Vision plays an important role in our lives and wheeled robot. However, accurate and rapid image recognition is difficult to obtain. This study develops an intelligent recognition algorithm, including image capturing, image pre-processing, fuzzy inference and pattern recognition. In the proposed recognition scheme, the advantages of moment invariant can be maintained and the bottleneck can be improved by fuzzy inference. To verify the performance of the proposed architecture, the proposed algorithm is applied to the wheeled robots (FESTO Robotino) under Matlab and Labview software packages. The developing stage contains following three steps:
Steps 1: Utilizing Labview to start the camera of the controlled wheeled robots (FESTO Robotino) and to capture the image on fixed interval time.
Step 2: Utilizing Matlab to pre-process (binary) and post-process (pattern recognition) the captured image. The type and location of the image can be obtained via the calculation of moment invariant and the distance comparison with the built-in samples.
Step 3: Writing Matlab program to establish the fuzzy inference to calculate the turning angles of the controlled wheeled robot (FESTO Robotino) and save the angles into a document file. Then, the Labview’s interface program is started to execute to connect the computer and the controlled wheeled robot (FESTO Robotino) through wireless WiFi. Then, the captured turning angles are saved into another document file by Labview. Finally, the driven voltages of the motor are converted by Labview and sent to robot through WiFi.
In this thesis, the technologies of Matlab, fuzzy inference and moment invariant has been combined successfully to classify and identify an image. And also the Labview interface program has been successfully executed to capture the image, access the turning angles and convert driving voltages of the wheeled robot (FESTO Robotino). In the future, the performance can be further enhanced by embedding additional image pre-processing mechanisms to promote the recognition capability and to improve tracking accuracy of the robot in practical application.
