Vehicle-mounted image processing device and automobile

Abstract

The invention provides a vehicle-mounted image processing device which specifically comprises at least one visible light camera and a micro-processing controller, the visible light camera is connected with the micro-processing controller, the micro-processing controller comprises a virtual exposure processing module used for image processing, the virtual exposure processing module comprises a virtual exposure processing process of an image. The process specifically comprises the steps of converting a collected image into a reversed image, obtaining the atmospheric illumination intensity of the image and the transmission rate of the image, generating a restored image according to a pre-configured image restoration model, and correcting and reversing the restored image to obtain a processed image. According to the technical scheme of the invention, by improving the calculation method of the atmospheric illumination intensity and the transmission rate in the prior art, the processed image is smoother and the imaging effect is better when the illumination is insufficient and the interference of a vehicle lamp, a street lamp or a relatively bright white object is caused.

Description

technical field [0001] The invention relates to the field of automobiles, in particular to an image processing device for vehicle-mounted images and an automobile. Background technique [0002] In the field of vehicle assisted driving, visible light cameras are generally used as imaging devices in front of the vehicle. However, most target detection technologies are based on daytime driving. When driving in daytime, the light is relatively sufficient and uniform, and the video data collected by the camera has less noise. Intelligent algorithms can be successfully applied to the scene to achieve efficient target detection. Provide drivers with assisted driving decisions. Visible light cameras often have poor imaging quality at night due to poor exposure conditions, especially when street lighting facilities are not perfect (such as on rural roads). In the case of poor light at night, the driver cannot see the surrounding environment clearly with the naked eye, and the visib...

AI Technical Summary

Problems solved by technology

Visible light cameras often have poor image quality at night due to poor exposure conditions, especially in situations where street lighting facilities are not perfect (such as on rural roads)

In the case of poor light at night, the driver cannot see the surrounding environment clearly with the naked eye. The visible light camera has poor imaging effect due to insufficient exposure, and cannot provide a better image effect for the driver.

Therefore, drivers often have traffic accidents because they cannot detect the pedestrians, vehicles, and road surface defects in front of them in time.

In order to better improve the imaging effect of the camera, the patent CN105991938A provides a virtual exposure method. In the virtual exposure method, in the process of calculating the image restoration model, when its t(x, y) is close to 0, it will directly cause the region The colors of the

In addition, in the application scenario of the car camera, in the case of insufficient light, it is necessary to turn on the car lights for road lighting. The light intensity is subject to the influence of car lights and street lights on both sides of the road. The calculation method of the atmospheric light intensity A is Take "cut the grayscale image into a preset number of grayscale image blocks according to the preset segmentation sequence, and the segmentation sequence includes from top to bottom and from left to right; when the grayscale image block is smaller than the preset When setting the size, obtain the mean value and variance of the brightness of each gray-scale image block; select the brightness of the pixel in the gray-scale image block with the largest mean value and the smallest variance as the atmospheric illumination intensity ", such The calculation method is feasible in the case of no light interference, but when the lighting is poor and there are many lights, using this method of using the maximum brightness of the pixel as the atmospheric light intensity will cause overestimation and improve the estimation of the atmospheric light intensity. value, making the final imaging effect poor

For example, the light intensity at a location with lights will be greater than that at a location without lights. When using this method to estimate, the intensity of the light location will be used to estimate the current atmospheric light intensity, so that the current estimated value deviates more from the real value.

Images