CN106331605B - A kind of floods detection system and method based on video - Google Patents
A kind of floods detection system and method based on video Download PDFInfo
- Publication number
- CN106331605B CN106331605B CN201610686278.8A CN201610686278A CN106331605B CN 106331605 B CN106331605 B CN 106331605B CN 201610686278 A CN201610686278 A CN 201610686278A CN 106331605 B CN106331605 B CN 106331605B
- Authority
- CN
- China
- Prior art keywords
- information
- ponding
- video
- time
- floods
- 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
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/10—Alarms for ensuring the safety of persons responsive to calamitous events, e.g. tornados or earthquakes
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A50/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE in human health protection, e.g. against extreme weather
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Emergency Alarm Devices (AREA)
- Image Processing (AREA)
- Alarm Systems (AREA)
- Traffic Control Systems (AREA)
Abstract
The floods detection system based on video that the present invention relates to a kind of, including water level acquisition module, video acquisition module and image analysis module；The video acquisition module is used for collection site video data, and sends image analysis module for video data；Described image analysis module is for handling the video frame in video data according to DRR algorithm, DWD algorithm and DFD algorithm, each video frame is analyzed and processed, according to the marginal information and color information of object in each video frame, judge that road surface whether there is ponding, and judges whether to generate flood information；The water level acquisition module is for acquiring water level information.The present invention, which realizes, to be passed through video analysis and detects whether flood information occurs, and guarantees that more intelligent information can be provided while video monitoring, to improve the accuracy and safety of video monitoring.
Description
Technical field
The present invention relates to image procossing and intelligent identification technology field, more specifically, it relates to a kind of based on video
Floods detection device and method.
Background technique
Today's society science and technology high speed development, artificial intelligence, computer machine vision technique, automatic technology become to get over
Come more mature, therefore the requirement of video monitoring is also higher and higher, not only before simple video monitoring, also to have base
This intelligent recognition such as automatically processes at the functions.Road or community road surface can generate product after many city summer heavy rainfalls now
Water, or even floods situation is generated, general city video monitoring can monitor in this case, but need personnel one by one
Camera is checked, therefore automatic detection floods are even more particularly important, not only wants to calculate these scenes with the presence or absence of product
Water also requires to identify the information such as personnel, the vehicle of ponding region, has detected whether floods situation, if having personnel
Vehicle is stranded phenomenon.
There are sudden, risk for floods accident, in some instances it may even be possible to the life security of crisis to personnel, so if energy
The generation that accidents happened is effectively detected by some way, can preferably complete to monitor, prompting related personnel earlier carries out
Relief.
Application No. is a kind of Chinese invention patent of CN201110364891.5 " railway floods based on radio frequency signal attenuation
Detection and prior-warning device ", it is mainly characterized by detecting using the variation of radio frequency unit received signal intensity value and early warning iron
The generation of floods on road；The present invention is that the present invention, can automatic identification area mainly using image recognition as foundation with its main distinction
Lane information in domain, and it is aided with water level information, it can determine whether to generate floods alarm, and when floods occur, can continue to know
Not there are personnel and vehicle is detained situation, and recognises that generation personnel are stranded and the trapped alarm of vehicle；Not only floods are known
Method for distinguishing is different, and identifies that floods generation also can detect that with the presence or absence of personnel and the trapped information of vehicle.
Summary of the invention
In view of this, it is necessary in view of the above-mentioned problems, providing a kind of floods detection device and method based on video, realize
The monitoring and alarm of floods accident information, and the device for being sent alert information by network.The present invention, which realizes, to be passed through
Video analysis simultaneously detects whether flood information occurs, and guarantees that more intelligent information can be provided while video monitoring, to improve
The accuracy and safety of video monitoring.
To achieve the goals above, technical scheme is as follows:
A kind of floods detection system based on video, including water level acquisition module, video acquisition module and image analysis mould
Block；The video acquisition module is used for collection site video data, and sends image analysis module for video data；The figure
As analysis module is used to according to DRR algorithm, DWD algorithm and DFD algorithm handle the video frame in video data, to each
Video frame is analyzed and processed, and according to the marginal information and color information of object in each video frame, judges that road surface whether there is
Ponding, and judge whether to generate flood information；The water level acquisition module is for acquiring water level information.
Preferably, the video acquisition module includes camera lens, video sensor and video processor；The camera lens is used
In collection site video image information；The video sensor is used to video pictures being converted to digital information, at the video
Reason device is used to receive and process the video data of video sensor acquisition.
Preferably, the water level acquisition module includes ultrasonic water level gauge, for acquiring water level information.
As shown in Fig. 2, described image analysis module includes image pre-processing unit, monitoring area self-identifying unit, ponding
Detection unit and floods detection unit；
Described image pretreatment unit handles each video frame, obtains video frame for obtaining video data
Object edge information and color information in picture；
The monitoring area self-identifying unit is used for according to marginal information and color information automatic identification road edge information,
The related objective information of automatic setting detection zone；
The water detection unit is for detecting ponding information；
The floods detection unit is used to detect whether that floods occur according to video and ponding information, if having personnel and
The trapped information of vehicle.
Preferably, further including video encoding module and video transmission module, the video encoding module is used for view
Frequency carries out compressed encoding processing, and the video transmission module is for sending coding module rear video data by network.
A kind of floods detection method based on video, comprising the following steps:
S1, acquisition video data, and water level information is acquired in real time；
S2, machine vision processing is carried out to the image of video data, mainly includes gray proces, binary conversion treatment, edge
Detection, color conversion；
S3, the road edge information within the scope of automatic identification picture is carried out using DRR algorithm, and detection range is set；
S4, it is analyzed in setting range using DWD algorithm with the presence or absence of ponding, ponding range, the color of ponding；
S5, analyzed using DFD algorithm, whether generate floods according in the interpretation of result floods of step S4, if exist by
Oppressive member or vehicle, and if so, generating warning information；
S6, data and warning information are sent by network, mainly includes that floods generate, floods are oppressive, floods are tired
The warning information of vehicle.
