151 results about "Retinex algorithm" patented technology

An image not subjected to color correction, gamma correction, MTF correction and others is used as an image to be subjected to Retinex image processing. Thus, a RAW image prepared only by converting an electric signal of a CCD image sensor by an A / D converter into a digital form is used as an image to be subjected to the Retinex processing. Predetermined gamma correction is performed after the Retinex processing. Thereby, Retinex algorithm substantially modeling functions of a retina and a cortex of a living body exhibits its original performance. The resolution of the original image is lowered, and processing is performed on an image thus produced to produce a blurred image. Then, the resolution of the blurred image is restored to the same resolution as that of the original image, and the Retinex processing is performed. Thereby, the Retinex processing can be performed without increasing memory consumption and extremely lowering an operation speed.

The invention discloses a multi-scale geometric analysis super-resolution processing method of a video blurred image, belonging to the technical field of intelligent information processing. Single-frame blurred images or multi-frame blurred images are acquired by surveillance videos, the input blurred images are decomposed into low-frequency coefficients and high-frequency coefficients by NSCT, the blurred images are de-noised by an HMT model in the NSCT domain, edge details are enhanced by visual suppression networks, sub-band images with low-frequency coefficients and high-frequency coefficients are interpolated nonlinearly by a HyperBF neural network model, the processed NSCT decompression coefficients are reconstructed by NSCT, and the multi-scale Retinex algorithm is introduced to regulate the image contrast in accordance with human eye visual consciousness. The processing of multi-frame blurred images is based on an image fusion method in the MGT domain and a non-uniform interpolation method in the MGT domain. Without changing the hardware of traditional video surveillance imaging system, the method can effectively restrain common noise in video images and further improve the resolution and the definition of the blurred images.

The invention discloses an image enhancement method on the basis of an improved multi-scale Retinex theory. The method comprises the following steps of: carrying out nonlinear adjustment on the details of dark areas and the brightness of highlighted areas by virtue of a global brightness adjustment function; enhancing an image by virtue of a canonical gain compensation multi-scale Retinex algorithm; and according to the mean brightness value of a selected area, calculating the parameters of an S curve, adaptively adjusting the S curve, and carrying out the procedures of nonlinear mapping and the like on the enhanced image, thus finishing the enhancement on the image. The method disclosed by the invention solves the problems that when the conventional multi-scale Retinex theory method is used, a halo phenomenon is caused, the overall brightness of an enhanced high dynamic range image is insufficient, and the contrast ratio of the image is low. According to the invention, the S curve is self-adaptively adjusted according to the brightness of the central area of the image, and then the nonlinear mapping is performed on the image, so that the gradation of the image is stretched, and the contrast ratio of the image is improved; and the robustness of the algorithm on a complex night vision image is improved.

An image processing system for implementing a variant of Retinex algorithm based on local illumination correction is provided to improve an intensity value of each position of an input image. The image processing system provides the ability to distinguish details when compressing dynamic range of a given image. Specifically, first the illumination of the image is estimated by performing a weighted low pass filter (WLPF) operation, and the original image is separated into the illumination and the reflectance images. Next, the estimated illumination is modified by using a non-linear mapping function based on local illumination at each pixel location. Then, the modified illumination image is combined with the reflectance image. Finally, an enhanced image is obtained after a contrast correction procedure.

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.

The invention discloses an enhancement method for a low-illumination image. The enhancement method comprises the steps that the low-illumination image to be processed is acquired and then the low-illumination image to be processed is transferred to an HSV color space from an RGB color space, and a chroma component, a saturation component and a brightness component are acquired; the brightness component is decomposed into a reflection component and an irradiation component by adopting an alternating minimization method based on a Retinex algorithm; the irradiation component and the reflection component are respectively enhanced and then synthesized into an enhanced brightness component; the saturation component is self-adaptively adjusted and then an enhanced saturation component is obtained; the chroma component, the enhanced brightness component and the enhanced saturation component are synthesized into a new HSV image; and the obtained new HSV image is converted into an RGB image, and an enhanced image is obtained through white balance processing. Definition of the low-illumination image can be greatly enhanced, and details are enabled to be reproduced so that the method is high in applicability and high in robustness and can be widely applied to the field of image processing.

The invention discloses a backlight image enhancement and denoising method based on foreground-background separation. The backlight image enhancement and denoising method comprises the steps that a backlight image is divided into a foreground area and a background area by adopting an interactive cutout algorithm; the pixel points in the foreground area are enhanced by adopting an improved Retinex algorithm; equalization processing is performed on the pixel points in the background area by adopting a CLAHE algorithm; denoising is performed on the foreground area after enhancement and the background area after equalization processing by adopting a multi-scale NLM algorithm; and weighted fusion is performed on the foreground area and the background area after denoising so that an enhanced and denoised backlight image is obtained. Different enhancement and denoising methods are respectively adopted for the foreground area and the background area of the backlight image so that detail enhancement of the foreground area of the backlight image can be realized, the background area can be protected from being excessively enhanced, the denoising effect is great and accuracy is high and thus the backlight image enhancement and denoising method can be widely applied to the field of backlight image processing.

The invention discloses a wavelet domain Retinex image defogging method, so as to improve the Retinex algorithm defogging effects and improve the algorithm efficiency. The method comprises the following steps: the image is transformed into HSV color space; wavelet transformation is carried out on a brightness component V, and a low-frequency subband in which a fog component is and a high-frequency subband in which noise and edge information are are acquired; the improved single-scale Retinex algorithm is used for processing the low-frequency subband, and a wavelet threshold method is used for processing the high-frequency subband; wavelet inverse transformation is carried out, and a reconstructed brightness component is acquired; a saturation component S is adaptively adjusted according to changes of the brightness component V; and according to each HSV color space component after change, a clear image after defogging is reconstructed. The defogging algorithm designed by the invention can effectively improve degeneracy of a foggy image, definition of the image is improved, and the processing efficiency is enhanced obviously.

The invention provides a Retinex image enhancement method based on a Laplacian pyramid. The method comprises the following three processes: firstly, carrying out multi-scale convolution on an originalimage, carrying out convolution by adopting three 3 * 3 Gaussian convolution kernels, and obtaining R, G and B channel feature maps with different scales; secondly, estimating an incident component by adopting a Retinex algorithm and enhancing colors; enhancing details of reflective components; and finally, carrying out linear weighted fusion on the color enhanced image and the detail enhanced image to obtain an enhanced image. According to the method, more details of the image are acquired by adopting multi-scale convolution, image enhancement is carried out by utilizing an MSRCR algorithm,and Gamma correction is carried out on an incident component to realize color enhancement. Bilateral filtering denoising is carried out on the reflection component to reduce noise interference, and image detail information is extracted through a Laplacian pyramid; and linear weighted fusion is carried out on the color enhanced image and the detail enhanced image to realize image enhancement.

The invention relates to a dual display device (DD2) for displaying an input image (I). The dual display device comprises a first display (D1) and a second display (D2). The first display is arranged for modulating an image from the second display. The dual display device further comprises a processor (Pr2) which comprises an image splitter (Sp) which splits the input image into an illumination image (Ii) and a reflection image (Ir) according to a retinex algorithm. The reflection image is displayed on the first display and the illumination image is displayed on the second display. Due to the series arrangement of the two displays the input image I is substantially recreated. The illumination image typically is a spatially low-resolution image derived from the input image. A benefit when using the illumination image at the second display is that the smoothed light intensity values of the illumination image lead to a lower average light intensity and thus to a lower power consumption compared to the prior art solution. Additional benefits when using the retinex algorithm for splitting the images are that parallax errors in dual display devices are reduced and that an improved usage of the dynamic range of the dual display device is obtained.

The invention discloses an image enhancement algorithm based on multi-algorithm fusion of different color spaces. Firstly, R, G and B three color channels of the image are stretched automatically in RGB color space; the original image is transformed from RGB color space to HSV color space, and the brightness component V is enhanced by using improved multi-scale Retinex algorithm, and the obtainedimage is transformed from HSV color space to RGB color space; then, two images in RGB color space are fused by weighting; finally, the fused image is Gamma corrected to realize color restoration and compensation. Experimental results show that the proposed algorithm improves the brightness and contrast of the image and retains more detailed information than other algorithms, which can effectivelyprevent false halo phenomenon.

The invention discloses an image enhancement method. The image enhancement method comprises steps that binary processing on original images is carried out, and the original images are segmented into background type images and target type images; according to the image types after segmentation, calculation processing on the images after segmentation is carried out through employing a corresponding scale retinex algorithm; double-side filtering processing on brightness images contained in the images after calculation processing is carried out, wavelet de-noising processing on reflection images contained in the images after calculation processing is carried out; the images after double-side filtering processing and the images after wavelet de-noising processing are merged, gamma correction is further carried out, and the images after enhancement are acquired. Two-scale transform is carried out, and reduction of convolutional calculation amount is facilitated; double-side filtering processing on the brightness images is carried out, wavelet de-noising processing on the reflection images is carried out, noise removal for the images is facilitated, and image detail loss is relatively small; gamma correction for the images after synthesis is carried out, and global image enhancement is facilitated.

The invention provides an image enhancement method and an image enhancement device. The method comprises the following steps: (1), performing multiscale Retinex algorithm treatment on a gray value of a pixel in a current image, and then obtaining RY(x, y) after EXP conversion; (2), revising RY (x, y) by a gain compensation method to obtain R'M (x, y), wherein the average value of the R'M (x, y) of the pixel in a selected area is Im; and (3) mapping the R'M (x, y) by a non-linear S-curve transfer function, wherein the coefficients (a and b) in the S-curve transfer function are parameters changing with the Im, a represents the increment speed of an S-curve, and b represents the position of the S-curve. The image enhancement method and the image enhancement device can not only enhance the contrast ratio of the overall brightness of the image, but also reduce the degree of the overall brightness toward to the average value easily resulted from algorithms like SSR (Single-Scale Retinex) and MSR (Multi-Scale Retinex) and the like, so that the contrast ratio of the local detail information is enhanced.

The invention discloses a high-quality quick tone mapping method and system, and the method comprises the steps: firstly carrying out the color space conversion, and converting an HDR image into an HSV color space from an RGB color space; secondly converting a brightness component of the image into a log-domain, and adaptively carrying out the global dynamic range compression through an arctan function with the sigmoidal function characteristics; thirdly improving a single-scale Retinex algorithm and carrying out the adaptive local detail adjustment of a global compression result through the improved algorithm; finally carrying out the adjustment of a saturation component, carrying out the color space conversion and obtaining a final mapping result. The method provided by the invention canachieve the high-quality display output of a high-dynamic range image on low-dynamic range display equipment, maintains a good overall effect, fully maintains the local details of the image, and enables the image to be more vivid and natural. Meanwhile, the method is high in calculation speed, is very high in efficiency, and can be used for a real-time system.

The invention discloses a mine image enhancement method based on bilateral filtering and a multi-scale Retinex algorithm. The method comprises the steps that 1 image wavelet decomposition is carried out to acquire high frequency and low frequency coefficients of an image; 2 the multi-scale Retinex algorithm and bilateral filtering are carried out on the low frequency coefficient of the image; 3 a soft threshold filtering algorithm is carried out on the high frequency coefficient of the image; 4 a discrete wavelet inverse transform formula is used to acquire an enhanced spatial domain image; and 5 local self-adaptive contrast enhancing is carried out on the enhanced spatial domain image. Compared with the existing enhancement method, the enhancement method provided by the invention has the advantage that based on wavelet transformation, the multi-scale Retinex algorithm and bilateral filtering are combined for processing; self-adaptive contrast enhancing is finally carried out; a processed picture has the advantages of rich detail, being structured and easy identification; and the shortcomings of halo, blurry detail and poor contrast of a result processed by a traditional scheme are overcome.

The invention relates to a low-contrast image enhancement method based on YCrCb. The low-contrast image enhancement method comprises the following steps: (1) performing image de-noising on an obtainedoriginal image by utilizing a trapezoidal low-pass filter; (2) performing conversion between a RGB color space and a YCrCb color space on the de-noised image to obtain an image brightness component;(3) performing piecewise linear transformation on a brightness component image by adopting a traditional spatial domain linear transformation method; (4) performing Gamma correction on the transformedimage; (5) performing Gaussian filtering on the corrected image by utilizing a single-scale Retinex algorithm and estimating illumination component and reflection component images; (6) performing image sharpening by utilizing a Gaussian sharpening algorithm aiming at the reflection component image; and (7) performing conversion between the YCrCb color space and the RGB color space on the sharpened image.

The invention discloses a color image enhancement method based on improved multi-scale Retinex. Firstly an unprocessed color image is read, a conventional multi-scale Retinex algorithm is improved, and the average brightness component of a current pixel is calculated by adopting bilateral filtering so that brightness component estimation based on the bilateral filtering method is obtained; secondly local contrast enhancement is performed according to the size relation of brightness of each pixel point and domain average brightness thereof; and finally the brightness image after enhancement is compared with the brightness component of the original image, RGB information of the recovered and enhanced image is linearly adjusted, and color image enhancement is obtained finally. The color image enhancement method has advantages of image detail recovery, elimination of the "halo" phenomenon, meeting the human eye visual effect and short operation time.

The invention relates to an image defogging method. To solve the problems that city streetscape / traffic monitoring images shot in foggy days are not clear, target objects such as vehicles and pedestrians in the images are not easy to recognize and the like, foggy-day images are integrally enhanced, and then high-frequency information of the images is extracted to strengthen details to finally acquire clear defogged images. The method includes collecting the foggy-day images, performing overall image enhancement on the foggy-day images by the Retinex algorithm, adjusting the details of the images and integrating the images. Tests prove that defogging effect of the Retinex<+> wavelet transform method is better than that of the Retinex algorithm and the Dark Channel algorithm. According to the image defogging method, auxiliary information such as depth information is not needed, and complex methods such as 3D (three-dimensional) modeling are not needed, so that defogging can be realized simply and quickly, and good real-time performance is realized.

The invention discloses an embedded foggy-weather real-time video image stabilization method. By using the method, problems of jitter and contrast ratio reduction, which exist in a foggy-weather video, are solved. The method comprises the following steps of A, simultaneously applying video image stabilization based on an Harris angular point and an Retinex algorithm with color recovery in an embedded system so as to carry out foggy-weather video processing, carrying out image stabilization processing of a YUV space based on the Harris angular point and defogging processing based on a multiscale Retinex algorithm with colors on an image collected in real time, and recovering a clear image without jitter; and B, realizing effective combination of defogging on the YUV space and the image stabilization and optimizing the algorithm based on an embedded system characteristic. By using the method, the jitter of the shooting video can be effectively eliminated, clear scene information can be obtained and a visual effect of the image is improved; and simultaneously, a real-time operation speed can be reached so that a good application value is provided for a real-time monitoring system.

The invention discloses a scanning line processing based automatic reading method for a pointer instrument. The method comprises the steps of firstly performing illumination processing on an original image by adopting a single-scale Retinex algorithm, binarizing the image, then extracting characteristic pixel points by a row scanning line processing based method, performing Hough conversion on the characteristic pixel points to detect a straight line of a pointer, and finally calculating a read number by adopting an angle method, wherein the row scanning line processing based method refers to a method for scanning each row in the image, extracting continuous segments in the row, searching for the continuous segments with the lengths smaller than or equal to the line width of the pointer and taking the line segment midpoints of the continuous segments as the characteristic pixel points. According to the method, due to the adoption of the Retinex algorithm, the influence of non-uniform illumination on subsequent algorithms can be reduced; and due to the adoption of the row scanning line processing based method, the number of pixel points participating in Hough conversion can be reduced, thereby reducing the calculation amount.

The invention provides an underwater colored image restoration method based on color correction and a red channel priori. A Retinex algorithm is used to eliminate non-uniform underwater illumination; a Lambert diffuse reflection model is combined with a gray world algorithm to estimate the color of a light source, light is introduced to improve an underwater attenuation imaging model, and the color of the light source is eliminated to realize color correction; and a prior method of a red color is used to restore a color corrected underwater image. Thus, the method provided by the invention can be used to restore the underwater colored image effectively, and eliminate non-uniform illumination and color distortion.

The invention discloses a haze weather image enhancement method. The haze weather image enhancement method comprises steps that an original haze weather color image is converted from RGB space to HSV space; filtering processing of brightness V is carried out, and color tone H, saturation S, and the brightness V are synthesized together to acquire an image 1; the image 1 is converted from the HSV space to the RGB space to acquire an RGB space image 2; the image 2 is processed by using a multi-dimensional Retinex algorithm, and the haze weather color image is divided into an incident component L and a reflection component R; the Gamma conversion of the incident component L is carried out, and the double-sided filtering of the reflection component R is carried out, and then an image 3 is acquired; the enhancement processing of the image 3 is carried out by using a Sigmoid function, and a final image 4 is acquired.

The invention relates to an image mist elimination device and a method, in particular to a real-time image enhancing device and a method aiming at image mist and haze elimination under a lower contrast ration condition of a mist condition or a haze condition. The real-time image enhancing device comprises a decoder, a field programmable gate array (FPGA), a storage and an encoder, wherein the decoder is used for decoding video signals and converting the video signals into an image sequence, the FPGA is used for carrying out an enhancement process of a fast Retinex algorithm on the image sequence, the storage is used for storing the image sequence, and the encoder is used for converting the image sequence on which the enhancing process is carried out into the video signals. According to the real-time image enhancing device and the method, large calculating resources and large data storage capacity are not needed, effects of mist and haze elimination can be met, video image details in mist and haze days are enhanced, image visual quality is improved, the requirement of real-time process of hardware can be met at the same time, initial delay is smaller than 40 milliseconds, and original video frame rate is not affected.

The invention discloses a pedestrian re-recognition method based on a retinex algorithm and a convolutional neural network. According to the method, a video frame sequence in a video database is extracted; the convolutional neural network is constructed, and a pedestrian network model is obtained through training; the trained network model is used to find out pedestrians from the video frame sequence; the retinex algorithm is used to perform image enhancement on the pedestrians; the enhanced pedestrians are inputted into the convolutional neural network, and the depth characteristics of the pedestrians at different levels are extracted; and classification is performed through the softmax classifier of the last layer of the convolutional neural network, so that a final matching similarity is obtained. Problems such as illumination change and shadow coverage in a real scene are fully considered; before recognition is performed, the retinex enhancement algorithm is introduced to simulate a human visual system, so that an image can be closer to what human eyes see, and therefore, a recognition effect can be effectively improved; and an end-to-end pedestrian re-recognition method is adopted, pedestrian detection and pedestrian recognition are combined through using the same convolutional neural network, and therefore, the alignment problem of pedestrian labels can be solved.

The invention discloses a foggy weather image enhancement method based on fractional differential and dark channel prior. The method herein includes the following steps: 1. inputting a foggy weather image I, conducting dark channel prior and Retinex algorithm processing on the I, obtaining an initial de-foggy image J (x,y); 2. segmenting the J (x,y) to a foreground image J1(x,y) and a background J2(x,y); 3. separately computing the optimal fractional differential order number v1 corresponding to J1(x,y) and the optimal fractional differential order value v2 corresponding to J2(x,y); 4. determining a mask coefficient and a mask size, constructing a fractional differential operator mask w (s,t); 5. separately introducing the fractional differential order number v1 and fractional differential order value v2 which are obtained from 3 to the w (s,t), obtaining w1(s,t) and w2(s,t), conducting convolution operation on the pixel points of w1(s,t) and J1(x,y), and conducting convolution operation on the pixel points of w2(s,t) and J2(x,y); and 6. outputting the image of I which after the image enhancement. According to the invention, the method herein addresses poor de-foggy effects which are often caused in the process of enhancing foggy images by using fractional differential algorithm which has single fractional differential order in the prior art.

The invention discloses a face recognition method. According to the method, a shot face image is input to a computer. First, the degree to which the face image is influenced by illumination is calculated by use of a Retinex algorithm, and the lightness value of the face image is adjusted according to the degree of influence; second, face feature attributes are extracted by use of an active appearance model; third, the face feature attributes are reduced by adopting a self-organizing neural network and a rough set reduction algorithm, and a support vector machine classifier is trained with the use of the reduced attribute data; and finally, the to-be-recognized face image is analyzed by use of the trained support vector machine classifier. The method eliminates subjective interference, and is high in rate of face image recognition.

The invention provides a system and a method for automatically enhancing underwater images in NSCT (non-subsampled contourlet transform) regions on the basis of human visual characteristics. The system and the method have the advantages that multi-scale and multi-direction edge expression merits and noise removal merits of NSCT are comprehensively utilized, perceptual characteristics of human visual systems are used for reference, accordingly, noise in the NSCT regions can be removed, uneven underwater illumination can be eliminated by the aid of introduced Retinex algorithms, and NSCT contrast mask coefficients on the basis of the human visual characteristics can be generated by the aid of introduced brightness mask and introduced contrast mask characteristics; the NSCT contrast mask coefficients and NSCT low-frequency sub-band coefficients can be respectively automatically regulated by two designed nonlinear mapping functions, accordingly, contrast can be enhanced, and parameters do not need to be manually regulated; the noise can be effectively removed, the contrast of the underwater images can be automatically enhanced, edges of the underwater images can be sharpened, and uneven illumination can be eliminated.

The invention relates to a weak light image enhancing method based on Retinex and Reinhard color migration. The method comprises the following steps of (1) reading optimal scale parameters of a gaussian function, (2) enhancing a weak light image through a single-scale Retinex algorithm, (3) calculating color information of the enhanced image through a Reinhard color migration algorithm, and meanwhile reading color information data of a reference image, (4) migrating the color information of the reference image to the enhanced image, and (5) obtaining an image with the rich colors and outstanding details. Compared with the prior art, the method has the advantages of being good in image enhancement effect, enriching image color information, good in real-time performance and the like.