CN105554494B - Snow point image detecting method and device and video quality detection device and system - Google Patents
Snow point image detecting method and device and video quality detection device and system Download PDFInfo
- Publication number
- CN105554494B CN105554494B CN201510900335.3A CN201510900335A CN105554494B CN 105554494 B CN105554494 B CN 105554494B CN 201510900335 A CN201510900335 A CN 201510900335A CN 105554494 B CN105554494 B CN 105554494B
- Authority
- CN
- China
- Prior art keywords
- image
- bright spot
- gray
- pixel
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/00002—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for
- H04N1/00005—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for relating to image data
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of snow point image detecting method and device and video quality detection devices and system, are related to field of image recognition.Method therein includes：Obtain the gray reference image of original image；The pixel for by gray value in gray reference image being more than luminance threshold is determined as bright spot pixel；Determine whether original image is snow point image according to the connectivity between distribution density of the bright spot pixel in gray reference image and bright spot pixel.By using the distribution density of bright spot pixel and the postsearch screening mode of connectivity in the gray reference image for calculating original image, the snow point image that can be accurately detected in original image, computation complexity is low and efficiency is higher.
Description
Technical field
The present invention relates to field of image recognition more particularly to a kind of snow point image detecting methods and device and video quality
Detection device and system.
Background technology
With the rapid development of video technique, video with video recording as safety-security area a kind of important means increasingly by
Pay attention to.Headend equipment is the basic component part of video monitoring system.However, headend equipment is often by dispersion mounted on wide
In outdoor utility scene, along with its quantity constantly increases, headend equipment is caused to be difficult to monitor and safeguard, easy tos produce various events
Barrier and quality problems, such as the snow point of monitor video, fuzzy, colour cast etc. are abnormal.
For video by the image construction of many frames, whether analysis video to a certain extent can be by analyzing certain frame with snow point
Whether image replaces with snow point.Snow point image is a kind of common image abnormity type, often headend equipment it is impaired or
Caused by junction contacts are bad.For the monitor video of magnanimity point, manually mode goes to select the equipment with snow point exception
Clearly extremely labor intensive, therefore, it is automatically to filter out the equipment with snow point image for security protection operation very much
It is necessary to.
Snow point evaluation algorithms common at present, such as classical edge detection, dust detection method, spatial domain detection algorithm nothing
With reference to evaluation algorithms, not only calculation amount is larger, and obtained evaluation result is also barely satisfactory, and is easy by image self-information
Amount enriches the influence of degree and image resolution ratio, is not suitable for video monitoring system.
Invention content
A technical problem to be solved of the embodiment of the present invention is：It is small, with high accuracy how calculation amount to be carried out to image
Snow point image detection.
The first aspect according to the ... of the embodiment of the present invention provides a kind of snow point image detecting method, including：It obtains original
The gray reference image of image；The pixel for by gray value in gray reference image being more than luminance threshold is determined as bright spot pixel；Root
Whether original image is determined according to the connectivity between distribution density of the bright spot pixel in gray reference image and bright spot pixel
For snow point image.
In one embodiment, the distribution density according to bright spot pixel in gray reference image and between bright spot pixel
Connectivity determine whether original image is that snow point image includes：In quantity and gray reference image by calculating bright spot pixel
The ratio of the quantity of all pixels obtains distribution density of the bright spot pixel in gray reference image, if distribution density is no more than
Distribution density threshold value, then original image is non-snow point image；If distribution density is more than distribution density threshold value, gray reference is calculated
The bright spot pixel centrifugal pump of image, if bright spot pixel centrifugal pump is more than discrete threshold values, original image is snow point image, if
Bright spot pixel centrifugal pump is less than discrete threshold values, then original image is non-snow point image.
In one embodiment, the bright spot pixel centrifugal pump of calculating gray reference image includes：Calculate gray reference image
The horizontal bright spot centrifugal pump and vertical bright spot centrifugal pump of middle bright spot pixel；By horizontal bright spot centrifugal pump and vertical bright spot centrifugal pump
Bright spot pixel centrifugal pump of the average value as image.