Preferably, the step S2 is specifically included:
S21, gray proces are done to the video frame images in video data, obtains gradation data；
S22, binary conversion treatment is carried out to gradation data；
S23, Canny algorithm process is carried out to the data after binary conversion treatment, obtains the object edge in image；
S24, HoughLines straight-line detection is carried out to obtained object edge data, detects and whether there is straight line in picture
Information；
S25, color conversion is carried out to image, switchs to RGB from yuv data, saves the image color information of picture, obtained every
The RGB color multimedia message of one pixel.
Preferably, the step S3 is specifically included:
S31, classify to obtained data, remove other interference informations, only retain the oblique line of picture middle section, this
A little oblique lines are the information at the edge of road；
S32, in-between color information, the comprehensive RGB information for obtaining the part among oblique line, with known road surface are extracted
Color information be compared, judge whether it is road surface, if not detecting information of road surface, it is new to return to detection again
Video frame；
S33, if it is detected that road surface, statistics continuously detects whether the number on road surface is greater than N；If the then automatic detection of judgement
Mark, unidentified road out if not and in M minute, it is determined that do not detect lane information, need manual setting lane line, repair
It is no for changing automatic detection mark；
S34, the automatic detection mark of judgement, if being then target monitoring region by pavement edge straight line information preservation, if not
Then manual setting pavement edge information and save as target monitoring region.
Preferably, the step S4 is specifically included:
S41, monitoring area water level information is obtained, ponding is judged whether there is according to the height of water level；It will if no ponding
No ponding information summarizes, and monitoring area is detected if having ponding with the presence or absence of marginal information
If S42, ponding do not cover sideline on both sides of the road, current ponding does not cover entire road, needs to judge to accumulate
The depth of water；If ponding covers entire road, need to judge the ponding region on road with the presence or absence of water stream characteristics, judgement
Whether its color information is consistent with original pavement of road color information；
If the color information of the ponding region identified in S43, picture is still consistent with the color information on original road surface, it is believed that
Without flowing water feature in current picture, depth of accumulated water need to be judged；Judge whether that water level is more than Hcm according to water level information, if it exceeds
Then it is judged as there is ponding, is not above, thinks no ponding, then will summarize whether there is or not ponding information.
Preferably, the step S5 is specifically included:
If S51, detecting ponding, the time T10 detected is recorded, and only record for the first time, records the number detected
C1；It continues to test, if detecting, the number of C1 is greater than M1 times, time T11 at this time is recorded, if T11-T10 is greater than N1 minutes,
Floods occur for judgement；If not detecting ponding, need to judge whether that floods had occurred；
If floods S52, have occurred, the time T20 of ponding is not detected in record, and the number of ponding is not detected in statistics
C2 is continued to test, if detecting, the number of C2 is greater than M2 times, records time T21 at this time, when T21-T20 is greater than N2 minutes,
Judge that floods restore；
If floods S53, have occurred, judge floods region whether with the presence of personnel, the stranded warning of personnel whether has occurred,
And it judges whether there is vehicle and is detained situation, the stranded alarm of vehicle whether has occurred；
S54, according in image after binaryzation whether there is humanoid profile to determine whether have personnel be detained situation, detect
When personnel are detained, the time T30 detected is recorded, and only record for the first time, records the number C3 detected, continues to record, if
It detects that the number of C3 is greater than M3 times, records time T31, if T31-T30 is greater than N3 minutes, judges that generation personnel are stranded and accuse
It is alert；When judgement had generated, personnel are stranded to be alerted, and currently detected that no personnel are detained situation, then record current time
T40 only records the time of first time, records the number C4 detected；It continues to test, if the number of C4 is greater than M4 times, when record
Between T41, T41-T40 is greater than N4 minute, then judges the stranded releasing of personnel；
S55, judge whether ponding region has vehicle to be detained feelings according to whether there is vehicle contour in image after binaryzation
Condition；Judge whether current region occurred the stranded alarm of vehicle, warning information was summarized if having occurred；Detect vehicle
When delay, the time T50 detected is recorded, and only record for the first time, records the number C5 detected, continues to test, time of C5
Number is greater than M5 times, and records time T51, and T51-T50 is greater than N5 minutes, then judges that the stranded alarm of vehicle occurs；Detecting does not have
Vehicle is detained situation, then records current time T60, only records the time of first time, records the number C6 detected, continue to test
Number to C6 is greater than M6 times, and records time T61, and T61-T60 is greater than N6 minutes, then judges that vehicle is stranded and release.
Compared with prior art, the beneficial effects of the present invention are:
1, the image got by analyzing camera, can be by the Function Extension of video camera in city video monitoring, not only
It is limited only to video output, can be with hidden danger such as intelligent recognition floods, and related warning information is complained to, enrich video monitoring
In video camera function.
2, by computer machine vision technique, automatic identification correlation circumstance avoids manually frequently checking each camera shooting
The video of head.
3, floods hidden danger and correlation circumstance can effectively be found, remind related personnel in time, far from relevant range, avoid thing
Therefore occur.
Detailed description of the invention
Fig. 1 is the system principle diagram of the embodiment of the present invention；
Fig. 2 is the flow chart of the image analysis module of the embodiment of the present invention；
Fig. 3 is the monitoring area self-identifying module DRR algorithm flow chart of the embodiment of the present invention；
Fig. 4 is the water detection cells D WD algorithm flow chart of the embodiment of the present invention；
Fig. 5 is the floods detection unit DFD algorithm flow chart of the embodiment of the present invention.
Fig. 6 is a specific implementation step figure of the embodiment of the present invention.
Specific embodiment
With reference to the accompanying drawings and examples to a kind of floods detection device and method work based on video of the present invention
It further illustrates.
It is a kind of preferred example of floods detection device and method based on video of the present invention below, not therefore
It limits the scope of protection of the present invention.
