them through wireless microwave receiver to the host computer. The main program written in Python keeps detecting several particular signs with meanings of acceleration, deceleration, reverse and stop in the receiving images. When a particular sign is detected and recognized, the main program gives motion command through Arduino UNO R3 board, which controls Arduino L298n board to generate PWM signals to railcar’s driving motor so that the railcar can properly response. In addition, the man machine interface that takes user command inputs and displays camera image is also provided. The particular signs’ identification program is developed based on Haar-like feature training and AdaBoost classifier from OpenCV.
To achieve the objective, comprehensive system integrations such as hardware, device drives, protocols, application program and man machine interface have been properly done. The experiment results successfully verify the proposed methodology and integrated system. In average, the railcar can response within one second when a particular sign is detected. Therefore, the real-time performance of the system is also assured. Since the pictures of particular signs can be easily reproduced at low cost, people can properly distribute those particular signs along the sides of the rails or hang those signs atop the rails to build up an autopilot railcar system with applications in mass rapid transit (MRT) and production line automation.
|
|---|
