Image real-time defogging method based on graphics processor

Abstract

The invention discloses an image real-time defogging method based on a graphics processor, mainly solving the problem that in the prior art, a computer CPU (Central Processing Unit) has a slow processing speed, and is not suitable for performing real-time defogging on high-definition images. The method comprises the steps of: 1, initializing a computer, inputting a fogging image and reading image data; 2, calculating a dark channel image for inputting the fogging image; 3, calculating a airglow value according to the dark channel image; 4, calculating the initial transmissivity image for inputting the fogging image; 5, guiding the initial transmissivity image to be filtered to obtain a final transmissivity image; and 6, restoring a photographic field according to the final transmissivity image, and outputting a clear and defogged image. The method can guarantee better defogging effects, has a fast processing speed, meets the real-time defogging requirement of a 10180p image, and can be used for high-definition image real-time defogging processing.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a real-time image dehazing method, which can be used in the fields of outdoor target recognition and tracking, road visual monitoring, satellite remote sensing monitoring and the like. Background technique [0002] The generation of fog map is mainly due to the scattering effect of turbid media such as fog and water vapor in the atmosphere, which seriously attenuates the light intensity received by the camera. Therefore, in order to ensure the clarity of the image and the identifiability of the information, and to correct the color shift of the image due to atmospheric light, the image dehazing algorithm is usually used to process the haze image. The image dehazing algorithm based on the dark channel prior physically dehazes the image, and the effect is obvious and natural, so it is widely used in the field of image dehazing. [0003] However, due to the high comp...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that the parallelism of the whole implementation scheme is not enough, some processes still use the CPU for serial processing, the GPU calculation part does not fully exploit the reusability of data, and does not use high-speed memory such as texture storage and shared storage to access data The mode is optimized so that it can only process low-resolution video images in real time. When the image resolution is slightly higher, it cannot achieve fast processing. For example, for a video image with a resolution of 720×576, its processing speed is only 21 frame / second

Obviously, it is difficult to achieve real-time processing for high-definition video images such as 1080p

Images