Fig. 1 to Fig. 5 shows a kind of floods detection system based on video, including water level acquisition module, video acquisition mould
Block and image analysis module；The video acquisition module is used for collection site video data, and sends image for video data
Analysis module；Described image analysis module is used for according to DRR algorithm, DWD algorithm and DFD algorithm to the video in video data
Frame is handled, and is analyzed and processed to each video frame, according to the marginal information and color information of object in each video frame,
Judge that road surface whether there is ponding, and judges whether to generate flood information；The water level acquisition module is for acquiring water level information.
Preferably, the video acquisition module includes camera lens, video sensor and video processor；The camera lens is used
In collection site video image information；The video sensor is used to video pictures being converted to digital information, at the video
Reason device is used to receive and process the video data of video sensor acquisition.
Preferably, the water level acquisition module includes ultrasonic water level gauge, for acquiring water level information.
As shown in Fig. 2, described image analysis module includes image pre-processing unit, monitoring area self-identifying unit, ponding
Detection unit and floods detection unit；
Described image pretreatment unit handles each video frame, obtains video frame for obtaining video data
Object edge information and color information in picture；
The monitoring area self-identifying unit is used for according to marginal information and color information automatic identification road edge information,
The related objective information of automatic setting detection zone；
The water detection unit is for detecting ponding information；
The floods detection unit is used to detect whether that floods occur according to video and ponding information, if having personnel and
The trapped information of vehicle.
Specifically, described image analysis module process flow includes: image capture module, real time video data is obtained, it is right
Image carries out gray proces, carries out gray proces to video data, carries out edge detection to image, the object wheel in detection image
Wide and edge carries out binaryzation to image, and carries out edge detection, and detection object edge contour information carries out color to image and turns
It changes, extracts the color information of relevant range, color conversion is carried out to video data and save the color information of relevant range；It is comprehensive
The object edge information and color information of image, integrate data and are saved；Judge that area to be tested whether there is, checks
Whether area to be tested has been arranged；Road edge in image is identified, and sets area to be tested automatically, when area to be detected
When domain is unable to automatic identification and comes out, needs manual setting and save area to be tested；
Water level acquisition module, water level acquisition module collects water level information, and gives water detection unit；
It according to DWD algorithm, detects whether there are ponding, this process belongs to water detection unit, information in summary, inspection
It surveys and whether there is ponding in current video；
The analysis of DFD algorithm detects whether the information such as generation floods, this process belongs to floods detection unit, according to image
The data informations such as object edge information and color information and ponding information detect whether that floods situation occurs；
When detection occurs without floods, judge whether that floods occurred, when testing result is that floods do not occur, it need to be detected
It is preceding whether floods to occur, if there is no can detecting new data again；
Once floods occurred, flood information should be prompted to restore, when judging that floods have occurred, then to judge that floods restore；It converges
Total floods warning information, summarizes flood information, and detect new data again.
Further include alarm transmission module in this implementation, is gone out for flood information by network.
Specifically, as shown in figure 3, the workflow of region self-identifying module specifically includes:
301, the image data that need to be analyzed is inputted, obtains video data from video acquisition module, and be input at video preprocessor
Manage module；
302, to image gray processing, gray proces are made to the image inputted from 301, obtain gradation data；
303, binaryzation is carried out to gradation data, binary conversion treatment is carried out to the gradation data obtained from 302；
304, Canny algorithms carry out edge extracting, carry out canny algorithm process to the binaryzation data obtained from 303, obtain
To object edge；
305, HoughLines straight-line detections carry out straight-line detection to the object edge data obtained from 304, detect picture
In whether there is straight line information；
306, color conversion is carried out to image, switchs to RGB from yuv data, the video data obtained from 301 is turned
It changes, obtains RGB data；
307, the image color information of picture is saved, the RGB color information for obtaining each pixel from 306 is saved；
308, road vertical line and other shapes on both sides of the road is excluded, the oblique straight line among picture is only retained, to from 307
Obtained data are classified, other interference informations are removed, and only retain the oblique line of picture middle section, these oblique lines are roads
The information at edge；
309, the straight line that comprehensive detection goes out lifts in-between color information, to the straight line information and 306 obtained from 308
Obtained RGB information, the comprehensive RGB information for obtaining the part among oblique line；
310, detect whether color information in straight line belongs to pavement of road, the color information obtained according to 309 is and known
The color information on road surface is compared, and judges whether it is road surface；
311, judge non-information of road surface, data in new images is detected again, according to 310 as a result, if not detecting
Information of road surface then returns and detects new frame again；
312, statistic mixed-state goes out the number on road surface, if n times is continuously identified, according to 310 as a result, if detection outlet
Whether face, the then number that statistic mixed-state goes out road surface are greater than N；
313, judge in M minutes it is still unidentified go out road, if 312 obtained conclusions be it is no, wait M minutes, such as
Fruit still can not find road information, then needs manual setting parameter；
314, judge automatic detection mark, the result that 312 obtain is judged to need to sentence if obtaining information of road surface
The automatic detection mark for breaking current；
315, it is target area by pavement edge straight line information preservation, 314 result is judged, if it is automatic inspection
Survey state then saves the road surface linear edge information that current detection goes out；
316, determine and do not detect lane information, prompt user's manual setting lane line, according to 313 as a result, if
It does not detect lane information, then prompt user is needed to carry out manual setting lane line；
317, the step of modified logo receives the information of manual setting, connects 316, modified logo is manual setting mode, together
Shi Yaoneng receives the information of manual setting；
318, target area manual setting module receives the parameter of user setting, and parameter is transmitted in algorithm；
