Colorized night vision image brightness enhancement method applicable to automotive assisted driving system

Abstract

The invention provides a colorized night vision image brightness enhancement method applicable to an automotive assisted driving system. The method comprises the following steps of: firstly, transforming a night vision colorized image from a red, green and blue (RGB) space to a luma and chroma (YUV) space to overcome the shortcoming of color distortion caused by direct processing of the night vision colorized image in the RGB space; then processing a brightness component image by using an S-curve correction Retinex algorithm so as to enhance the detail and the brightness of the image; enhancing the brightness component image by a selective nonlinear grey level mapping method and keeping good shadow information; and finally, performing weighed fusion on the two enhanced images by a weighed fusion method and inversely transforming the weighed fusion brightness component image and a UV component image to the RGB space for displaying. The enhanced image acquired by the method keeps the necessary shadow information and has the detail and the brightness applicable to vision observation; and the night vision colorized image has a good enhancement effect.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to a night vision color image brightness adjustment method based on an S-curve correction Retinex algorithm, which is suitable for an automobile driving assistance system at night. Background technique [0002] Under the conditions of street lights and car lights, the biggest safety hazard of driving at night is that the driver cannot fully and accurately grasp the road conditions in time due to insufficient light. At present, driving safety is already an important evaluation index of automobiles, and it is extremely important to improve the safety of driving at night. Research on the enhancement technology of color night vision light image brightness can extend the driver's visual distance, improve the visual conditions of night driving, and provide the necessary technical means for the night assisted driving system of the car. [0003] The current patents for nighttime ima...

AI Technical Summary

Problems solved by technology

[0004] The current patents for color image enhancement include: A fast color image enhancement method based on Retinex theory (CN200810116385.2) discloses a color image that constructs a new color space, an average value template, and adjusts pixel value distribution by selecting adaptive parameters. Image enhancement method; a Retinex-based nonlinear color image enhancement method (CN201010578402.1) discloses a method of converting an image from RGB space to YCbCr space, and then using an improved Retinex illumination reflection model for local adaptive enhancement and reuse Gamma correction is a non-linear color image enhancement method for global brightness adjustment, but the above are all for color images in daylight. For night vision color images with insufficient brightness, the images obtained by these methods will cause visibility due to the lack of necessary brightness information. very poor

[0005] To sum up, the current image enhancement methods are not suitable for the brightness enhancement of color night vision images under lights, and cannot meet the requirements of the car's night-time assisted driving system in terms of enhancing the invisible details of the image, the color and brightness of the night vision image

Images