In one embodiment, method further includes：Calculate the average value and gray scale of the three primary colors value of color of original image
Value, if the difference of the average value of three primary colors value of color and gray value is less than colored threshold value, and gray value is more than gray threshold,
Original image is then determined as non-snow point image.
In one embodiment, method further includes：High-pass filtering processing is carried out to gray reference image, to reduce gray scale
Low-frequency noise in reference picture.
The second aspect according to the ... of the embodiment of the present invention provides a kind of video quality detection method, including：Obtain video flowing
In several frame images；Aforementioned any method is used to judge several frame images whether for snow point image；If calculating video flowing
The quantity of dry frame image moderate snow point image；When the quantity of snow point image is more than snow point image threshold, video flowing is snow point video
Stream.
In terms of third according to the ... of the embodiment of the present invention, a kind of device for snow point image detection is provided, including：Gray scale
Conversion module, the gray reference image for obtaining original image；Bright spot pixel extraction module, being used for will be in gray reference image
Gray value is more than that the pixel of luminance threshold is determined as bright spot pixel；Snow point image determining module is used for according to bright spot pixel in ash
The connectivity between distribution density and bright spot pixel in degree reference picture determines whether original image is snow point image.
In one embodiment, snow point image determining module includes：Distribution density computational submodule, for bright by calculating
The quantity of point pixel obtains bright spot pixel in gray reference image with the ratio of the quantity of all pixels in gray reference image
Distribution density；Discrete threshold values computational submodule, the bright spot pixel centrifugal pump for calculating gray reference image；Judge submodule
Block, for when distribution density is no more than distribution density threshold value, judging original image for non-snow point image；When distribution density is more than
Distribution density threshold value, and bright spot pixel centrifugal pump be more than discrete threshold values when, judge original image for snow point image；It is close when being distributed
Degree is less than distribution density threshold value, and bright spot pixel centrifugal pump is less than discrete threshold values, judges original image for non-snow point figure
Picture.
In one embodiment, discrete threshold values computational submodule further includes：Horizontal bright spot centrifugal pump computing unit, based on
Calculate the horizontal bright spot centrifugal pump of bright spot pixel in gray reference image；Vertical bright spot centrifugal pump computing unit, for calculating gray scale
The vertical bright spot centrifugal pump of bright spot pixel in reference picture；Average calculation unit, for by horizontal bright spot centrifugal pump with it is vertical bright
Bright spot pixel centrifugal pump of the average value of point centrifugal pump as image.
In one embodiment, device further includes white screen detection module, the three primary colors value of color for calculating original image
Average value and gray value, if the difference of the average value of three primary colors value of color and gray value is less than colored threshold value, and ash
Angle value is more than gray threshold, then original image is determined as non-snow point image.
In one embodiment, device further includes high-pass filtering module, for carrying out high-pass filtering to gray reference image
Processing, to reduce the low-frequency noise in gray reference image.
4th aspect according to the ... of the embodiment of the present invention provides a kind of system for video quality detection, including：Frame figure
As acquisition module, several frame images for obtaining video flowing；It is aforementioned any one be used for snow point image detection device, be used for
Whether judgment frame image is snow point image；Snow point image statistics module, several frame image moderate snow point diagrams for calculating video flowing
The quantity of picture；Snow point video flowing judging submodule, for when the quantity of snow point image is more than snow point image threshold, judging video
Stream is snow point video flowing.
By using calculate original image gray reference image in bright spot pixel distribution density and connectivity it is secondary
Screening mode, the snow point image that can be accurately detected in original image, computation complexity is low and efficiency is higher.
By referring to the drawings to the detailed description of exemplary embodiment of the present invention, other feature of the invention and its
Advantage will become apparent.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.
Fig. 1 shows the flow diagram of one embodiment of snow point image detecting method of the present invention.
Fig. 2 shows the partial pixel schematic diagrames of the gray reference image with bright spot pixel.
Fig. 3 shows the flow diagram of one embodiment of the discrete value calculating method of bright spot pixel of the present invention.