319, receive manual setting pavement edge information, the judgement to 314 when indicating for manual setting, receives and comes 318 and set
The target area information set；
320, floods detection unit saves the data in 315, and is arranged into floods detection unit.Ponding inspection
It is as shown in Figure 4 to survey unit:
401, image pre-processing unit obtains video data, and handles it, obtain picture object edge information and
Its color information；
402, monitoring area self-identifying unit supports the identification and calibration of automatic detection monitoring area, supports manual setting
The parameter of monitoring area；
403, water level acquisition module mainly acquires water level information by ultrasonic water level gauge；
404, the color information of marginal information and centre that image pre-processing unit detects is obtained, is obtained from 401 steps
The object edge information and color information for taking it detected；
405, obtain object color component information in monitoring area and region；Obtained from 402 steps need the region that monitors and
Object color component information in the region for needing to monitor；
406, the water level information of water level acquisition module is obtained, obtains water level information from 403 steps；
407, ponding is judged whether there is according to the data of water level acquisition module, water level information, root are obtained from 406 steps
Ponding is judged whether there is according to the height of water level；
408, judge no ponding, when 407 conclusion is no, then judges the information of no ponding need to currently be given without ponding
Ponding summarizing module；
409, it whether there is the sideline of target road in detection data, detect 404 marginal information, if there are in 405
Monitoring area sideline information；
410, ponding non-covering path both sides sideline, 409 conclusion is when being, then it is assumed that current ponding does not cover entirely
Road needs to judge the depth of ponding；
411, whether detection data color information is water stream characteristics, detect 409 conclusion be it is no when, then it is assumed that current ponding
Entire road is covered, at this moment needs to judge whether the ponding region on road with the presence or absence of water stream characteristics, judges its color information
It is consistent with original pavement of road color information；
412, ponding color information is similar to original road surface color information, when 411 result is no, i.e., identifies in picture
Ponding region color information it is still consistent with the color information on original road surface, it is believed that without flowing water feature in current picture, need
Judge depth of accumulated water；
413, water level is judged whether more than Hcm, when the step of 410 and 412, which recognizes no ponding, to be detected again, in root
Judge whether that water level is more than that Hcm is not above if it exceeds being then judged as there is ponding according to water level information, thinks no ponding, then
It will whether there is or not ponding information to give ponding summarizing module；
414, summarize and whether generate seeping phenomenon, summarize 411,413 and 408 judging result, whether judgement finally generates
Seeping phenomenon；
415,414 testing result is transmitted to floods detection unit by floods detection unit, and detection is new again after having handled
Data.
Floods detection unit is as shown in Figure 5:
501, image pre-processing module obtains video data, and handles it, obtain picture object edge information and
Its color information；
502, monitoring area self-identifying module supports the identification and calibration of automatic detection monitoring area, supports manual setting
The parameter of monitoring area；
503, water level acquisition module mainly acquires water level information by ultrasonic water level gauge；
504, water detection unit mainly summarizes 501,502 and 503 information and judges whether to produce ponding and ponding
Situation；
505, judge water detection unit as a result, mainly judge the inspection of 504 ponding with the presence or absence of ponding and ponding situation
Survey the result of unit；
506, judge whether to have generated floods alarm, if generated should be directly entered detect whether generation personnel it is stranded and
In the stranded detection of vehicle；
507, record detects the time T10 of ponding, and the number C1 that statistic mixed-state arrives, and T20 and C2 is emptied, when 505
As a result to record the time T10 detected when detecting ponding, and only record for the first time, records the number C1 detected, simultaneously
Need to empty the value of T20 and C2；
508, judge that the number of C1 is greater than M1 times, and record time T11, T11-T10 is greater than N1 minutes, then judges that water occurs
Calamity, by the duration N1 and number M1 for the ponding being consecutively detected to determine whether generating floods, the value of N1 and M1 can be set；
509, judge whether to have generated floods warning, empty the value of T10 and C1, when 505 testing result is no, it is believed that
It currently without ponding, then needs to judge whether to have occurred and that floods situation, empties the value of T10 and C1, if there is no floods, directly
Return detects new data again；
510, floods situation has such as occurred, has recorded time T20, statistics number C2, has been alerted when judgement had generated floods, and
It currently detects no ponding, then records current time T20, only record the time of first time, record the number C2 detected；
511, judge that the number of C2 is greater than M2 times, and record time T21, T21-T20 is greater than N2 minutes, then judges that floods are extensive
It is multiple, restored by not detecting the duration N2 and number M2 of ponding continuously to determine whether generating floods, the value of N2 and M2 can be set；
512, judge whether ponding region there are personnel to be detained whether situation need to judge floods region after floods have occurred
With the presence of personnel, mainly judge according to whether there is humanoid profile in image after binaryzation；
513, judge whether ponding region has occurred the stranded alarm of personnel, when 512 judgements are detained situation there are personnel, needs
Judge whether current region occurred the stranded alarm of personnel, warning information was directly summarized into mould to alarm if having occurred
Block；
514, the time T30 that record personnel are detained, and the number C3 that statistic mixed-state arrives, empty T40 and C4, when 513 knot
When fruit is that the personnel that detect are detained, the time T30 detected is recorded, and only record for the first time, records the number C3 detected, together
When need to empty the value of T40 and C4；
515, judge that the number of C3 is greater than M3 times, and record time T31, T31-T30 is greater than N3 minutes, then judges that people occurs
The stranded alarm of member, by being consecutively detected the duration N3 and number M3 of personnel's delay to determine whether the personnel of generation are stranded, N3 and
The value of M3 can be set；
516, judge whether that having generated personnel is stranded, and empties the value of T30 and C3, when 512 testing result is no, it is believed that
Currently it is detained without personnel, then needs to judge whether to have occurred and that personnel are stranded situation, empty the value of T30 and C3, if there is no
Personnel are stranded, directly return and detect new data again；
