| Summary: | 碩士 === 國立臺灣師範大學 === 電機工程學系 === 104 === In recent years, automation technology is reaching into its maturity stage, and could be combined with various fields. Especially the use of LEDs is growing steadily within each year, which made our everyday life more convenient. In order to differentiate the quality of LEDs, LED probes are needed for measuring its characteristics. But to date, most probes are produced and tested artificially, which might lead to impreciseness of probes and are of variable quality. In view of this, we propose an image measuring technique using on LED probes with combination of computer vision and uniaxial stepper motor mobile platform. Two techniques are used in this thesis: image processing and automatic measuring, which are for the simulation of automatic measuring and analyzing of LED probes in factories. Finally, an automatic measuring platform for LED probes system is implemented.
Due to the complicated environment in factories, including noise, dust, etc., these factors will undoubtedly cause certain effect on measurement accuracy of probes. Thus, we first proposed the various segmentation averaging sub-pixels (VSAS) method, to reduce disturbance caused by potential environmental factors. Second, we proposed an efficient method to diminish reflection points on probe. Last, we extracted the base of probe by using ROI method. By using the above mentioned image processing methods, we are able to greatly improve the accuracy and stability of the automatic measuring system.
|
|---|
