| dc.description.abstract | This master thesis features the implementation of a complete image processing, mapping and communication system on a Raspberry Pi for the characterization and mapping of a maze from the air. The system was implemented in Python and has been verified in real-life testing scenarios.
The system captures an image, detects the wall segments of the maze in the image and through the integration of sensors the system calculates the real life lengths and position of the walls. The system is designed to be run remotely, and incorporates a communication platform for the exchange of mapping data.
Image Processing and Mapping:
By translating an existing implementation of image processing and mapping algorithm from MATLAB and further developed in Python, and through the use of a new image processing library, OpenCV, the system is able to detect and extract information about wall position and length.
Sensor implementation:
A camera and a range sensor has been integrated in the hardware and software of the system and provides sensor data to the system for the image processing and mapping algorithm. The camera has been determined to give excellent sensor data, while the sensor data from the range sensor can be improved to obtain better accuracy.
Communication:
A communication system, together with a defined communication protocol has been implemented using SSH. Libraries that facilitate for a smooth, fast and secure connection between the system and a host computer has been installed and verified. | |