517, time T40, and the number C4 that statistic mixed-state arrives are recorded, personnel are stranded to be alerted when judgement had generated, and was worked as
Before detect that no personnel are detained situation, then record current time T40, only record the time of first time, record and detect
Number C4；
518, judge that the number of C4 is greater than M4 times, and record time T41, T41-T40 is greater than N4 minutes, then judges personnel's quilt
It is tired to release, by duration N4 and number M4 that continuously non-testing staff is detained to determine whether the personnel of generation are stranded restores, N4 and
The value of M4 can be set；
519, judge whether ponding region has vehicle to be detained whether situation need to judge floods region after floods have occurred
With the presence of vehicle, mainly judge according to whether there is vehicle contour in image after binaryzation；
520, judge whether ponding region has occurred the stranded alarm of vehicle, when 519 judgements are detained situation there are vehicle, needs
Judge whether current region occurred the stranded alarm of vehicle, warning information was directly summarized into mould to alarm if having occurred
Block；
521, the time T50 that record vehicle is detained, and the number C5 that statistic mixed-state arrives, empty T60 and C6, when 520 knot
Fruit is when detecting that vehicle is detained, to record the time T50 detected, and only record for the first time, records the number C5 detected, together
When need to empty the value of T60 and C6；
522, judge that the number of C5 is greater than M5 times, and record time T51, T51-T50 is greater than N5 minutes, then judges that vehicle occurs
Stranded alarm, by being consecutively detected the duration N5 and number M5 of vehicle delay to determine whether to generate vehicle stranded, N5 and
The value of M5 can be set；
523, judge whether that having generated vehicle is stranded, and empties the value of T50 and C5, when 519 testing result is no, it is believed that
Currently it is detained without vehicle, then needs to judge whether to have occurred and that vehicle is stranded situation, empty the value of T50 and C5, if there is no
Vehicle is stranded, directly returns and detects new data again；
524, time T60, and the number C6 that statistic mixed-state arrives are recorded, vehicle is stranded to be alerted when judgement had generated, and was worked as
Before detect that no vehicle is detained situation, then record current time T60, only record the time of first time, record and detect
Number C6；
525, judge that the number of C6 is greater than M6 times, and record time T61, T61-T60 is greater than N6 minutes, then judges vehicle quilt
It is tired to release, by not detecting the duration N6 and number M6 of vehicle delay continuously to determine whether generating that vehicle is stranded to be restored, N6 and
The value of M6 can be set；
526, floods are integrated, personnel are stranded, and vehicle is stranded, the alerts such as water level, and comprehensive 515,518,522,525,
513,520,511 as a result, judging whether that floods situation occurs, floods restore, and personnel are stranded, and personnel are stranded to be released, vehicle quilt
It is tired, the stranded releasing of vehicle and water level conditions；
527, the flood information that will test is sent to transmission module, and detects new data again.Preferably, further including
Video encoding module and video transmission module, the video encoding module are used to carry out compressed encoding processing, the view to video
Frequency transmission module is for sending coding module rear video data by network.
Fig. 6 shows a kind of floods detection method based on video, comprising the following steps:
S1, acquisition video data, and water level information is acquired in real time；
S2, machine vision processing is carried out to the image of video data, mainly includes gray proces, binary conversion treatment, edge
Detection, color conversion；
S3, the road edge information within the scope of automatic identification picture is carried out using DRR algorithm, and detection range is set；
S4, it is analyzed in setting range using DWD algorithm with the presence or absence of ponding, ponding range, the color of ponding；
S5, analyzed using DFD algorithm, whether generate floods according in the interpretation of result floods of step S4, if exist by
Oppressive member or vehicle, and if so, generating warning information；
S6, data and warning information are sent by network, mainly includes that floods generate, floods are oppressive, floods are tired
The warning information of vehicle.
Preferably, the step S2 is specifically included:
S21, gray proces are done to the video frame images in video data, obtains gradation data；
S22, binary conversion treatment is carried out to gradation data；
S23, Canny algorithm process is carried out to the data after binary conversion treatment, obtains the object edge in image；
S24, HoughLines straight-line detection is carried out to obtained object edge data, detects and whether there is straight line in picture
Information；
S25, color conversion is carried out to image, switchs to RGB from yuv data, saves the image color information of picture, obtained every
The RGB color multimedia message of one pixel.
Preferably, the step S3 is specifically included:
S31, classify to obtained data, remove other interference informations, only retain the oblique line of picture middle section, this
A little oblique lines are the information at the edge of road；
S32, in-between color information, the comprehensive RGB information for obtaining the part among oblique line, with known road surface are extracted
Color information be compared, judge whether it is road surface, if not detecting information of road surface, it is new to return to detection again
Video frame；
S33, if it is detected that road surface, statistics continuously detects whether the number on road surface is greater than N；If the then automatic detection of judgement
Mark, unidentified road out if not and in M minute, it is determined that do not detect lane information, need manual setting lane again
Line, it is no for modifying automatic detection mark；
S34, the automatic detection mark of judgement, if being then target monitoring region by pavement edge straight line information preservation, if not
Then manual setting pavement edge information and save as target monitoring region.
Preferably, the step S4 is specifically included:
S41, monitoring area water level information is obtained, ponding is judged whether there is according to the height of water level；It will if no ponding
No ponding information summarizes, and monitoring area is detected if having ponding with the presence or absence of marginal information
If S42, ponding do not cover sideline on both sides of the road, current ponding does not cover entire road, needs to judge to accumulate
The depth of water；If ponding covers entire road, need to judge the ponding region on road with the presence or absence of water stream characteristics, judgement
Whether its color information is consistent with original pavement of road color information；
If the color information of the ponding region identified in S43, picture is still consistent with the color information on original road surface, it is believed that
Without flowing water feature in current picture, depth of accumulated water need to be judged；Judge whether that water level is more than Hcm according to water level information, if it exceeds
Then it is judged as there is ponding, is not above, thinks no ponding, then will summarize whether there is or not ponding information.
