Image defogging adaptive parameter calculation method, and image defogging method and system
A technology of adaptive parameters and calculation methods, applied in the field of image processing, can solve the problems of limited real-time processing, slow running speed, and particularly large impact, and achieve the effect of increasing the edge-preserving filtering process, good contrast, and avoiding black borders.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0053] Such as figure 1 and figure 2 : An image defogging system, including a front-end module, a cloud defogging module and a back-end module. The front-end module, including the video data collection module and MPEG-4.10-based encoder, transmits the collected video data and sensor device style to the MPEG-4.10 encoder for compression, and the compressed code stream is sent to the cloud through network transmission for the next step processing.
[0054]The back-end module, including MPEG-4.10 code collection module and MPEG-4.10 decoder for video data, decodes the data transmitted through the network and plays it through the video player.
[0055] Such as figure 2 : the defogging module includes a dark channel processing module, a sky area processing module and a special defogging processing module;
[0056] The dark channel processing module is used to identify the dark channel image; the sky area processing module is used to identify the sky area image;
[0057] The ...
Embodiment 2
[0059] There is an existing input image I, the length of the image is m, the width is n, and the image bit width is Nbit.
[0060] An image defogging method, first of all, according to machine learning based on the Radviz algorithm, the gain circle of the sensor device with the distribution of defogging adaptive parameters is trained.
[0061] Obtain the device number through the image decoding data, call out the gain circle of the corresponding sensor device, recognize the image, and output the adaptive parameter k matched in the gain circle.
[0062] Then, through the dark channel prior algorithm, output the identified dark channel image, the calculation formula of the dark channel prior: Idark=min y∈s(x) (min c∈{r,g,b} I c (y)); c is an independent variable, the purpose is to select one of the 3 channels of RGB, and obtain the minimum value of each pixel compared with the 3 channels; Ic(y) is any color channel of image I; Idark is the dark channel data, s(x) is a sliding...
Embodiment 3
[0076] The Radviz algorithm is a concept of tension and spring, and the entire algorithm operates on a circle.
[0077] A method for calculating an adaptive parameter k for image defogging, comprising the following steps:
[0078] Step 1. Acquire three image data sets P={A, X, Ymean}, where A represents the global atmospheric light value, X represents the complexity of the image texture, and Ymean represents the average brightness of the image. Note: Radviz can support the existence of many data sets, but the optimal number is around 3-7.
[0079] Step 2, such as image 3 : The data set P is normalized, and P is substituted into the Radviz algorithm, and the position of the equilibrium point is obtained in the circle of the Radviz algorithm, and the adaptive parameter k is obtained.
[0080] Step 3, using several images with different gains, repeating the method of steps 1-2 to obtain several adaptive parameters k, and distribute the output values of the adaptive parameter...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More