Thermal infrared image super-resolution reconstruction evaluation method

Abstract

The invention discloses a thermal infrared image super-resolution reconstruction evaluation method, which is used for evaluating the performance of a super-resolution reconstruction algorithm at a thermal infrared image application level, and comprises the following steps of: performing temperature inversion on a thermal infrared image at a first stage, and converting the thermal infrared image into a temperature image representing temperature; and utilizing the obtained temperature image for comparing the similarity between the temperature image of the thermal infrared image after super division and a temperature image of a real high-resolution thermal infrared image by calculating the Bhattacharyya distance of the temperature image at a second stage, and therefore evaluation can be givenon the application level of thermal infrared image temperature inversion. According to the thermal infrared image super-resolution reconstruction evaluation method, evaluation indexes of the thermalinfrared image super-resolution reconstruction algorithm are provided at the application level by utilizing the application characteristics of the thermal infrared image, so that the super-resolutionreconstruction algorithm evaluation and application are combined, the method has more practical significance, the method is not limited by the size of the image, and calculation can further be carriedout due to different sizes of temperature maps.

Description

technical field [0001] The invention relates to the technical field of digital image super-resolution reconstruction, and more specifically relates to a thermal infrared image super-resolution reconstruction evaluation method. Background technique [0002] At present, the image super-resolution reconstruction algorithm evaluates the performance of the algorithm through some image quality evaluation indicators. Commonly used reference objective evaluation indicators include peak signal-to-noise ratio (PSNR), structural similarity (SSIM), no reference evaluation index NIQE, subjective evaluation Index mean subjective score (MOS) and so on. [0003] However, the commonly used quality evaluation indicators for image super-resolution reconstruction are only evaluated from the image itself, without considering the application of the image. For different application scenarios, the selection of evaluation indicators should be inconsistent. For example, for scenes that do not pursue...

AI Technical Summary

Problems solved by technology

Similarly, thermal infrared images have their own application scenarios. Its main function is to invert the temperature of the shooting area. The quality evaluation indicators mentioned above are not specific to this application of thermal infrared images.

Images