Preferably, the step S5 is specifically included:
If S51, detecting ponding, the time T10 detected is recorded, and only record for the first time, records the number detected
C1；It continues to test, if detecting, the number of C1 is greater than M1 times, time T11 at this time is recorded, if T11-T10 is greater than N1 minutes,
Floods occur for judgement；If not detecting ponding, need to judge whether that floods had occurred；
If floods S52, have occurred, the time T20 of ponding is not detected in record, and the number of ponding is not detected in statistics
C2 is continued to test, if detecting, the number of C2 is greater than M2 times, is recorded time T21 at this time and is sentenced when T21-T20 is greater than N2 minutes
Calamity of cutting off the water supply is restored；
If floods S53, have occurred, judge floods region whether with the presence of personnel, the stranded warning of personnel whether has occurred,
And it judges whether there is vehicle and is detained situation, the stranded alarm of vehicle whether has occurred；
S54, according in image after binaryzation whether there is humanoid profile to determine whether have personnel be detained situation, detect
When personnel are detained, the time T30 detected is recorded, and only record for the first time, records the number C3 detected, continues to record, if
It detects that the number of C3 is greater than M3 times, records time T31, if T31-T30 is greater than N3 minutes, judges that generation personnel are stranded and accuse
It is alert；When judgement had generated, personnel are stranded to be alerted, and currently detected that no personnel are detained situation, then record current time
T40 only records the time of first time, records the number C4 detected；It continues to test, if the number of C4 is greater than M4 times, when record
Between T41, T41-T40 is greater than N4 minute, then judges the stranded releasing of personnel；
S55, judge whether ponding region has vehicle to be detained feelings according to whether there is vehicle contour in image after binaryzation
Condition；Judge whether current region occurred the stranded alarm of vehicle, warning information was summarized if having occurred；Detect vehicle
When delay, the time T50 detected is recorded, and only record for the first time, records the number C5 detected, continues to test, time of C5
Number is greater than M5 times, and records time T51, and T51-T50 is greater than N5 minutes, then judges that the stranded alarm of vehicle occurs；Detecting does not have
Vehicle is detained situation, then records current time T60, only records the time of first time, records the number C6 detected, continue to test
Number to C6 is greater than M6 times, and records time T61, and T61-T60 is greater than N6 minutes, then judges that vehicle is stranded and release.
The embodiments described above only express several embodiments of the present invention, and the description thereof is more specific and detailed, but simultaneously
Limitations on the scope of the patent of the present invention therefore cannot be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to guarantor of the invention
Protect range.Therefore, the scope of protection of the patent of the invention shall be subject to the appended claims.
Claims (8)
1. a kind of floods detection system based on video, which is characterized in that including water level acquisition module, video acquisition module and figure
As analysis module；The video acquisition module is used for collection site video data, and sends image analysis mould for video data
Block；Described image analysis module be used for according to DRR algorithm, DWD algorithm and DFD algorithm to the video frame in video data at
Reason, is analyzed and processed each video frame, according to the marginal information and color information of object in each video frame, judges road surface
With the presence or absence of ponding, and judge whether to generate flood information；The water level acquisition module is for acquiring water level information；
Described image analysis module is used to carry out the road edge information within the scope of automatic identification picture using DRR algorithm, and sets
Set detection range；
Described image analysis module is used to whether there is ponding, ponding range, ponding using in DWD algorithm analysis setting range
Color, specific steps include:
S41, monitoring area water level information is obtained, ponding is judged whether there is according to the height of water level；It will be without product if without ponding
Water information summarizes, and monitoring area is detected if having ponding with the presence or absence of marginal information；
If S42, ponding do not cover sideline on both sides of the road, current ponding does not cover entire road, needs to judge ponding
Depth；If ponding covers entire road, need to judge that the ponding region on road with the presence or absence of water stream characteristics, judges its color
Whether multimedia message is consistent with original pavement of road color information；
If the color information of the ponding region identified in S43, picture is still consistent with the color information on original road surface, it is believed that current
Without flowing water feature in picture, depth of accumulated water need to be judged；Judge whether that water level is more than Hcm according to water level information, if it exceeds then sentencing
Break to there is ponding, be not above, thinks no ponding, then will summarize whether there is or not ponding information；
Described image analysis module is used to analyze using DFD algorithm, whether generates water according to the interpretation of result that DWD algorithm is analyzed
Calamity, if there are trapped person or vehicles, and if so, generating warning information, specific steps include:
If S51, detecting ponding, the time T10 detected is recorded, and only record for the first time, records the number C1 detected；After
Continuous detection, if detecting, the number of C1 is greater than M1 times, records time T11 at this time, if T11-T10 is greater than N1 minutes, judges
Floods occur；If not detecting ponding, need to judge whether that floods had occurred；
If floods S52, have occurred, the time T20 of ponding is not detected in record, and the number C2 of ponding is not detected in statistics, after
Continuous detection, if detecting, the number of C2 is greater than M2 times, records time T21 at this time and judges water when T21-T20 is greater than N2 minutes
Calamity is restored；
If floods S53, have occurred, judge floods region whether with the presence of personnel, the stranded warning of personnel whether has occurred, and sentence
It is disconnected to be detained situation with the presence or absence of vehicle, the stranded alarm of vehicle whether has occurred；
S54, according in image after binaryzation whether there is humanoid profile to determine whether have personnel be detained situation, detect personnel
When delay, the time T30 detected is recorded, and only record for the first time, records the number C3 detected, continues to record, if detection
Number to C3 is greater than M3 times, records time T31, if T31-T30 is greater than N3 minutes, judges the stranded alarm of generation personnel；When
Judgement had generated the stranded alarm of personnel, and currently detected that no personnel are detained situation, then record current time T40, only
The time for recording first time, record the number C4 detected；It continues to test, if the number of C4 is greater than M4 times, records time T41,
T41-T40 is greater than N4 minutes, then judges that personnel are stranded and release；
S55, judge whether ponding region has vehicle to be detained situation according to whether there is vehicle contour in image after binaryzation；Sentence
Whether disconnected current region occurred the stranded alarm of vehicle, summarized if having occurred to warning information；Detect that vehicle is detained
When, the time T50 detected is recorded, and only record for the first time, records the number C5 detected, continues to test, the number of C5 is big
In M5 times, and time T51 is recorded, T51-T50 is greater than N5 minutes, then judges that the stranded alarm of vehicle occurs；Detect no vehicle
It is detained situation, then records current time T60, only records the time of first time, record the number C6 detected, continue to test C6
Number be greater than M6 time, and record time T61, T61-T60 be greater than N6 minute, then judge vehicle be stranded releasing.
2. the floods detection system according to claim 1 based on video, which is characterized in that the video acquisition module packet
Include camera lens, video sensor and video processor；The camera lens is used for collection site video image information；The video sensor
For video pictures to be converted to digital information, the video processor is used to receive and process the video of video sensor acquisition
Data.
3. the floods detection system according to claim 1 based on video, which is characterized in that the water level acquisition module packet
Ultrasonic water level gauge is included, for acquiring water level information.
4. the floods detection system according to claim 1 based on video, which is characterized in that described image analysis module packet
Include image pre-processing unit, monitoring area self-identifying unit, water detection unit and floods detection unit；
Described image pretreatment unit handles each video frame, obtains video frame picture for obtaining video data
Middle object edge information and color information；
The monitoring area self-identifying unit is used for according to marginal information and color information automatic identification road edge information, automatically
Set the related objective information of detection zone；
The water detection unit is for detecting ponding information；
The floods detection unit is used to detect whether that floods occur, if having personnel and vehicle according to video and ponding information
Trapped information.
5. the floods detection system according to claim 1 based on video, which is characterized in that further include video encoding module
And video transmission module, the video encoding module are used to carry out video compressed encoding processing, the video transmission module is used
In coding module rear video data are sent by network.
6. a kind of floods detection method based on video, which comprises the following steps:
S1, acquisition video data, and water level information is acquired in real time；
S2, machine vision processing is carried out to the image of video data, mainly includes gray proces, binary conversion treatment, edge detection,
Color conversion；
S3, the road edge information within the scope of automatic identification picture is carried out using DRR algorithm, and detection range is set；
S4, it is analyzed in setting range using DWD algorithm with the presence or absence of ponding, ponding range, the color of ponding；
The step S4 is specifically included:
S41, monitoring area water level information is obtained, ponding is judged whether there is according to the height of water level；It will be without product if without ponding
Water information summarizes, and monitoring area is detected if having ponding with the presence or absence of marginal information；
If S42, ponding do not cover sideline on both sides of the road, current ponding does not cover entire road, needs to judge ponding
Depth；If ponding covers entire road, need to judge that the ponding region on road with the presence or absence of water stream characteristics, judges its color
Whether multimedia message is consistent with original pavement of road color information；
If the color information of the ponding region identified in S43, picture is still consistent with the color information on original road surface, it is believed that current
Without flowing water feature in picture, depth of accumulated water need to be judged；Judge whether that water level is more than Hcm according to water level information, if it exceeds then sentencing
Break to there is ponding, be not above, thinks no ponding, then will summarize whether there is or not ponding information；
S5, it is analyzed using DFD algorithm, whether floods is generated according to the interpretation of result of step S4, if there are trapped person or vehicles
, and if so, generating warning information；
The step S5 is specifically included:
If S51, detecting ponding, the time T10 detected is recorded, and only record for the first time, records the number C1 detected；After
Continuous detection, if detecting, the number of C1 is greater than M1 times, records time T11 at this time, if T11-T10 is greater than N1 minutes, judges
Floods occur；If not detecting ponding, need to judge whether that floods had occurred；
If floods S52, have occurred, the time T20 of ponding is not detected in record, and the number C2 of ponding is not detected in statistics, after
Continuous detection, if detecting, the number of C2 is greater than M2 times, records time T21 at this time and judges water when T21-T20 is greater than N2 minutes
Calamity is restored；
If floods S53, have occurred, judge floods region whether with the presence of personnel, the stranded warning of personnel whether has occurred, and sentence
It is disconnected to be detained situation with the presence or absence of vehicle, the stranded alarm of vehicle whether has occurred；
S54, according in image after binaryzation whether there is humanoid profile to determine whether have personnel be detained situation, detect personnel
When delay, the time T30 detected is recorded, and only record for the first time, records the number C3 detected, continues to record, if detection
Number to C3 is greater than M3 times, records time T31, if T31-T30 is greater than N3 minutes, judges the stranded alarm of generation personnel；When
Judgement had generated the stranded alarm of personnel, and currently detected that no personnel are detained situation, then record current time T40, only
The time for recording first time, record the number C4 detected；It continues to test, if the number of C4 is greater than M4 times, records time T41,
T41-T40 is greater than N4 minutes, then judges that personnel are stranded and release；
S55, judge whether ponding region has vehicle to be detained situation according to whether there is vehicle contour in image after binaryzation；Sentence
Whether disconnected current region occurred the stranded alarm of vehicle, summarized if having occurred to warning information；Detect that vehicle is detained
When, the time T50 detected is recorded, and only record for the first time, records the number C5 detected, continues to test, the number of C5 is big
In M5 times, and time T51 is recorded, T51-T50 is greater than N5 minutes, then judges that the stranded alarm of vehicle occurs；Detect no vehicle
It is detained situation, then records current time T60, only records the time of first time, record the number C6 detected, continue to test C6
Number be greater than M6 time, and record time T61, T61-T60 be greater than N6 minute, then judge vehicle be stranded releasing；
S6, data and warning information are sent by network, it is main oppressive, the tired vehicle of floods including floods generation, floods
Warning information.
7. the floods detection method according to claim 6 based on video, which is characterized in that the step S2 is specifically wrapped
It includes:
S21, gray proces are done to the video frame images in video data, obtains gradation data；
S22, binary conversion treatment is carried out to gradation data；
S23, Canny algorithm process is carried out to the data after binary conversion treatment, obtains the object edge in image；
S24, HoughLines straight-line detection is carried out to obtained object edge data, detects in picture and believes with the presence or absence of straight line
Breath；
S25, color conversion is carried out to image, switchs to RGB from yuv data, saves the image color information of picture, obtain each
The RGB color multimedia message of pixel.
8. the floods detection method according to claim 7 based on video, which is characterized in that the step S3 is specifically wrapped
It includes:
S31, classify to obtained data, remove other interference informations, only retain the oblique line of picture middle section, these are tiltedly
Line is the information at the edge of road；
S32, in-between color information, the comprehensive RGB information for obtaining the part among oblique line, the color with known road surface are extracted
Multimedia message is compared, and judges whether it is road surface, if not detecting information of road surface, returns and detect new video again
Frame；
S33, if it is detected that road surface, statistics continuously detects whether the number on road surface is greater than N；If the then automatic detection mark of judgement
Will, if not and in M minutes it is unidentified go out road, it is determined that do not detect lane information, need manual setting lane line, modify
Automatic detection mark is no；
S34, the automatic detection mark of judgement, if being then target monitoring region by pavement edge straight line information preservation, if otherwise hand
Dynamic setting pavement edge information simultaneously saves as target monitoring region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610686278.8A CN106331605B (en) | 2016-08-17 | 2016-08-17 | A kind of floods detection system and method based on video |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610686278.8A CN106331605B (en) | 2016-08-17 | 2016-08-17 | A kind of floods detection system and method based on video |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106331605A CN106331605A (en) | 2017-01-11 |
| CN106331605B true CN106331605B (en) | 2019-11-15 |
Family
ID=57744289
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610686278.8A Active CN106331605B (en) | 2016-08-17 | 2016-08-17 | A kind of floods detection system and method based on video |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106331605B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107657822A (en) * | 2017-11-17 | 2018-02-02 | 嘉兴四维智城信息科技有限公司 | Highway moisture film early warning system and method for traffic video automatic identification |
| CN107909070A (en) * | 2017-11-24 | 2018-04-13 | 天津英田视讯科技有限公司 | A kind of method of road water detection |
| CN109377725B (en) * | 2018-12-06 | 2020-07-21 | 合肥海诺恒信息科技有限公司 | System for emergency response and scheduling of urban road waterlogging |
| CN110491088A (en) * | 2019-07-26 | 2019-11-22 | 安徽泛米科技有限公司 | A kind of area stay warning device |
| CN110956783B (en) * | 2019-12-04 | 2021-11-16 | 江河瑞通（北京）技术有限公司 | Urban waterlogging monitoring method and electronic equipment |
| CN111726582B (en) * | 2020-06-22 | 2021-12-14 | 广东技安科技有限公司 | Waterproof outdoor monitoring system and method applying numerical analysis |
| CN117218801A (en) * | 2023-10-23 | 2023-12-12 | 华北科技学院(中国煤矿安全技术培训中心) | Urban flood disaster monitoring and early warning method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003284051A (en) * | 2002-03-26 | 2003-10-03 | Fujitsu Ltd | Video transmission system and method for controlling video transmission |
| KR20140024945A (en) * | 2014-02-11 | 2014-03-03 | 주식회사 스마트에어 챔버 | 4 season and real-time and scene image relay system and image relay method by using portable working vehicle |
| CN104180869A (en) * | 2014-09-05 | 2014-12-03 | 奥璞特智能科技（上海）有限公司 | Road surface water on-line monitoring and early-warning method and system |
| CN104727423A (en) * | 2015-02-06 | 2015-06-24 | 天津大学 | Intelligent dispatching system and method for urban inland inundation water drainage based on traffic videography system |
| CN105389561A (en) * | 2015-11-13 | 2016-03-09 | 深圳华中科技大学研究院 | Method for detecting bus lane based on video |
-
2016
- 2016-08-17 CN CN201610686278.8A patent/CN106331605B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003284051A (en) * | 2002-03-26 | 2003-10-03 | Fujitsu Ltd | Video transmission system and method for controlling video transmission |
| KR20140024945A (en) * | 2014-02-11 | 2014-03-03 | 주식회사 스마트에어 챔버 | 4 season and real-time and scene image relay system and image relay method by using portable working vehicle |
| CN104180869A (en) * | 2014-09-05 | 2014-12-03 | 奥璞特智能科技（上海）有限公司 | Road surface water on-line monitoring and early-warning method and system |
| CN104727423A (en) * | 2015-02-06 | 2015-06-24 | 天津大学 | Intelligent dispatching system and method for urban inland inundation water drainage based on traffic videography system |
| CN105389561A (en) * | 2015-11-13 | 2016-03-09 | 深圳华中科技大学研究院 | Method for detecting bus lane based on video |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106331605A (en) | 2017-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106331605B (en) | A kind of floods detection system and method based on video | |
| CN101639894B (en) | Method for detecting train driver behavior and fatigue state on line and detection system thereof | |
| CN101655907B (en) | Trainman driving state monitoring intelligent alarm system | |
| CN101593425B (en) | Machine vision based fatigue driving monitoring method and system | |
| CN102521578B (en) | Method for detecting and identifying intrusion | |
| CN111091098B (en) | Training method of detection model, detection method and related device | |
| CN108474866A (en) | A kind of manhole cover loss detecting system and method based on deep learning | |
| JP7407198B2 (en) | Driving monitoring methods, systems and electronic equipment | |
| CN109040693B (en) | Intelligent alarm system and method | |
| CN103065121B (en) | The engine driver's method for monitoring state analyzed based on video human face and device | |
| CN113052029A (en) | Abnormal behavior supervision method and device based on action recognition and storage medium | |
| CN110287906A (en) | Method and system based on image/video detection people " playing mobile phone " | |
| US8301577B2 (en) | Intelligent monitoring system for establishing reliable background information in a complex image environment | |
| CN102164270A (en) | Intelligent video monitoring method and system capable of exploring abnormal events | |
| CN113642474A (en) | Hazardous area personnel monitoring method based on YOLOV5 | |
| CN103049740A (en) | Method and device for detecting fatigue state based on video image | |
| CN112819068A (en) | Deep learning-based real-time detection method for ship operation violation behaviors | |
| CN101803928A (en) | Video-based driver fatigue detection device | |
| CN107330414A (en) | Act of violence monitoring method | |
| CN111401310B (en) | Kitchen sanitation safety supervision and management method based on artificial intelligence | |
| EP2000998B1 (en) | Flame detecting method and device | |
| CN104102921A (en) | Video detection method and video detection device based on face recognition | |
| CN112183219A (en) | Public safety video monitoring method and system based on face recognition | |
| CN115205581A (en) | Fishing detection method, fishing detection device and computer readable storage medium | |
| CN110728212A (en) | Road well lid monitoring device and method based on computer vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |