CN115546047A - Video image noise reduction method, device and medium based on improved local filtering algorithm - Google Patents
Video image noise reduction method, device and medium based on improved local filtering algorithm Download PDFInfo
- Publication number
- CN115546047A CN115546047A CN202211070387.9A CN202211070387A CN115546047A CN 115546047 A CN115546047 A CN 115546047A CN 202211070387 A CN202211070387 A CN 202211070387A CN 115546047 A CN115546047 A CN 115546047A
- Authority
- CN
- China
- Prior art keywords
- pixel
- frame
- noise reduction
- current
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001914 filtration Methods 0.000 title claims abstract description 102
- 230000009467 reduction Effects 0.000 title claims abstract description 86
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 62
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000003068 static effect Effects 0.000 claims abstract description 20
- 238000001514 detection method Methods 0.000 claims abstract description 7
- 230000002146 bilateral effect Effects 0.000 claims description 23
- 230000008859 change Effects 0.000 claims description 23
- 238000013139 quantization Methods 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 6
- 238000009825 accumulation Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 230000002123 temporal effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 5
- 238000004364 calculation method Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 241000218645 Cedrus Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Picture Signal Circuits (AREA)
Abstract
The invention relates to a video image noise reduction method, a device and a medium based on an improved local filtering algorithm, wherein the method comprises the steps of judging whether an obtained current video frame is a first frame or not, and if so, carrying out spatial domain noise reduction on the current video frame based on the improved local filtering algorithm; if the current video frame is not the first frame, performing motion region detection on the current video frame according to the previous video noise reduction frame to obtain the motion condition of each pixel point; judging whether the current pixel is static or not according to the motion condition of the pixel, and if the current pixel is static, performing time domain noise reduction based on an improved local filtering algorithm on the current pixel; and if the current pixel point is not static, denoising the motion area, and storing the filtered current video frame. The invention utilizes the pixel correlation of the image in time domain and space domain and the randomness of noise to reduce noise, and ensures that the original details and textures of the image are not damaged and blurred while the image quality is improved. The invention can reduce the operation amount and improve the image noise reduction effect.
Description
Technical Field
The invention relates to the technical field of image noise reduction processing, in particular to a video image noise reduction method, a video image noise reduction device and a video image noise reduction medium based on an improved local filtering algorithm.
Background
The image noise is a kind of random variation of brightness or color information in the image, and the photographed object itself does not have such variation, and is usually an expression of electronic noise. It is typically generated by the sensor and circuitry of a scanner or digital camera and may also be affected by film grain or shot noise that is unavoidable in an ideal photodetector. Image noise is an undesirable by-product of the image capture process, giving the image errors and additional information.
The video is composed of a plurality of continuous video frames, and one video frame is an image, so the video noise reduction comprises the noise reduction of the image in units of frames in the video. The image denoising processing refers to a technology of adopting a certain method, suppressing or eliminating noise in an image, and maintaining original textures and details of the image as much as possible to improve the visual quality of the image. Video noise reduction differs from the noise reduction of single frame images in that video can exploit the temporal and/or spatial correlation between video frames.
The existing noise reduction technologies include time domain noise reduction, space domain noise reduction, motion adaptive noise reduction and the like, and the existing time domain noise reduction technologies mainly include two types. The first method uses motion matching to search for similar blocks between the current frame and other frames of the video, and filters these similar blocks to remove noise. The second method is to detect a moving area in the current frame, for example, calculate the mean square error between image blocks to determine whether the image blocks move, then filter a static area, and reduce the filtering weight for the moving area or not. Existing spatial noise reduction uses pixel or block correlation to approximate, which mainly includes local and non-local two kinds of filtering algorithms. When estimating any pixel in an image, the filtering algorithm is a local algorithm if its reference points are spatially adjacent and the coefficients are affected by spatial distance. Many local filtering algorithms have been proposed in succession for removing noise, such as gaussian filtering, wiener filtering, least mean square filtering, training filtering, bilateral filtering, anisotropic filtering, guided kernel regression filtering, etc.
