A novel image enhancement method for uncooled infrared thermal imaging system

In the uncooled infrared imaging system by the optical readout method with the micro-cantilever, the images obtained have obvious noise. The noise is generated due to the imaging device technology and stray light effects. The noise has a bad affect for the image quality, and even the continuous noise pixels form holes, which make the target difficult to be identified. Therefore, the novel infrared image enhancement method based on the mask convolution is presented. In the method, the information of convolution mask like size can be obtained by Hough transform, and the convolution mask can be built based on the characteristics of optical readout method in uncooled infrared imaging system. The position of mask on the image is changed pixel by pixel, and convolution is made between mask and pixels in the image. Then, an evaluation parameter threshold is established for eliminating the noise of the image. In the image after denoising, a mean filter is made to fill noise pixels and unit intervals of micro-cantilever. Finally, the Enhanced infrared image can be gained.