Development of a Vision System and Basic Drawing with NAO Robot
Sun, Xiao (2016)
Vaasan ammattikorkeakoulu
2016
All rights reserved
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2016061012662
https://urn.fi/URN:NBN:fi:amk-2016061012662
Tiivistelmä
This thesis introduces the development of the vision system and basic drawing of a NAO robot. In this thesis three kinds of detection systems are discussed in order to test the algorithm for the vision system. Each detection corresponds to a kind of drawing, but all of the drawings use the same method, which is introduced in this thesis in detail.
This thesis can be divided into two main sections, the vision module section and the motion module section. The vision module is used for detection. As mentioned above, three kinds of detection systems are included namely contours detection, character detection, and face detection. In the detection of contours and characters, the robot will recognize the contours and characters that it detects. The motion module is based on a mathematic model that is built by two 2D coordinate systems. The two 2D coordinate systems have the same x-axis, but the unit length in the physical significance are not necessarily the same.
The main carrier of this thesis is the fifth generation NAO robot developed with Python language under the Windows platform. OpenCV and Tessract are used as the libraries of image processing and character recognition. In the character recognition, a technique named optical character recognition (OCR) is also used. Matlab is used to do some mathematical calculations and generate functions.
In this application the NAO robot can detect shapes, characters, and faces successfully and can draw corresponding graphs. However, due to the unstable hardware, the drawing cannot be very accurate, and the robot has to learn how to draw classes of shapes. In further study, the robot is expected to calculate the equations of motion by itself.
This thesis can be divided into two main sections, the vision module section and the motion module section. The vision module is used for detection. As mentioned above, three kinds of detection systems are included namely contours detection, character detection, and face detection. In the detection of contours and characters, the robot will recognize the contours and characters that it detects. The motion module is based on a mathematic model that is built by two 2D coordinate systems. The two 2D coordinate systems have the same x-axis, but the unit length in the physical significance are not necessarily the same.
The main carrier of this thesis is the fifth generation NAO robot developed with Python language under the Windows platform. OpenCV and Tessract are used as the libraries of image processing and character recognition. In the character recognition, a technique named optical character recognition (OCR) is also used. Matlab is used to do some mathematical calculations and generate functions.
In this application the NAO robot can detect shapes, characters, and faces successfully and can draw corresponding graphs. However, due to the unstable hardware, the drawing cannot be very accurate, and the robot has to learn how to draw classes of shapes. In further study, the robot is expected to calculate the equations of motion by itself.