Wherein, the bilateral filtering is a classic nonlinear spatial domain filtering. The method simultaneously considers the spatial domain information and the pixel gray level similarity, is a compromise treatment for the spatial proximity and the pixel value similarity, and has the characteristic of edge protection and denoising. The principle of bilateral filtering is: local weighted averaging is performed, but unlike general gaussian filtering, the weighting coefficients of bilateral filtering are obtained by multiplying the spatial proximity factor and the luminance similarity factor. Bilateral filtering can well protect high-frequency detail information, and has great practical significance in the image preprocessing stage.
Specifically, in the process of shooting an image video, noise including gaussian noise, cedar noise and the like is inevitably generated due to various interference factors, so that the quality of the shot image is reduced. Because the method using motion matching usually adopts an optical flow method or a block matching method, the complexity is high, the computation amount is large, and the reduction of the time domain noise reduction weight of the motion area can cause the noise of the motion area not to be eliminated.
In the process of executing the algorithm, the local filtering sets a pixel frame by taking a pixel point to be filtered as a center, then traverses all pixel points in the pixel frame and executes the algorithm of the local filtering, and the defect of the local filtering is that the calculation amount of the algorithm is large, all the pixel points in the pixel frame do not need to be selected under many conditions, and all the pixel points can be represented only by selecting representative partial pixel points.
Disclosure of Invention
The invention provides a video image noise reduction method, a device and a medium based on an improved local filtering algorithm, which can at least solve one of the technical problems.
In order to realize the purpose, the invention adopts the following technical scheme:
a video image noise reduction method based on improved local filtering algorithm comprises the following steps,
judging whether the obtained current video frame is a first frame, if so, performing spatial domain noise reduction on the current video frame based on an improved local filtering algorithm;
if the current video frame is not the first frame, performing motion region detection on the current video frame according to the previous video noise reduction frame to obtain the motion condition of each pixel point;
judging whether the current pixel is static according to the motion condition of the pixel, and if so, performing time domain noise reduction on the current pixel based on an improved local filtering algorithm; if the current pixel point is not static, the following operations are carried out when the noise of the motion area is reduced:
subtracting the value of the same pixel point in the previous video noise reduction frame from the value of the current pixel point to obtain a change value of the current pixel point, processing the change value by using an improved local filtering algorithm, improving the local filtering algorithm on the change value to obtain a change value after filtering, and adding the change value after filtering and the value of the same pixel point in the previous video noise reduction frame to obtain a value of the current pixel point after filtering;
storing the filtered current video frame;
the improved local filtering algorithm is to change each pixel in the traversing set pixel frame into the mode that the odd pixel points of all odd rows are selected from the pixel points at the upper left corner of the pixel frame.
Further, the local filtering algorithm before the improvement of the improved local filtering algorithm is one of a bilateral filtering algorithm, a guided kernel regression filtering algorithm and a gaussian filtering algorithm.
Further, the detecting the motion region of the current video frame according to the previous video noise reduction frame to obtain the motion condition of each pixel point includes:
obtaining the motion quantization value of each pixel point of the current video frame through a formula (3):
in the above formula, move _ ref p Is the motion quantization value of pixel p, S is the pixel frame centered at p, w is the sampled pixel summation, I2 q Is the pixel value of the q point in the previous video noise reduction frame, I1 q Is the pixel value of the q point in the current frame;
and q is a pixel point obtained by sampling from the pixel frame S, and the sampling mode is that the odd pixel points of all odd rows are selected from the pixel point at the upper left corner of the pixel frame.
Further, the determining whether the current pixel is static according to the motion condition of the pixel includes:
comparing the motion quantization value of the pixel point with a preset threshold value, and if the motion quantization value of the pixel point is smaller than the preset threshold value, judging that the pixel point is static; and if the motion quantization value of the pixel point is larger than the preset threshold value, judging that the pixel point is not static.
Further, if the current pixel point is static, performing time domain noise reduction based on an improved local filtering algorithm on the current pixel point, where the improved local filtering algorithm is an improved bilateral filtering algorithm, and specifically includes:
the formula (4) and the formula (5) are utilized to reduce noise of the current pixel point,
res p is the noise-reduced result of pixel point p, W p The method is characterized in that the method is an intermediate variable of an algorithm, S is a pixel frame taking a p point as a center, S is sampled to reduce the operation amount, the sampling mode is that all odd numbered pixel points in odd rows are selected to participate in operation from pixel points at the upper left corner of the pixel frame S, | p-q | refers to the position difference value between the pixel point p and the pixel point q, and is the accumulation of the distance in the horizontal direction and the distance in the vertical direction, and I1 p Refers to the pixel value, σ, of the current frame p point s Mean variance of position, σ r Refers to the variance of the pixel difference, G (. Degree.) is a Gaussian function, I2 q Is the pixel value of the q point in the previous video noise reduction frame, I1 q Is the pixel value of the q point in the current frame.
Further, the denoising of the motion region uses an improved bilateral filtering algorithm, which specifically includes
diff＝I1–I2 (6)
res p ＝diff_denoise p +I2 P (9)
Wherein diff is the difference between the current frame and the previous video noise reduction frame, diff p Is the difference, diff, of a pixel p and the same pixel p in the previous video noise reduction frame q Is the difference of a pixel q and the same pixel q in the previous video noise reduction frame, W p Is an intermediate variable of the algorithm, I1 is the current frame, I2 is the previous video noise reduction frame, σ s Mean variance of position, σ r Refers to the variance of the pixel difference, G () is a Gaussian function, diff _ dense p Is to perform bilateral filtering on the p-point difference value, W p Is an intermediate variable of the algorithm, res p Is the result of the noise reduction of the pixel point p.
Furthermore, the pixel frame set by the improved local filtering algorithm is an N × N square, and each small block includes a pixel point, where N is an odd number.
Further, the preset threshold is adjustable, and the value of N in the N × N block is adjustable.
On the other hand, the invention also discloses a video image noise reduction device based on time domain, comprising:
the acquisition module is used for acquiring video frame data;
the judging module is used for judging whether the current video frame is a first frame or not;
a motion detection module, configured to perform the method according to claim 3, to determine whether a current pixel is stationary according to a motion condition of the pixel;
a local filtering module for performing the improved local filtering algorithm of claim 4;
and the storage module is used for storing the filtered current video frame.
Further, the preset threshold in the motion detection module is adjustable, and the value of N in the nxn block in the local filtering module is adjustable.
In yet another aspect, the present invention also discloses a computer readable storage medium storing a computer program, which when executed by a processor causes the processor to perform the steps of the method as described above.
According to the technical scheme, the image blocks are sampled, so that the calculated amount is reduced while the noise reduction quality is not reduced, and the noise of the motion area is reduced by filtering the change condition of the motion area. The noise reduction of the motion area can be realized by searching for similar blocks in front and back frames by using a motion matching method, but the calculation complexity is higher, and particularly the noise reduction efficiency of a high-resolution video is lower, so that the real-time performance cannot be met. The invention greatly reduces the calculated amount by carrying out interlaced and alternate sampling on the image blocks, and simultaneously carries out noise reduction on the change condition of the motion area so as to avoid motion matching.
The video image noise reduction method based on the improved local filtering algorithm utilizes the pixel correlation of the image in a time domain and a space domain and the randomness of noise to reduce noise, improves the image quality and simultaneously ensures that original details and textures of the image are not damaged and blurred. The invention can reduce the operation amount and improve the image noise reduction effect.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
fig. 2 is a schematic diagram of the structure of the apparatus of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention.
As shown in fig. 1, the method for reducing noise of a video image based on an improved local filtering algorithm according to this embodiment includes first capturing a video frame by a camera or obtaining the video frame from a storage device, and then determining whether the video frame is a first frame. If the frame is the first frame, the spatial filtering is performed, in order to maintain the edge information while reducing the noise, the spatial filtering of the embodiment adopts bilateral filtering, and in order to reduce the calculation amount while maintaining the noise reduction effect, the bilateral filtering is improved. The bilateral filtering principle is referred to as formula (1) and formula (2).
Wherein, BF [ I ]] p Representing the pixel value W of a pixel point p in an image after spatial filtering p Is an intermediate variable of the algorithm, S is a pixel frame with a p point as the center, | p-q | refers to the position difference between a pixel point p and a pixel point q, and the accumulation of the distance in the horizontal direction and the distance in the vertical direction, I p Refers to the pixel value, σ, of the p point s Mean variance of position, σ r Refers to the variance of the pixel difference, and G (is) is a gaussian function. In order to reduce the calculation amount, the embodiment of the application samples S, and the specific implementation mode is that the odd-numbered pixel points of all odd-numbered rows are selected to participate in the operation from the pixel point at the upper left corner of the pixel frame S.
And if the current frame is not the first frame, performing time-domain filtering. The temporal filtering firstly estimates the motion condition of each pixel point of the current frame, and the specific method is to obtain the motion matrix of the image by using the previous video noise reduction frame through a formula 3.
move_ref p The motion quantization value of a pixel point p is adopted, S is a pixel frame with p as the center, S is sampled to reduce the operation amount in the specific implementation, q is a pixel point obtained by sampling in S, the sampling mode is that the odd number pixel points of all odd lines are selected to participate in the operation from the pixel point at the upper left corner of the pixel frame S, w is the sum of the pixel points obtained by sampling, and I2 q Is the pixel value of the q point in the previous video noise reduction frame, I1 q Is the pixel value of the q point in the current frame. The motion matrix of the current frame is obtained through calculation, and then motion judgment is carried out, the specific implementation mode is that a preset threshold value is compared with a motion quantization value of a pixel point, if the threshold value is smaller than the threshold value, the pixel point is judged to be static, and static time domain noise reduction is carried out on the current pixel point; if the current pixel point is larger than the threshold value, the current pixel point is judged to be not static, and motion time domain noise reduction is carried out on the current pixel point. Furthermore, the preset threshold N is adjustable, and the size of the threshold N is adjusted according to the noise reduction effect of the video image before the video image is subjected to noise reduction.
The noise reduction is performed on the static area, and the specific implementation is to perform time domain noise reduction by using improved bilateral filtering, as shown in formula 4 and formula 5.
res p Is the noise-reduced result of pixel point p, W p Is an intermediate variable of the algorithm, S is a pixel box centered at the p-pointSampling S to reduce the operation amount, wherein the sampling mode is to select odd pixels in all odd lines to participate in operation starting from the pixel at the upper left corner of the pixel frame S, | p-q | refers to the position difference between the pixel p and the pixel q, the accumulation of the distance in the horizontal direction and the distance in the vertical direction, and I1 p Refers to the pixel value, σ, of the current frame p point s Mean variance of position, σ r Means the variance of the pixel difference, G (is) is a Gaussian function, I2 q Is the pixel value of the q point in the previous video noise reduction frame, I1 q Is the pixel value at point q in the current frame.
The invention firstly subtracts the previous video noise reduction frame from the current frame to obtain the change matrix of the previous video frame, because the time domain noise is superposed in the change of the two frames, the invention filters the change matrix between the two frames to eliminate the noise, and finally adds the filtered change pixel value to the previous video noise reduction frame to obtain the final noise reduction result.
The specific implementation method is that the value of the current pixel is subtracted by the value of the same pixel in the previous video denoising frame to obtain the change value of the current pixel, the change value is subjected to improved bilateral filtering to obtain the change value after filtering, the change value after filtering is added with the value of the same pixel in the previous video denoising frame to obtain the value of the current pixel after filtering, as shown in formula 6, formula 7, formula 8 and formula 9,
diff＝I1–I2 (6)
res p ＝diff_denoise p +I2 P (9)
wherein diff is the difference between the current frame and the previous video noise reduction frame, diff p Is the difference, diff, of a pixel p and the same pixel p in the previous video noise reduction frame q Is the difference of a pixel q and the same pixel q in the previous video noise reduction frame, W p Is an intermediate variable of the algorithm, I1 is the current frame, I2 is the previous video noise reduction frame, σ s Mean variance of position, σ r Refers to the variance of the pixel difference, G () is a Gaussian function diff _ dense p Is to perform bilateral filtering on the p-point difference, res p Is the result of the noise reduction of the pixel point p.
And finally, storing the video frame subjected to noise reduction in a storage device and carrying out a noise reduction process of the next video frame.
The pixel frame in the above formula is an N × N square, and each small block includes a pixel point, where N is an odd number. Furthermore, the size of N is adjustable, the numerical value of N is preset before the video image is subjected to noise reduction, and then the numerical value of N is adjusted according to the noise reduction effect of the video image.
Meanwhile, the filtering algorithm of the invention adopts bilateral filtering, which can be replaced by other airspace filters such as directional filtering, anisotropic filtering and the like.
The motion area judgment method adopts a mean square error mode, and can be replaced by other modes such as absolute difference sum or square difference sum and the like.
Fig. 2 is a video image noise reduction device based on improved bilateral filtering, which includes an obtaining module: the video frame data acquisition unit is used for acquiring video frame data; a judging module: the video frame judging module is used for judging whether the current video frame is a first frame or not; a motion detection module: the system is used for judging whether the current pixel is static according to the motion condition of the pixel, wherein the motion condition and the motion condition of the pixel are judged as above; a local filtering module: for performing a modified local filtering algorithm; a storage module: for storing the filtered current video frame.
In summary, the embodiment of the present invention samples the image blocks to reduce the amount of computation while not reducing the noise reduction quality, and filters the change situation of the motion region to reduce the noise of the motion region. Specifically, the method for reducing noise of the motion region can reduce noise of the motion region by searching for similar blocks in front and back frames by using a motion matching method, but has high computational complexity, particularly low noise reduction efficiency for high-resolution videos, and cannot meet the real-time property. The invention reduces the calculated amount to 1/4 of the original amount by carrying out interlaced and alternate sampling on the image blocks, and simultaneously carries out noise reduction on the change condition of the motion area so as to avoid motion matching.
In yet another aspect, the present invention also discloses a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of any of the methods described above.
In yet another aspect, the present invention also discloses a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of any of the methods described above.
In a further embodiment provided by the present application, there is also provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the steps of any of the methods of the embodiments described above.
It can be understood that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and for the explanation, examples and beneficial effects of the relevant contents, reference may be made to the corresponding parts in the above method.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by a computer program, which may be stored in a non-volatile computer readable storage medium, and when executed, may include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
In the embodiment, a bilateral filtering algorithm in a local filtering algorithm is used, and the local filtering algorithm includes gaussian filtering, wiener filtering, minimum mean square filtering, training filtering, bilateral filtering, anisotropic filtering, guided kernel regression filtering, and the like.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A video image noise reduction method based on improved local filtering algorithm is characterized by comprising the following steps,
judging whether the obtained current video frame is a first frame, if so, carrying out spatial domain noise reduction on the current video frame based on an improved local filtering algorithm;
if the current video frame is not the first frame, performing motion region detection on the current video frame according to the previous video noise reduction frame to obtain the motion condition of each pixel point;
judging whether the current pixel is static according to the motion condition of the pixel, and if so, performing time domain noise reduction on the current pixel based on an improved local filtering algorithm; if the current pixel point is not static, the following operations of noise reduction of the motion area are carried out:
subtracting the value of the same pixel point in the previous video noise reduction frame from the value of the current pixel point to obtain a change value of the current pixel point, processing the change value by using an improved local filtering algorithm to obtain a filtered change value, and adding the filtered change value and the value of the same pixel point in the previous video noise reduction frame to obtain a filtered value of the current pixel point;
storing the filtered current video frame;
the improved local filtering algorithm is to change each pixel in the traversing set pixel frame into the mode that the odd pixel points of all odd rows are selected from the pixel points at the upper left corner of the pixel frame.
2. The method of claim 1, wherein the modified local filtering algorithm is one of a bilateral filtering algorithm, a guided kernel regression filtering algorithm and a gaussian filtering algorithm.
3. The method of claim 1, wherein the detecting a motion region of a current video frame according to a previous video denoising frame to obtain a motion condition of each pixel point comprises:
obtaining a motion quantization value of each pixel point of the current video frame through a formula (3):
in the above formula, move _ ref p Is the movement of a pixel point pThe quantization value, S is the pixel frame centered at p, w is the sum of sampled pixel points, I2 q Is the pixel value of the q point in the previous video noise reduction frame, I1 q Is the pixel value of the q point in the current frame; and q is a pixel point obtained by sampling from the pixel frame S, and the sampling mode is that the odd pixel points of all odd lines are selected from the pixel point at the upper left corner of the pixel frame.
4. The method of claim 3, wherein the determining whether the current pixel is still according to the motion of the pixel comprises:
comparing the motion quantization value of the pixel point with a preset threshold value, and if the motion quantization value of the pixel point is smaller than the preset threshold value, judging that the pixel point is static; and if the motion quantization value of the pixel point is larger than the preset threshold value, judging that the pixel point is not static.
5. The method according to claim 3, wherein if the current pixel is still, performing temporal denoising based on the improved local filtering algorithm on the current pixel, wherein the improved local filtering algorithm is an improved bilateral filtering algorithm, and specifically comprises:
as shown by the formula (4) and the formula (5),
res p is the noise-reduced result of pixel point p, W p Is the intermediate variable of the algorithm, S is the pixel frame with p point as the center, and S is sampled to reduce the operation amount, the sampling mode is that the odd number of all the odd lines is selected from the pixel point at the upper left corner of the pixel frame SThe pixel points participate in the operation, | p-q | | refers to the position difference value between the pixel point p and the pixel point q, and is the accumulation of the distance in the horizontal direction and the distance in the vertical direction, I1 p Refers to the pixel value, σ, of the current frame p point s Mean variance of position, σ r Means the variance of the pixel difference, G (is) is a Gaussian function, I2 q Is the pixel value of the q point in the previous video noise reduction frame, I1 q Is the pixel value of the q point in the current frame.
6. The method for reducing noise of a video image based on an improved local filtering algorithm according to claim 1, wherein the reducing noise of the motion region by using the improved bilateral filtering algorithm specifically comprises:
the formulas (6), (7), (8) and (9) are used to reduce noise of the pixel points in the motion area,
diff＝I1–I2 (6)
res p ＝diff_denoise p +I2 P (9)
wherein diff is the difference between the current frame and the previous video noise reduction frame, diff p Is the difference, diff, of a pixel p and the same pixel p in the previous video noise reduction frame q Is the difference of a pixel q and the same pixel q in the previous video noise reduction frame, W p Is the intermediate variable of the algorithm, I1 is the current frame, I2 is the previous video noise reduction frame, σ s Mean variance of position, σ r Refers to the variance of the pixel difference, G () is a Gaussian function, diff _ dense p Is to perform bilateral filtering on the p-point difference, res p And the noise reduction result of the pixel point p is obtained.
7. The method of claim 1, wherein the improved local filtering algorithm has a frame of N × N pixels, each pixel comprises a pixel point, and N is an odd number.
8. The method of claim 7, wherein the predetermined threshold is adjustable, and the value of N in the NxN blocks is adjustable.
9. A time-domain based video image noise reduction apparatus, comprising:
the acquisition module is used for acquiring video frame data;
the judging module is used for judging whether the current video frame is a first frame or not;
a motion detection module, configured to perform the method according to claim 4, to determine whether a current pixel is stationary according to a motion condition of the pixel;
a local filtering module for performing the improved local filtering algorithm of claim 8;
and the storage module is used for storing the filtered current video frame.
10. A computer-readable storage medium, storing a computer program which, when executed by a processor, causes the processor to carry out the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211070387.9A CN115546047A (en) | 2022-09-02 | 2022-09-02 | Video image noise reduction method, device and medium based on improved local filtering algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211070387.9A CN115546047A (en) | 2022-09-02 | 2022-09-02 | Video image noise reduction method, device and medium based on improved local filtering algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115546047A true CN115546047A (en) | 2022-12-30 |
Family
ID=84726448
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211070387.9A Pending CN115546047A (en) | 2022-09-02 | 2022-09-02 | Video image noise reduction method, device and medium based on improved local filtering algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115546047A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116523765A (en) * | 2023-03-13 | 2023-08-01 | 湖南兴芯微电子科技有限公司 | Real-time video image noise reduction method, device and memory |
-
2022
- 2022-09-02 CN CN202211070387.9A patent/CN115546047A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116523765A (en) * | 2023-03-13 | 2023-08-01 | 湖南兴芯微电子科技有限公司 | Real-time video image noise reduction method, device and memory |
| CN116523765B (en) * | 2023-03-13 | 2023-09-05 | 湖南兴芯微电子科技有限公司 | Real-time video image noise reduction method, device and memory |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10963993B2 (en) | Image noise intensity estimation method, image noise intensity estimation device, and image recognition device | |
| WO2021217643A1 (en) | Method and device for infrared image processing, and movable platform | |
| US9202263B2 (en) | System and method for spatio video image enhancement | |
| US8508605B2 (en) | Method and apparatus for image stabilization | |
| EP3488388B1 (en) | Video processing method and apparatus | |
| KR100739753B1 (en) | Method and apparatus of bidirectional temporal noise reduction | |
| KR20060076176A (en) | A method of temporal noise reduction in video sequence ans system therefore | |
| CN111161172B (en) | Infrared image column direction stripe eliminating method, system and computer storage medium | |
| CN110796615B (en) | Image denoising method, device and storage medium | |
| US20170178293A1 (en) | Method for performing super-resolution on single images and apparatus for performing super-resolution on single images | |
| WO2017100971A1 (en) | Deblurring method and device for out-of-focus blurred image | |
| CN114586337B (en) | Video anti-shake optimization processing method and device and electronic equipment | |
| CN111652814A (en) | Video image denoising method and device, electronic equipment and storage medium | |
| CN108270945B (en) | Motion compensation denoising method and device | |
| CN116416268B (en) | Method and device for detecting edge position of lithium battery pole piece based on recursion dichotomy | |
| CN115546047A (en) | Video image noise reduction method, device and medium based on improved local filtering algorithm | |
| CN115294149A (en) | Astronomical image background extraction method and system | |
| CN112967206B (en) | Self-adaptive infrared image and video vertical line removing method based on image blocking | |
| CN112805989A (en) | Time domain filtering method and device based on global motion estimation and storage medium | |
| CN111652821A (en) | Low-light-level video image noise reduction processing method, device and equipment based on gradient information | |
| CN114372938A (en) | Image self-adaptive restoration method based on calibration | |
| KR20030082249A (en) | Motion adaptive spatial-temporal deinterlacing method | |
| CN114240778A (en) | Video denoising method and device and terminal | |
| CN112801903A (en) | Target tracking method and device based on video noise reduction and computer equipment | |
| CN112614078A (en) | Image noise reduction method and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |