Traditional homomorphic filtering function is based on high pass filler, and images processed by the function are dark. The paper analyzed image filtering result by traditional low pass filter at cut-off frequency radius of 5, and got a conclusion that illumination component mainly distributes in low-frequency range of 0~5 radius. It introduced a design scheme of novel homomorphic filter, and proposed a processing method for underground image of uneven illumination based on homomorphic filtering theory. The method uses traditional low pass filter to restrain low-frequency components and traditional high pass filter to strengthen high-frequency components, divides range of low-frequency components of underground images into two sections, and uses different processing methods in the two low-frequency sections, namely restraining low-frequency components in the range of restraining radius of low-frequency and making low-frequency components invariant which are between restraining radius of low-frequency and strengthening radius of high-frequency, so as to avoid restraining low-frequency component overly. The testing result showed that the method solves uneven illumination problem as well as avoids weakening overall illumination of underground images overly, which has good effect.
DownLoad: