CN104378629A - Camera fault detection method - Google Patents
Camera fault detection method Download PDFInfo
- Publication number
- CN104378629A CN104378629A CN201410718592.0A CN201410718592A CN104378629A CN 104378629 A CN104378629 A CN 104378629A CN 201410718592 A CN201410718592 A CN 201410718592A CN 104378629 A CN104378629 A CN 104378629A
- Authority
- CN
- China
- Prior art keywords
- camera
- value
- partially
- illumination
- regression model
- 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.)
- Granted
Links
Landscapes
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a camera fault detection method. The camera fault detection method includes the following steps: analyzing the degrees of brightness and darkness of a camera video picture by an image processing algorithm to acquire an original diagnosis value; acquiring an illuminance value of an environment, where a camera is located, by an illuminometer; creating a low-illuminance Logistic regression model, normal-illuminance model and a high-illuminance Logistic regression model of camera faults, and integrating the three models into an illuminance regression model; calculating an illuminance regression value according to the illuminance regression model; correcting the original diagnosis value of the degrees of brightness and darkness according to the illuminance regression value; finally, judging whether the camera has a fault or not according to a corrected diagnosis result of the degrees of brightness and darkness. The camera fault detection method is combined with a hardware sensing method on the basis of the image processing algorithm, and thus, the defects in pure-software algorithm detection are overcome.
Description
Technical field
The present invention relates to safety-security area, particularly relate to fault of camera detection method.
Background technology
Along with the development of supervisory control system network, video monitoring system, at the future development towards scale, also strengthens gradually to the maintenance of supervisory control system and the difficulty of guarantee monitor video quality.It is video pictures Quality Down that fault of camera the most directly shows, and occurs partially bright, partially dark situation.In some specific occasions, higher to the quality requirement of monitor video, such as in sensitizing ranges such as mass activity place, traffic intersection, military base, harbour and airports, need the normal operation of real-time ensuring video camera.If CCTV camera occurs that abnormal or experience fault causes video quality decline but do not diagnosed in time again and solve, likely great loss can be caused.For large-scale supervisory control system, according to manually investigating fault, not only inefficiency, and the restriction being vulnerable to mankind's self character.Therefore, design a kind of fault of camera automatic testing method to be very important.
In current monitor product, the methods analyst video image quality adopting image procossing is conventional fault of camera detection mode.Whether whether this method abnormal abnormal with judgment device from video pictures analysis image.But affect video pictures quality because have the many factors such as environmental factor, equipment fault and human factor.Adopt the mode of video image analysis can not judge that picture is caused by environmental factor extremely, or equipment fault or human factor cause, and most of abnormal alarm is because the wrong report that produces of environmental factor.Ambient light illumination is partially bright or partially dark on the direct result that camera video picture affects.When ambient light illumination is excessive, the easy overexposure of video camera, video pictures will be partially bright; When ambient light illumination is too small, its brightness value is less than the minimal illumination of video camera, and video pictures is partially dark.When ambient light illumination is crossed bright or excessively dark, the fault of camera result adopting image algorithm to detect often is affected, and the unit exception result be now diagnosed by camera video picture is very insecure.Therefore, a kind of more effective mode of necessary invention detects fault of camera.
Summary of the invention
The object of the invention is the wrong report that minimizing environmental factor causes fault of camera to detect, increase the reliability of fault of camera detection method.
Technical scheme of the present invention is a kind of detection method of fault of camera.First this new method utilizes image processing algorithm to carry out analyzing the diagnostic value obtaining partially bright and partially dark degree to video pictures quality, then utilize illuminance meter to obtain ambient light illumination and to set up the regression model of ambient light illumination, finally utilize the partially bright of the regression model correcting video picture of illumination and partially dark degree diagnostic result, and judge whether video camera exists fault according to the partially bright and partially dark degree diagnostic result after correction.The method comprises following step:
(1) the partially bright of image processing method diagnosis camera video picture and partially dark degree is used
First, obtain continuous 32 two field pictures in camera video stream, calculate the average image of this 32 two field picture;
Secondly, calculate the luminance graph of the average image, and calculate the mean flow rate L of luminance graph
avg, calculate the standard deviation L of luminance graph
σ;
Finally, partially bright, partially dark degree is calculated respectively;
The computational methods of partially bright degree are as follows:
Wherein, α ∈ [0,1], if L
l< 0, then L
l=0;
In formula, L
lrefer to partially bright degree, L
avgrefer to the mean flow rate of luminance graph, L
σrefer to the standard deviation of luminance graph.L
llarger, partially bright degree is higher.If when the variance of luminance graph is larger, then in image, the brightness in some region is lower, can reduce partially bright degree to a certain extent.
The computational methods of partially dark degree are as follows:
Wherein, β ∈ [0,1], if L
d< 0, then L
d=0;
In formula, L
drefer to partially dark degree, L
avgrefer to the mean flow rate of luminance graph, L
σrefer to the standard deviation of luminance graph.L
d＝0。L
dlarger, partially dark degree is higher.If when the variance of luminance graph is larger, then in image, the brightness in some region is higher, can reduce partially dark degree to a certain extent.
(2) illuminance meter is used to obtain the brightness value of video camera place environment and set up illumination regression model
For partially bright and partially dark abnormity diagnosis value, the meeting higher or on the low side in ambient light illumination has influence on the partially bright and partially dark diagnostic value of camera views.The minimal illumination of video camera and overexposure illumination are exactly the threshold value higher and on the low side defined in ambient light illumination.When ambient light illumination is less than the minimal illumination of video camera or is greater than the overexposure illumination of video camera, if camera itself is normal, the ambient light illumination order of severity that partially bright or inclined shadow rings to picture increases gradually.The order of severity of picture exception represents the probability that video camera breaks down, and the order of severity is higher, and probability of malfunction is larger.In conjunction with object of the present invention, if picture occurs abnormal, be really now that the probability that video camera breaks down can reduce gradually along with the on the low side or higher of ambient light illumination, and be reduce to 0 from 1.Logistic regression model just can meet this relation, and video camera to break down be 0,1 two grouped datas.The present invention utilizes the minimal illumination of video camera to set up the low-light (level) Logistic regression model of fault of camera; The overexposure illumination of video camera is utilized to set up the high illumination Logistic regression model of fault of camera; Illumination is normal and picture is abnormal time fault of camera probability be 1.
Integrating low-light (level) Logistic regression model, normally illuminance model, high illumination Logistic regression model is the regression model of illumination.By the principle of the raw diagnostic result of the regression model correcting camera of illumination be: after given ambient illumination values and raw diagnostic value, be equivalent to the probability that when two conditions occur simultaneously, video camera can break down, adopt the product of illumination regression model value and raw diagnostic value as correction result thus.
The regression model of illumination in the present invention:
Wherein, l is the brightness value obtained from illuminance meter, L
minthe minimal illumination value of video camera, L
maxthe overexposure brightness value of video camera, e ≈ 2.71828.Low-light (level) Logistic regression model, normal illuminance model, high illumination Logistic regression model combine by this model.
(3) diagnostic result of illumination regression model correcting image Processing Algorithm is utilized
Camera views is use the regression model of brightness value to correct diagnostic value in the bearing calibration of partially bright and partially dark degree diagnostic result:
V(s
1)＝W
1(l)·D(s
1)
Wherein, s
1what represent is partially bright or partially dark; V (s
1) represent diagnosis item s
1diagnostic value after being corrected; W
1l () is the value that brightness value regression model obtains, be the equal of weight; D (s
1) be diagnosis item s
1raw diagnostic value in image processing algorithm; Work as s
1when representing partially bright, D (s
1) what represent is partially bright raw diagnostic value, D (s
1)=L
l; Work as s
1when representing partially dark, D (s
1) what represent is partially dark raw diagnostic value, D (s
1)=L
d.When brightness value is less than the minimal illumination of video camera, brightness value increases gradually, and regression model trends towards 1 gradually, and the diagnostic value of correction also trends towards raw diagnostic value gradually; When brightness value is in normal range (NR), the value of regression model is 1, and now correcting diagnostic value is raw diagnostic value; When brightness value is greater than the overexposure illumination of video camera, brightness value increases gradually, and regression model trends towards 0 gradually, corrects diagnostic value and trends towards 0.
(4) diagnostic value after correcting is utilized to judge whether video camera breaks down
With thresholding method, the diagnostic value after correction is dealt with, to judge whether video camera breaks down.To partially bright degree setting threshold T
l, T
l=0.7, if the partially bright diagnostic value after correcting is greater than 0.7, then illustrate that video camera breaks down; To partially bright degree setting threshold T
d, T
d=0.7, if the partially dark diagnostic value after correcting is greater than 0.7, then illustrate that video camera breaks down.
Beneficial effect: the detection method of fault of camera in the technical program, decreases the wrong report that environmental factor causes fault of camera to detect, and adds the reliability of fault of camera detection method.
Accompanying drawing explanation
Fig. 1 is the connection block diagram of the embodiment of the present invention;
Fig. 2 is the workflow diagram of another kind of embodiment of the present invention.
Mark in figure: 1-video camera; 2-ball machine; 3-pinhole camera; 4-DVR; 5-streaming media server; 6-diagnosis server; 7-Alarm Server; 8-wireless network; 9-illuminance meter.
Embodiment
Below in conjunction with accompanying drawing, preferably embodiment of the present invention is described in further detail:
The connection block diagram of the specific embodiment of the invention, as shown in Figure 1.Video image uses camera acquisition, and the front-end camera used comprises the various video cameras such as common camera 1, ball machine 2, pinhole camera 3, and each video camera is composed of unique ID.Ambient illumination values is obtained by illuminance meter 9, is placed with illuminance meter 9 in the environment residing for each video camera, and illuminance meter 9 also has unique ID, and the illuminance meter 9 of this environment associates according to ID with video camera.Video image stores and uses DVR 4, wherein comprises built-in hard disk video tape recorder and mixed type DVR.Monitor video is forwarded to diagnosis server 6 by streaming media server 5, diagnoses out raw diagnostic value with the diagnostic software based on image processing algorithm.Diagnosis server 6 inquires about the ID of corresponding illuminance meter 9 according to the video camera ID of Current Diagnostic, the brightness value of corresponding illuminance meter 9 is inquired about by wireless network 8, then correct raw diagnostic value with illumination regression model, then send warning message to Alarm Server 7 if there is abnormal.Diagnosis server 6 and illuminance meter 9 are that the installation of illuminance meter 9 is more convenient, wiring method is also simple by the benefit that wireless network 8 communicates.
See Fig. 1 and Fig. 2, a kind of detection method of fault of camera, concrete steps are as follows:
(1) camera video image is obtained
Diagnosis server 6 obtains the camera video stream of specifying from streaming media server 5.
(2) by the partially bright of image processing algorithm diagnosis camera views and partially dark degree
Diagnostic software takes out continuous 32 frame video images of wherein video camera, calculates the average image of this 32 two field picture, then calculates the luminance graph of the average image; Mean flow rate L is obtained to luminance graph
avgwith standard deviation L
σ; Finally calculate partially bright, partially dark degree respectively.
The computational methods of partially bright degree are as follows:
Wherein, α ∈ [0,1], if L
l< 0, then L
l=0;
In formula, L
lrefer to partially bright degree, L
avgrefer to the mean flow rate of luminance graph, L
σrefer to the standard deviation of luminance graph.
The computational methods of partially dark degree are as follows:
Wherein, β ∈ [0,1], if L
d< 0, then L
d=0;
In formula, L
drefer to partially dark degree, L
avgrefer to the mean flow rate of luminance graph, L
σrefer to the standard deviation of luminance graph.
