158results about How to "Good fog removal effect" patented technology

The invention discloses an image defogging method and a system based on a deep learning neural network. The method comprises the following steps of inputting an image with fog into a deep learning neural network system; using the deep learning neural network system to carry out characteristic extraction on the image with fog, and carrying out autonomous learning and extracting a fog correlation characteristic; carrying out multiscale mapping on the image with fog, extracting the characteristic of the image with fog in a concentrative mode under different scales and forming a characteristic graph; carrying out local extremum on each pixel in the characteristic graph, maintaining a resolution to be unchanged and acquiring the processed image; carrying out nonlinear regression operation on the processed image and acquiring initial transmissivity t(x); using a guided filter to optimize the transmissivity and carrying out image smoothing processing on the processed image; calculating an atmospheric light parameter; and according to the initial transmissivity t(x) and the atmospheric light parameter, recovering a fogless image. In the invention, connection is established between the system and an existing defogging method, and under the condition that efficiency and easy implementation are guaranteed, compared with the existing method, the method has better defogging performance.

The invention discloses a haze image rapid haze removal method. Dark channel rapid computation comprises the steps of establishing a cross-shaped dark channel formwork; computing pixel values of a transverse dark channel window and a longitudinal dark channel window; and comparing the pixel values and assigning a minimum value to an image dark channel. A mean value of a maximum gray value of 0.1% of a pixel in the dark channel is selected and serves as an atmosphere light value; a transmission image is obtained according to a dark channel image; and the transmission image is refined through a rapid Gaussian filter. Due to the fact that the cross-shaped formwork is adopted, computing area and computing amount are reduced; repeated computation is avoided and computing time is saved; the atmosphere light value is computed and obtained in the dark channel in the process of estimation and computing speed is accelerated under the premise of guaranteeing haze removal effect; and a transmission rate is optimized through a recursion Gaussian filter method and an optimized result is good.

The invention provides an image defogging method based on the concentration feature of fog, comprising the following steps: calculating the fog concentration feature value of each pixel in a foggy image; segmenting the foggy image to get a sub scene set through an image segmentation method based on the fog concentration feature values; screening out a sky-like area from the sub scene set; selecting the first 1% pixels with lowest saturation component in the sky-like area to form a candidate pixel set, selecting the first 10% pixels with highest saturation component in the candidate pixel set to form an atmospheric light area, and calculating the average intensity value of all the pixels in the atmospheric light area as a global atmospheric light value; calculating the transmittance of each pixel in the foggy image; and getting a defogged image according to the global atmospheric light value and the transmittance. The atmospheric light area can be located accurately in the defogging process. The method is less susceptible to highlighted noise points or interference in the foggy image. Therefore, an accurate global atmospheric light value is obtained, and a better defogging effect is achieved. The image defogging method is used to defog a variety of foggy images, and is of high robustness.

The invention discloses an image defogging method based on dark channel prior. The image defogging method based on dark channel prior comprises steps of 1) reading an original fogged image I(x), 2 ) performing an operation of taking a minimum value on two sides of the original fogged image I (x), using the dark channel prior knowledge to initially estimate an atmosphere light value and a medium transmissivity value through a grey scale opening operation, performing white balance processing on the atmosphere illumination so as to obtained a simplified image I'(x) of the original fogged image, 3) introducing a threshold T to obtain a mixed dark channel I<~><dark> of the simplified image I'(x), 4) using a sectional type mapping equation to perform mapping processing on the mixed dark channel I<~><dark> to obtain a rough estimation value t''(x) of the medium transmissivity and 5) obtaining an accurate estimation value t''' (x) of the medium transmissivity through a guided filtering method,, 6) using an adaptive parameter regulation method to perform revising on the medium transmissivity of a sky area to obtain a medium transmissivity t(x)m and 7) calculating a defogged image according to a final estimation value of the medium transmissivity t(x).

The invention discloses a real-time video de-fogging method using gray-scale image guided filtering. The method comprises the following steps that: an image is obtained in a video stream; the original image is converted into a gray-scale image; guided filtering is performed on the gray-scale image; downsampling is performed on the original image, so that a thumbnail is obtained; the dark primary color image of the thumbnail is obtained according to a dark primary color definition formula; a quadtree method is adopted to find a dense fog area or a sky area; and an improved atmospheric light scattering model is used to restore an image, and a foggy day video is outputted. With the method adopted, the defogging effect of the restored image is good, and the colors of the restored image are bright and vivid; problems of white edge phenomena occurring in places where depth of field changes greatly, and color distortion in a sky area which is caused by a dark primary color theory-based de-fogging algorithm can be solved. The method has a very good de-fogging effect for light fog, moderate fog and dense fog; the computation amount of the method is small; and the processing speed of the method is high and stable.

The invention discloses an image defogging method and method based on image partial content characteristics. The method comprises the following steps: dividing an initial image into a plurality of small block regions; calculating the image of each small block region to obtain a grey level histogram; calculating an information entropy the image of each small block region; calculating a self-adapted cutting parameter of the image of each small block region; cutting the information entropy the image of each small block region by using the self-adapted cutting parameter; dividing the initial image into a plurality of small block regions again; and carrying out histogram equalization processing on each small block region according to the cut grey level histogram. By the aid of the image defogging method and, the initial image can be effectively defogged; meanwhile, noises are not enhanced in a processing process and the image color is not distorted.

The invention discloses an adaptive global dark channel prior dehazing method for a bright area. The method comprises the following steps; 1, a global dark channel value of a haze image I(x) is solved; 2, a fuzzy logic controller is built to obtain a tolerance parameter and a transmittance adjustment factor; 3, the tolerance parameter is used for dividing the haze image into a bright area and a non-bright area; 4, the atmospheric light intensity A is solved; 5; the transmittance is solved; and 6, the tolerance parameter, the transmittance and the atmospheric light intensity are used to obtain a restored image. Problems of block effects, Halo phenomena, color distortion and the like caused by bright area distortion and block processing can be effectively solved, good dehazing effects can be acquired in a condition of not increasing exposure treatment, objective evaluation results for three aspects of the image contrast ratio, the information entropy and the average gradient are obviously better than those of other comparison algorithms, and the operation efficiency is greatly enhanced.

The invention discloses a fast surveillance video haze removal method, aiming at addressing high computing complexity and fuzzy details of haze removed images in the prior art. The technical solution includes the following steps: upon computing an atmospheric light, halving a haze image into an upper half and a lower half, obtaining the atmospheric light from the upper half of the haze image by using dark channel prior; in the process of computing the transmittance of a single frame haze image, estimating the transmittance by using the improved fast transmittance estimation method which introduces adaptive parameters; in the process of processing a video, dividing the video into a background and foreground varying part by using the adjacent frame difference method, only computing the transmittance of the foreground varying part, and making the transmittance computing result at the same position of a previous frame as the transmittance of the background part of the current frame. According to the invention, the method herein not only reduces the complexity in computing transmittance, and increases the clarity of details of the haze removal image. The method can be applied to surveillance situations which have complex moving targets and large moving displacement.