Fig. 4 shows the flow diagram of one embodiment of video quality detection method of the present invention.
Fig. 5 shows structural schematic diagram of the present invention for one embodiment of the system of video quality detection.
Fig. 6 shows structural schematic diagram of the present invention for one embodiment of the device of snow point image detection.
Fig. 7 shows the structural schematic diagram of one embodiment of snow point image determining module of the present invention.
Fig. 8 shows structural schematic diagram of the present invention for another embodiment of the device of snow point image detection.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Below
Description only actually at least one exemplary embodiment is illustrative, is never used as to the present invention and its application or makes
Any restrictions.Based on the embodiments of the present invention, those of ordinary skill in the art are not making creative work premise
Lower obtained every other embodiment, shall fall within the protection scope of the present invention.
The snow point image detecting method of one embodiment of the invention is described below with reference to Fig. 1.
Fig. 1 is the flow chart of one embodiment of snow point image detecting method of the present invention.As shown in Figure 1, the embodiment
Method includes：
Step S102 obtains the gray reference image of original image.
Gray reference image is the gray-scale map that original image obtains after gradation conversion.
Step S104, the pixel for by gray value in gray reference image being more than luminance threshold are determined as bright spot pixel.
Step S106, according to the connection between distribution density of the bright spot pixel in gray reference image and bright spot pixel
Property determines whether original image is snow point image.
In the gray reference image of original image, if the distribution density of bright spot pixel is relatively low, do not go out in image
Now a large amount of bright spot pixel, it is possible to determine that image is non-snow point image, if the distribution density of bright spot pixel is higher, original graph
As there is a possibility that snow point occur；If the connectivity between bright spot pixel is higher, illustrate that most of bright spot pixel is original
The snow point abnormal conditions that the content rather than original image of image itself occur illustrate if connectivity is relatively low between bright spot pixel
Mutually separated, bright spot pixel is that the possibility of snow point is higher.By using bright spot in the gray reference image for calculating original image
The distribution density of pixel and the postsearch screening mode of connectivity, the snow point image that can be accurately detected in original image, meter
Calculation complexity is low and efficiency is higher.
After step s 102, can also include the following steps：High-pass filtering processing is carried out to gray reference image, so as to
Reduce the low-frequency noise in gray reference image.By using this method, the proportion of high-frequency noise in the calculation can be protruded.
Since the snow point itself in gray reference image is also a kind of high-frequency noise, the proportion of snow point in the calculation can be protruded,
Promote the precision of snow point image detection.
It is close with reference to distribution of the bright spot pixel in gray reference image when whether determine original image is snow point image
Connectivity between degree and bright spot pixel.More specifically determine that method can refer to following steps：First, bright spot picture is calculated
The ratio of the quantity of element and the quantity of all pixels in gray reference image obtains point of the bright spot pixel in gray reference image
Cloth density；When distribution density is no more than distribution density threshold value, determine that original image is non-snow point image；When distribution density is more than
When distribution density threshold value, continue the bright spot pixel centrifugal pump for calculating gray reference image.If bright spot pixel centrifugal pump be more than from
Threshold value is dissipated, then original image is snow point image, if bright spot pixel centrifugal pump is less than discrete threshold values, original image is non-snow
Point image.
The calculating process of the distribution density of bright spot pixel is fairly simple, by being more than threshold value by the distribution density of bright spot pixel
The corresponding original image of gray reference image be determined as non-snow point image, to which to enter bright spot pixel discrete for only parts of images
The calculation stages of value, therefore the complexity of method can be reduced.
Wherein it is possible to calculate the centrifugal pump of the bright spot pixel of gray reference image using following methods：Gray scale is calculated first
The horizontal bright spot centrifugal pump of bright spot pixel and vertical bright spot centrifugal pump in reference picture, then by horizontal bright spot centrifugal pump with it is vertical bright
Bright spot pixel centrifugal pump of the average value of point centrifugal pump as gray reference image.Below with reference to Fig. 2 and Fig. 3, to bright spot pixel
Horizontal bright spot centrifugal pump and the computational methods of vertical bright spot centrifugal pump and the bright spot pixel centrifugal pump of gray reference image into
Row illustrates.
Fig. 2 is the partial pixel schematic diagram of the gray reference image with bright spot pixel.As shown in Fig. 2, the white in figure
Pixel is bright spot pixel, respectively bright spot pixel 22,24,26,28, wherein bright spot pixel 22,24 and 26 is located at same a line, bright
Point pixel 22 and 28 is located at same row.Bright spot pixel 22 is connected in the horizontal direction with 24, and the coordinate difference of horizontal direction is 1；
Bright spot pixel 24 is not connected in the horizontal direction with pixel 26, and the coordinate difference of horizontal direction is 3.Therefore, in the horizontal direction,
It can reflect the connection situation of bright spot pixel in the horizontal direction by the difference of the abscissa of the bright spot pixel positioned at same a line；
Similarly, in vertical direction, it can reflect that bright spot pixel is being hung down by the difference of the ordinate of the bright spot pixel positioned at same row
The upward connection situation of histogram.Pass through the vertical of the bright spot pixel of the abscissa difference and Ge Lie of the bright spot pixel of each row of COMPREHENSIVE CALCULATING
Coordinate difference can obtain the bright spot pixel centrifugal pump of gray reference image, to reflect bright spot pixel in gray reference image
Connectivity.
In the horizontal direction, if in two adjacent bright spot pixels, left side bright spot pixel is the previous of right side bright spot pixel
In pixel, such as Fig. 2, bright spot pixel 22 is 24 previous pixels in the horizontal direction, and bright spot pixel 24 is 26 in the horizontal direction
Previous pixel；In vertical direction,
If in two adjacent bright spot pixels, upside bright spot pixel is the previous pixel of downside bright spot pixel, such as Fig. 2
In, bright spot pixel 22 be 28 vertical direction previous pixel.The bright of one embodiment of the invention is specifically described below with reference to Fig. 3
The point discrete value calculating method of pixel.
Fig. 3 is the flow chart of one embodiment of the discrete value calculating method of bright spot pixel of the present invention.As shown in figure 3, the reality
The method for applying example includes：
Step S302, by previous bright spot picture in each bright spot pixel in the one-row pixels of gray reference image and horizontal direction
The summation of the difference of the abscissa of element is as row centrifugal pump.
Step S304, by the row centrifugal pump of each row pixel of gray reference image and with gray reference image picture
Horizontal bright spot centrifugal pump of the quotient of plain line number as image；
Step S306, by previous bright spot picture in each bright spot pixel in a row pixel of gray reference image and vertical direction
The summation of the difference of the ordinate of element is as row centrifugal pump；
Step S308, by the row centrifugal pump of each row pixel of gray reference image and with gray reference image picture
Vertical bright spot centrifugal pump of the quotient of plain columns as image；
Step S310, using the horizontal bright spot centrifugal pump with the average value of the vertical bright spot centrifugal pump as gray reference
The bright spot pixel centrifugal pump of image.
By using the above method, horizontally and vertically upper bright spot pixel and previous bright spot pixel can be passed through
Coordinate difference reflects that the connectivity of the bright spot pixel of gray reference image, the complexity of calculating are relatively low.
In the above-mentioned methods, the difference that can have both calculated bright spot pixel and previous bright spot pixel, can also be as needed,
When several bright spot pixels are the adjacent pixel in gray reference image, several bright spot pixels are connected as bright spot line segment, then count
Calculate bright spot line segment and previous bright spot pixel or the coordinate difference of previous bright spot line segment.
Further, it is also possible to directly following formula (1) be used to calculate bright spot pixel centrifugal pump：
Wherein, f (M) indicates the bright spot pixel centrifugal pump of gray reference image；I is the mark of bright spot pixel in vertical direction
Know, j is the mark of bright spot pixel in horizontal direction；M is the pixel columns of gray reference image, and n is the picture of gray reference image
Plain line number.By using above-mentioned formula, the ratio shared in total pixel of the coordinate difference between each bright spot pixel can be calculated
Mean value, to reflect the connectivity between bright spot pixel.
Original image is possible to as white screen image.If the noise in white screen image is less, bright spot is judged in the above method
White screen image can be correctly determined as non-snow point image by the step of connectivity between pixel.However, actual video prison
Control or the white screen image that defines of other Video Applications scenes not only only have the white screen image of full-mesh, also doped portion stain
Image.When only with the method for foregoing individual embodiments, this image is likely to be mistaken for snow point image.It therefore, can be with
By additional screening technique, the erroneous judgement of dialogue screen image is prevented.This method specifically includes：The three primary colors for calculating original image are color
The average value and gray value of color value, if the difference of the average value of three primary colors value of color and gray value is less than colored threshold value, and
And gray value is more than gray threshold, then original image is determined as non-snow point image.The gray value of white screen image is larger, and connects
It is bordering on the gray value of its gray reference image.By using the above method, white screen image can be identified from original image,
It prevents from white screen image being mistaken for snow point image, improves the accuracy of system.
The above method is also applied in video quality detection method.Video is by several group of picture at therefore, passing through
Whether every frame image of detection video has snow point to may determine that whether video quality problems occurs extremely.Have below with reference to Fig. 4
Body describes the video quality detection method of one embodiment of the invention.
Fig. 4 is the flow chart of one embodiment of video quality detection method of the present invention.As shown in figure 4, the embodiment
Method includes：
Step S402 obtains several frame images in video flowing.
Whether step S404 uses snow point image detecting method to judge several frame images for snow point image.
Specifically, the snow point image detecting method in previous embodiment may be used, such as including step in Fig. 1 or Fig. 3
Snow point image detecting method.
Step S406 calculates the quantity of several frame image moderate snow point images of video flowing.
Step S408 judges video flowing for snow point video flowing when the quantity of snow point image is more than snow point image threshold.
By way of the snow point detection of frame image in being converted to the detection of video quality to video flowing, it can be easy
Detect the quality problems that video occurs in ground.
The system for video quality detection of one embodiment of the invention is specifically described below with reference to Fig. 5.
Fig. 5 is structure chart of the present invention for one embodiment of the system of video quality detection.As shown in figure 5, the reality
The system for applying example includes：Frame image collection module 52, several frame images for obtaining video flowing；For snow point image detection
Whether device 54 is snow point image for judgment frame image；Snow point image statistics module 56, several frames for calculating video flowing
The quantity of image moderate snow point image；Snow point video flowing judging submodule 58, for being more than snow point image when the quantity of snow point image
When threshold value, judge video flowing for snow point video flowing.The frame image of acquisition is sent to for snow point figure by frame image collection module 52
As the device 54 of detection, in the video flowing that the statistics of snow point image statistics module 56 is detected for the device 54 of snow point image detection
Snow point image quantity, snow point video flowing judging submodule 58 determines that video is further according to the quantity of the snow point image detected
It is no defective in quality.
The device for snow point image detection of one embodiment of the invention is specifically described below with reference to Fig. 6.
Fig. 6 is structure chart of the present invention for one embodiment of the device of snow point image detection.As shown in fig. 6, the reality
The device 54 for snow point image detection for applying example includes：Gradation conversion module 642, the gray reference for obtaining original image
Image；Bright spot pixel extraction module 644, it is bright for gray value in gray reference image to be determined as more than the pixel of luminance threshold
Point pixel；Snow point image determining module 646, for the distribution density and bright spot according to bright spot pixel in gray reference image
Connectivity between pixel determines whether original image is snow point image.After gradation conversion module 642 will carry out gradation conversion
Gray reference image is sent to bright spot pixel extraction module 644, and bright spot pixel extraction module 644 determines in gray reference image
After bright spot pixel, snow point image determining module 646 carries out postsearch screening further according to the bright spot pixel of extraction.By using above-mentioned knot
Structure, the snow point image that can be accurately detected in original image, complexity is low and computational efficiency is high.
The snow point image determining module of one embodiment of the invention is specifically described below with reference to Fig. 7.
Fig. 7 is the structure chart of one embodiment of snow point image determining module of the present invention.As shown in fig. 7, the embodiment
Snow point image determining module 646 may include：Distribution density computational submodule 7462, for the quantity by calculating bright spot pixel
Distribution density of the bright spot pixel in gray reference image is obtained with the ratio of the quantity of all pixels in gray reference image；From
Dissipate threshold calculations submodule 7464, the bright spot pixel centrifugal pump for calculating gray reference image；Judging submodule 7466, is used for
When distribution density is no more than distribution density threshold value, judge original image for non-snow point image；When distribution density is more than that distribution is close
Spend threshold value, and bright spot pixel centrifugal pump be more than discrete threshold values when, judge original image for snow point image；When distribution density does not surpass
Distribution density threshold is crossed, and bright spot pixel centrifugal pump is less than discrete threshold values, judges original image for non-snow point image.Judge
Submodule 7466, can be comprehensive according to the result of calculation of distribution density computational submodule 7462 and discrete threshold values computational submodule 7464
Conjunction judges whether image is snow point image.
Wherein, discrete threshold values computational submodule 7464 can also include：Horizontal bright spot centrifugal pump computing unit 74642 is used
In the horizontal bright spot centrifugal pump for calculating bright spot pixel in gray reference image；Vertical bright spot centrifugal pump computing unit 74644, is used for
Calculate the vertical bright spot centrifugal pump of bright spot pixel in gray reference image；Average calculation unit 74646, for by horizontal bright spot from
Dissipate bright spot pixel centrifugal pump of the value with the average value of vertical bright spot centrifugal pump as image.Horizontal bright spot centrifugal pump and vertical bright spot
The circular of centrifugal pump is referred to the embodiment of aforementioned snow point image detecting method.
The device for snow point image detection of another embodiment of the present invention is specifically described below with reference to Fig. 8.
Fig. 8 is structure chart of the present invention for another embodiment of the device of snow point image detection.As shown in figure 8, should
The device 54 for snow point image detection of embodiment can also include white screen detection module 842, for calculating original image
The average value and gray value of three primary colors value of color, if the difference of the average value of three primary colors value of color and gray value is less than colour
Threshold value, and gray value is more than gray threshold, then original image is determined as non-snow point image.By the way that white screen detection module is arranged
842, it can prevent from white screen image being mistaken for snow point image.
In addition, the device 54 for snow point image detection can also include high-pass filtering module 844, for gray reference
Image carries out high-pass filtering processing, to reduce the low-frequency noise in gray reference image.By the way that high-pass filtering module is arranged
844, the snow point in gray reference image can be protruded, so that other modules carry out bright spot pixel extraction and snow point image really
It is fixed.
In addition, it is also implemented as a kind of computer program product according to the method for the present invention, the computer program product
Including computer-readable medium, be stored on the computer-readable medium for execute the present invention method in limit it is above-mentioned
The computer program of function.Those skilled in the art will also understand is that, various exemplary in conjunction with described in disclosure herein
Logical block, module, circuit and algorithm steps may be implemented as the combination of electronic hardware, computer software or both.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of snow point image detecting method, including：
Obtain the gray reference image of original image；
The pixel for by gray value in the gray reference image being more than luminance threshold is determined as bright spot pixel；
Institute is obtained by the ratio of the quantity and the quantity of all pixels in the gray reference image that calculate the bright spot pixel
State distribution density of the bright spot pixel in the gray reference image；
If the distribution density is no more than distribution density threshold value, the original image is non-snow point image；
If the distribution density is more than distribution density threshold value, the bright spot pixel centrifugal pump of the gray reference image is calculated；
If the bright spot pixel centrifugal pump is more than discrete threshold values, the original image is snow point image；
If the bright spot pixel centrifugal pump is less than discrete threshold values, the original image is non-snow point image.
2. according to the method described in claim 1, it is characterized in that, the bright spot pixel centrifugal pump for calculating gray reference image
Including：
Calculate the horizontal bright spot centrifugal pump of bright spot pixel described in the gray reference image and vertical bright spot centrifugal pump；
Using the horizontal bright spot centrifugal pump with the average value of the vertical bright spot centrifugal pump as the described bright of gray reference image
Point pixel centrifugal pump.
3. according to the method described in claim 1, it is characterized in that, further including：
The average value and gray value of the three primary colors value of color of the original image are calculated, if the three primary colors value of color is flat
Mean value and the difference of gray value are less than colored threshold value, and the gray value is more than gray threshold, then the original image is true
It is set to non-snow point image.
4. according to the method described in claim 1, it is characterized in that, further including：High pass filter is carried out to the gray reference image
Wave processing, to reduce the low-frequency noise in the gray reference image.
5. a kind of video quality detection method, including：
Obtain several frame images in video flowing；
The method described in any one of claim 1-4 is used to judge several frame images whether for snow point image；
Calculate the quantity of several frame image moderate snow point images of the video flowing；
When the quantity of the snow point image is more than snow point image threshold, the video flowing is snow point video flowing.
6. a kind of device for snow point image detection, including：
Gradation conversion module, the gray reference image for obtaining original image；
Bright spot pixel extraction module, it is bright for gray value in the gray reference image to be determined as more than the pixel of luminance threshold
Point pixel；
Snow point image determining module, including：
Distribution density computational submodule owns for the quantity by calculating the bright spot pixel with the gray reference image
The ratio of the quantity of pixel obtains distribution density of the bright spot pixel in the gray reference image；
Discrete threshold values computational submodule, the bright spot pixel centrifugal pump for calculating the gray reference image；
Judging submodule, for when the distribution density is no more than distribution density threshold value, judging the original image for non-snow
Point image；When the distribution density is more than distribution density threshold value, and the bright spot pixel centrifugal pump is more than discrete threshold values, sentence
The original image that breaks is snow point image；When the distribution density is less than distribution density threshold value, and the bright spot pixel from
Scattered value is less than discrete threshold values, judges the original image for non-snow point image.
7. device according to claim 6, which is characterized in that the discrete threshold values computational submodule further includes：
Horizontal bright spot centrifugal pump computing unit, for calculate the horizontal bright spot of bright spot pixel described in the gray reference image from
Dissipate value；
Vertical bright spot centrifugal pump computing unit, for calculate the vertical bright spot of bright spot pixel described in the gray reference image from
Dissipate value；
Average calculation unit is used for using the horizontal bright spot centrifugal pump with the average value of the vertical bright spot centrifugal pump as described in
The bright spot pixel centrifugal pump of gray reference image.
8. device according to claim 6, which is characterized in that further include white screen detection module, it is described original for calculating
The average value and gray value of the three primary colors value of color of image, if the difference of the average value and gray value of the three primary colors value of color
Value is less than colored threshold value, and the gray value is more than gray threshold, then the original image is determined as non-snow point image.
9. device according to claim 6, which is characterized in that further include high-pass filtering module, for joining to the gray scale
It examines image and carries out high-pass filtering processing, to reduce the low-frequency noise in the gray reference image.
10. a kind of system for video quality detection, including：
Frame image collection module, several frame images for obtaining video flowing；
The device for snow point image detection described in any one of claim 6-9, for judge the frame image whether be
Snow point image；
Snow point image statistics module, the quantity of several frame image moderate snow point images for calculating the video flowing；
Snow point video flowing judging submodule is used for when the quantity of the snow point image is more than snow point image threshold, described in judgement
Video flowing is snow point video flowing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510900335.3A CN105554494B (en) | 2015-12-09 | 2015-12-09 | Snow point image detecting method and device and video quality detection device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510900335.3A CN105554494B (en) | 2015-12-09 | 2015-12-09 | Snow point image detecting method and device and video quality detection device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105554494A CN105554494A (en) | 2016-05-04 |
| CN105554494B true CN105554494B (en) | 2018-09-11 |
Family
ID=55833404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510900335.3A Active CN105554494B (en) | 2015-12-09 | 2015-12-09 | Snow point image detecting method and device and video quality detection device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105554494B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107067392B (en) * | 2017-04-11 | 2020-11-10 | 北京小米移动软件有限公司 | Method and device for identifying screen-patterned image |
| CN109101912B (en) * | 2018-08-01 | 2019-04-16 | 南京旭羽睿材料科技有限公司 | Driving behavior intelligence adjustment system |
| CN109298228A (en) * | 2018-09-13 | 2019-02-01 | 安徽天尚清洁能源科技有限公司 | A kind of Intelligence Diagnosis method and system based on photovoltaic group string current anomaly |
| CN112118439B (en) * | 2019-06-20 | 2024-01-23 | 瑞昱半导体股份有限公司 | Video quality detection method and image processing circuit |
| CN111191517B (en) * | 2019-12-06 | 2023-02-28 | 清华大学合肥公共安全研究院 | Snow intensity detection image processing method and system |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5189649B2 (en) * | 2007-07-25 | 2013-04-24 | キャンデラ マイクロシズテムズ（エス）プライベート リミテッド | Exposure control for imaging systems |
| CN101312545A (en) * | 2008-07-02 | 2008-11-26 | 中兴通讯股份有限公司 | Image detection method and apparatus |
| CN101902557A (en) * | 2009-05-26 | 2010-12-01 | 南京敏思科技有限公司 | Reconstruction method and device of video image background |
| CN102421008A (en) * | 2011-12-07 | 2012-04-18 | 浙江捷尚视觉科技有限公司 | Intelligent video quality detecting system |
| CN103763515B (en) * | 2013-12-24 | 2017-08-11 | 浙江工业大学 | A kind of video abnormality detection method based on machine learning |
| CN103996177A (en) * | 2014-05-27 | 2014-08-20 | 武汉科技大学 | Snow noise removing algorithm free of reference detection |
| CN104794702B (en) * | 2014-10-08 | 2016-02-17 | 江苏省电力公司电力科学研究院 | A kind of based on the image blurring detection method of electric power video under the special scenes of image complexity analytical approach |
| CN104539936B (en) * | 2014-11-12 | 2017-09-12 | 广州中国科学院先进技术研究所 | There are the system and method for snow noise in a kind of monitoring video |
-
2015
- 2015-12-09 CN CN201510900335.3A patent/CN105554494B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105554494A (en) | 2016-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105554494B (en) | Snow point image detecting method and device and video quality detection device and system | |
| CN104737202B (en) | fire detection method and device | |
| CN103763515B (en) | A kind of video abnormality detection method based on machine learning | |
| KR100858140B1 (en) | Method and system for detecting a fire by image processing | |
| CN102348128B (en) | Surveillance camera system having camera malfunction detection function | |
| CN105453153B (en) | Traffic lights detects | |
| US20060221181A1 (en) | Video ghost detection by outline | |
| CN104394377B (en) | A kind of fuzzy abnormal recognition methods of monitoring image and device | |
| CN106686377B (en) | A kind of video emphasis area determination method based on deep-neural-network | |
| CN106067177A (en) | HDR scene method for detecting and device | |
| CN101944267A (en) | Smoke and fire detection device based on videos | |
| JP2012098045A (en) | Crack detector and crack detection program | |
| TW201322144A (en) | Apparatus and method for foreground detection | |
| KR101956250B1 (en) | Coastline monitoring apparatus and method using ocean color image | |
| CN105812618B (en) | A kind of method for testing motion and motion detection apparatus | |
| CN108364282A (en) | Image-mosaics detection method, image-mosaics detecting system | |
| JP2000295617A (en) | Method for detecting error block in compressed video sequence | |
| JP2020077165A (en) | Image processing device, image processing method, and program | |
| CN101727673A (en) | Method and unit for motion detection based on a difference histogram | |
| KR100234593B1 (en) | Method for high inspection of fabric | |
| JP2000137877A (en) | Fire detecting device | |
| JP2014229156A (en) | Fire detection device and fire detection method | |
| TWI465699B (en) | Method of water level measurement | |
| JP2024008990A (en) | Monitoring device, monitoring system, and monitoring method | |
| CN104112266B (en) | Image edge blurring detecting method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |