Infrared image detail enhancement method and device

Abstract

The invention discloses an infrared image detail enhancement method, which comprises the following steps that: adopting a platform histogram equalization algorithm to carry out compression processing on an original infrared image I to obtain image tone data IB; according to the original infrared image I, designing a detail preserving filter for the original infrared image I, and utilizing the detail preserving filter to carry out detail preserving filtering on the original infrared image I to obtain image detail data ID; calculating the amplification coefficient of the image detail data ID, processing the image detail data ID according to the amplification coefficient and a preset detail gain factor G to obtain final detail data I'D, wherein the detail gain factor G is used for controlling image detail proportion weight; and according to a formula IF=IB+G*I'D, carrying out fusion on the image tone data IB and the final detail data I'D to obtain a combined infrared image IF, and processing a non-positive and overflow numerical part in the combined infrared image IF to obtain an output infrared image I'F. By use of the method, the global comparison of the infrared image is kept, local details are enhanced, and an image display effect is improved. In addition, an algorithm is low in complexity and easy in implementation in real time.

Description

technical field [0001] The invention belongs to the technical field of infrared image processing, and more particularly relates to an infrared image detail enhancement method and device. Background technique [0002] The dynamic range of the output data of mainstream infrared thermal imaging and temperature measurement equipment is 14bits bit width, while the dynamic range of display equipment and human eye recognition ability is generally 8bits. In order to facilitate the display and observation of infrared image details, it is necessary to compress the dynamic range of the original output data under the premise of retaining global and local details. [0003] The traditional histogram equalization algorithm is to make the pixels with high probability density in the image occupy more gray levels, and vice versa, allocate less gray levels, so as to effectively use the 8bits gray level and improve the contrast of infrared images. However, this algorithm will lead to problems ...

AI Technical Summary

Problems solved by technology

However, this algorithm will cause problems such as over-amplification of background noise, over-enhancement of images, and whitening of images.

[0004] The platform histogram equalization algorithm introduces the concept of platform value. By calculating the appropriate platform value, the pixels with too large probability density in the histogram are restricted, which solves the problem of image over-enhancement in the traditional histogram equalization algorithm, and the global display effect of the image is obtained. Effective improvement, but for scenes with a wide temperature range, the algorithm does not significantly enhance local details

However, the bilateral filter algorithm for local detail enhancement has high complexity and a large amount of calculation, which is not easy to realize in real time and has the problem of edge virtual image.

Images