(3) reading of illuminance meter 9 is obtained
Each video camera is composed of unique ID, and illuminance meter 9 also has unique ID.Video camera and illuminance meter 9 associate according to ID.Diagnosis server 6 inquires about the illuminance meter 9 of video camera association, is obtained the reading of corresponding illuminance meter 9 by wireless network 8.
(4) regressand value of illumination is calculated with illumination regression model
The regression model of illumination is a piecewise function, and it is combined low-light (level) Logistic regression model, normal illuminance model, high illumination Logistic regression model.The regression model of illumination is as follows:
Wherein, l is the brightness value obtained from illuminance meter 9.L
minthe minimal illumination value of video camera, L
min=100; L
maxthe overexposure brightness value of video camera, L
max=700; E ≈ 2.71828.
The minimal illumination of video camera is when the illuminance of subject is to a certain extent low and make the video level of the output of video camera low to scene light illumination during a certain setting.The present invention is the minimum ambient light illumination of the outputting video signal level of requirement video camera when remaining on 100%.By this requirement, the minimal illumination of the video camera that different company produces has certain difference, but most video camera is being less than L
minwhen=100, the outputting video signal level of video camera can be little by 100%.Therefore, in an embodiment, L
min=100.
The overexposure illumination of video camera is when the illuminance height of subject makes camera views produce the minimum ambient light illumination of Solarization effects to a certain extent.In experimentation, close the automatic gain function of all video cameras, when ambient light illumination reach 700 and above time, most of video camera easily produces overexposure situation.L in embodiment
max=700.
(5) partially bright and partially dark raw diagnostic value is corrected
After using illumination regression model to calculate the regressand value of illumination, can correct partially bright and partially dark raw diagnostic value, bearing calibration is as follows:
V(s
1)＝W
1(l)·D(s
1)
Wherein, s
1what represent is partially bright or partially dark; V (s
1) represent diagnosis item s
1diagnostic value after being corrected; W
1l () is the value that brightness value regression model obtains, be the equal of weight; D (s
1) be diagnosis item s
1raw diagnostic value in image processing algorithm.Work as s
1when representing partially bright, D (s
1) what represent is partially bright raw diagnostic value, D (s
1)=L
l; Work as s
1when representing partially dark, D (s
1) what represent is partially dark raw diagnostic value, D (s
1)=L
d.
(6) judge whether video camera breaks down
After partially bright and partially dark diagnostic value is corrected, span is [0,1] still.Getting partially bright abnormal threshold value is T
l=0.7, if the partially bright diagnostic value after correcting is greater than T
l, then illustrate that video camera breaks down; Getting partially dark abnormal threshold value is T
d=0.7, if the partially dark diagnostic value after correcting is greater than T
d, then illustrate that video camera breaks down.
Diagnosis server 6 obtains the correction result of all abnormal raw diagnostic values by above step.When correcting result and still occurring abnormal, then send fault of camera early warning to Alarm Server 7.
Above content is in conjunction with concrete preferred implementation further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations.For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, some simple deduction or replace can also be made, all should be considered as belonging to protection scope of the present invention.
Claims (4)
1. a detection method for fault of camera, is characterized in that, comprises following steps:
S1, use image processing algorithm are diagnosed camera video picture, and obtain the diagnostic value of partially bright and partially dark degree, the computational methods of partially bright degree are as follows:
Wherein, α ∈ [0,1], if L
l< 0, then L
l=0;
In formula, L
lrefer to partially bright degree, L
avgrefer to the mean flow rate of luminance graph, L
σrefer to the standard deviation of luminance graph;
The computational methods of partially dark degree are as follows:
Wherein, β ∈ [0,1], if L
d< 0, then L
d=0;
In formula, L
drefer to partially dark degree, L
avgrefer to the mean flow rate of luminance graph, L
σrefer to the standard deviation of luminance graph;
S2, use illuminance meter obtain the brightness value of video camera place environment, calculate the regressand value of illumination with illumination regression model; Described illuminance model is as follows:
In formula, W
1l () is the value that brightness value regression model obtains, l is the brightness value obtained from illuminance meter, L
minthe minimal illumination value of video camera, L
maxthe overexposure brightness value of video camera, e ≈ 2.71828;
The regression model of S3, use brightness value corrects diagnostic value L
land L
d:
V(s
1)＝W
1(l)·D(s
1)
In formula, s
1what represent is partially bright or partially dark; V (s
1) represent diagnosis item s
1diagnostic value after being corrected; W
1l () is the value that brightness value regression model obtains, be the equal of weight; D (s
1) be diagnosis item s
1raw diagnostic value in image processing algorithm.Work as s
1when representing partially bright, D (s
1) what represent is partially bright raw diagnostic value, D (s
1)=L
l; Work as s
1when representing partially dark, D (s
1) what represent is partially dark raw diagnostic value, D (s
1)=L
d;
Diagnosis after S4, utilization correct judges whether video camera breaks down, if the partially bright or partially dark degree diagnostic value after correcting is greater than certain threshold value, then illustrates that fault has appearred in this video camera.
2. the detection method of fault of camera according to claim 1, is characterized in that: described step S1 comprises step by step following:
S101, obtains continuous 32 two field pictures in camera video stream, calculates the average image of this 32 two field picture;
S102, calculates the luminance graph of the average image, and calculates the mean flow rate L of luminance graph
avg, calculate the standard deviation L of luminance graph
σ;
S103, calculates partially bright, partially dark degree respectively.
3. the detection method of fault of camera according to claim 1, is characterized in that: described step S2 comprises step by step following:
S201, acquisition illuminance meter reading;
The minimal illumination of S202, use video camera sets up the low-light (level) Logistic regression model of fault of camera; Use video camera overexposure illumination to set up the high illumination Logistic regression model of fault of camera, in conjunction with these two kinds of models and normal illuminance model, integrate out the regression model of illumination;
S203, calculate the regressand value of illumination by illumination regression model.
4. the detection method of fault of camera as claimed in claim 1, is characterized in that: arranging partially bright abnormal threshold value is 0.7, and arranging partially dark abnormal threshold value is 0.7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410718592.0A CN104378629B (en) | 2014-12-01 | 2014-12-01 | A kind of detection method of fault of camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410718592.0A CN104378629B (en) | 2014-12-01 | 2014-12-01 | A kind of detection method of fault of camera |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104378629A true CN104378629A (en) | 2015-02-25 |
| CN104378629B CN104378629B (en) | 2016-08-24 |
Family
ID=52557242
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410718592.0A Active CN104378629B (en) | 2014-12-01 | 2014-12-01 | A kind of detection method of fault of camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104378629B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107272637A (en) * | 2017-06-06 | 2017-10-20 | 武汉瑞科兴业科技有限公司 | A kind of video monitoring system fault self-checking self- recoverage control system and method |
| CN107705334A (en) * | 2017-08-25 | 2018-02-16 | 北京图森未来科技有限公司 | A kind of video camera method for detecting abnormality and device |
| CN108737741A (en) * | 2017-12-21 | 2018-11-02 | 西安工业大学 | A kind of auto Anti-Blooming system of night Computer Vision |
| CN116866545A (en) * | 2023-06-30 | 2023-10-10 | 荣耀终端有限公司 | Mapping relation adjustment method, equipment and storage medium of camera module |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101783970A (en) * | 2009-12-22 | 2010-07-21 | 新太科技股份有限公司 | Methods, devices and systems for automatically detecting and managing fault of camera |
| CN102348128A (en) * | 2010-07-30 | 2012-02-08 | 株式会社日立制作所 | Surveillance camera system having camera malfunction detection function |
| CN103021138A (en) * | 2012-12-28 | 2013-04-03 | 广州市浩云安防科技股份有限公司 | Method for sensing shielding of video camera and device for implementing method |
-
2014
- 2014-12-01 CN CN201410718592.0A patent/CN104378629B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101783970A (en) * | 2009-12-22 | 2010-07-21 | 新太科技股份有限公司 | Methods, devices and systems for automatically detecting and managing fault of camera |
| CN102348128A (en) * | 2010-07-30 | 2012-02-08 | 株式会社日立制作所 | Surveillance camera system having camera malfunction detection function |
| CN103021138A (en) * | 2012-12-28 | 2013-04-03 | 广州市浩云安防科技股份有限公司 | Method for sensing shielding of video camera and device for implementing method |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107272637A (en) * | 2017-06-06 | 2017-10-20 | 武汉瑞科兴业科技有限公司 | A kind of video monitoring system fault self-checking self- recoverage control system and method |
| CN107272637B (en) * | 2017-06-06 | 2019-08-30 | 武汉瑞科兴业科技有限公司 | A kind of video monitoring system fault self-checking self- recoverage control system and method |
| CN107705334A (en) * | 2017-08-25 | 2018-02-16 | 北京图森未来科技有限公司 | A kind of video camera method for detecting abnormality and device |
| CN107705334B (en) * | 2017-08-25 | 2020-08-25 | 北京图森智途科技有限公司 | Camera abnormity detection method and device |
| CN108737741A (en) * | 2017-12-21 | 2018-11-02 | 西安工业大学 | A kind of auto Anti-Blooming system of night Computer Vision |
| CN116866545A (en) * | 2023-06-30 | 2023-10-10 | 荣耀终端有限公司 | Mapping relation adjustment method, equipment and storage medium of camera module |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104378629B (en) | 2016-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021042816A1 (en) | Method and device for detecting fault in monitoring apparatus | |
| JP4718253B2 (en) | Image abnormality detection device for surveillance camera | |
| CN109299703B (en) | Method and device for carrying out statistics on mouse conditions and image acquisition equipment | |
| US20170195556A1 (en) | Monitoring system and camera device | |
| CN104598895A (en) | Method and device for flame detection based on video image analysis | |
| US9076037B2 (en) | Image processing apparatus and method | |
| CN103702111B (en) | A kind of method detecting camera video color cast | |
| CN105828065B (en) | A kind of video pictures overexposure detection method and device | |
| CN104378629A (en) | Camera fault detection method | |
| CN102176758A (en) | Video quality diagnosis system and realization method thereof | |
| KR102592231B1 (en) | Method for diagnosing fault of camera | |
| CN104168478A (en) | Video image off-color detection method based on Lab space and correlation function | |
| KR20210097500A (en) | Apparatus for Monitoring Fire And System having the same | |
| WO2018149322A1 (en) | Image identification method, device, apparatus, and data storage medium | |
| CN110929608A (en) | Safety helmet wearing detection method and system | |
| US11113798B2 (en) | Collating system, collating method, and camera device | |
| CN110928305B (en) | Patrol method and system for patrol robot of railway passenger station | |
| JP5710230B2 (en) | Monitoring system and monitoring method | |
| JP5792671B2 (en) | Abnormality detection method and abnormality detection device | |
| CN106067963B (en) | Distributed automatic report method for remote large monitor abnormity | |
| JP2011061651A (en) | Suspicious object detection system | |
| JP2008271329A (en) | Institution monitoring device and institution monitoring method | |
| CN117152890B (en) | Designated area monitoring method, system and terminal | |
| KR200497533Y1 (en) | Livestock disease monitoring server | |
| CN103517063A (en) | Video quality diagnosis system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C56 | Change in the name or address of the patentee | ||
| CP03 | Change of name, title or address |
Address after: 511400, No. 2201, building 2, Cheonan headquarters, No. 555, Panyu Avenue, east ring street, Guangzhou, Guangdong, Panyu District Patentee after: Polytron Technologies Inc Address before: 511400 Guangdong city of Guangzhou province Panyu District City Bridge Street Liang Lu Gold Garden East two Street No. 16 101 Patentee before: Guangzhou Haoyun Security Technology Co., Ltd. |