| Summary: | 碩士 === 元智大學 === 光電工程學系 === 100 === In this research, using Minimum average correlation energy (MACE) method and shifted to recognize the polychromatic images. In our past research, training images are disposed the center. However, the total sidelobe energy may not be the minimum. In order to improve it, we decomposed the polychromatic images into 3 R, G and B components, and then rotated from 0° to 360° in steps of 5°, shifted from -5 pixel to 5 pixel in both the vertical direction and the horizontal direction to generate many training images. Finally, by using the MACE method and cross correlation operation, the filter with minimum sidelobe energy on the output plane can be obtained, so as to increase the recognition ability.
|
|---|
