| Summary: | 碩士 === 國立成功大學 === 機械工程研究所 === 81 === The purpose of this thesis is to develop an automatic
inspection system of robot. Using image processing technique
and the method of fuzzy-neural networks, this system can detect
the major axis and the number of pins of IC chips. It can find
the symbols printed on the face. Furthermore, it can determine
the orientation and recognize the symbols from the individual
characters. The input device is a CCD camera, so it can take
the image of an IC chip at an arbitrary position or
orientation. Througu the analysis of the image, we can find the
number of pins and the major axis of the chip so that the robot
can frasp it properly. After rotating the image to align the
the screen axis, the vision system can detect the blocks of
labels and find the relationships and profile of the characters
to recognize the symbols. The fuzzy neural networks was used to
classify symbols and determine the orientation of IC chips from
the consensus of individual symbols. Finally, it can read the
recognized symbols and combine them into meaningful data. This
system work automatically, and it can apply to the inspection
of the engineering component in manufacture process to assure
the production quality.
|
|---|
