CN104392630A - Throw-out intelligent detection device and method - Google Patents
Throw-out intelligent detection device and method Download PDFInfo
- Publication number
- CN104392630A CN104392630A CN201410688800.7A CN201410688800A CN104392630A CN 104392630 A CN104392630 A CN 104392630A CN 201410688800 A CN201410688800 A CN 201410688800A CN 104392630 A CN104392630 A CN 104392630A
- Authority
- CN
- China
- Prior art keywords
- information
- foreground
- intelligent detection
- processing chip
- color camera
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 62
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000012544 monitoring process Methods 0.000 claims abstract description 19
- 230000008569 process Effects 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims description 62
- 239000011521 glass Substances 0.000 claims description 14
- 230000035945 sensitivity Effects 0.000 claims description 9
- 230000005484 gravity Effects 0.000 claims description 7
- 229910001220 stainless steel Inorganic materials 0.000 claims description 7
- 239000010935 stainless steel Substances 0.000 claims description 7
- 235000013399 edible fruits Nutrition 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 239000013307 optical fiber Substances 0.000 claims description 3
- 238000004806 packaging method and process Methods 0.000 claims 1
- 206010039203 Road traffic accident Diseases 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000001681 protective effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000005538 encapsulation Methods 0.000 description 3
- 238000003909 pattern recognition Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
Disclosed is a throw-out intelligent detection device and method. A throw-out intelligent detection device and method comprises an intelligent detection module, a display module, and a warning module. The intelligent detection module monitors and collects traffic conditions of tunnels and roads in real time, and transmits the traffic conditions to the display module, meanwhile, an internally embedded intelligent chip processor processes and judges a video image signal information through a background modeling step, a foreground detection step, a protected zone setting step, a foreground matching and updating step, and a throw-out judging step; if the throw-out exists, the display module displays the information of the throw-out in real time and sends a warning signal to the warning module, so as to remind the monitoring personnel to deal with the throw-out, thereby avoiding an accident. The throw-out intelligent detection device and method is not only able to intelligently detect the information of the throw-out, but the detection method is also able to avoid underreporting and false alarm at a certain degree, thereby providing an effective and executable solution for ensuring the traffic safety and the traffic order.
Description
Technical Field
The invention belongs to the field of intelligent video analysis and monitoring, and relates to technologies of image processing, video analysis, pattern recognition, intelligent monitoring, artificial intelligence and the like.
Background
With the development of economy in China, the transportation industry is developed increasingly, so that the traffic problem is more severe, traffic accidents frequently occur, and the national and people pay attention to the traffic accidents deeply. Especially on tunnels and highways, the traffic flow is large, the speed of the vehicle is high, and sometimes objects fall off from the vehicle or the vehicle with faults cannot be timely processed and moved, so that traffic accidents are caused. The object throwing event becomes a frequent traffic event, so that not only are a plurality of vehicles affected by a primary accident, but also secondary accidents are caused, the life and property safety of people is seriously damaged, and the loss which cannot be estimated is caused sometimes, so that the caused traffic accident and the caused potential safety hazard become problems which need to be solved urgently. Therefore, how to quickly and accurately detect the sprinkled objects on the tunnel and the highway and give an alarm, find potential safety hazards as early as possible and remove the potential safety hazards in time, and keeping the tunnel, the highway and the like safe and smooth becomes an important problem in the field of traffic security and protection. The intelligent detection device for the road and tunnel sprinkles can work continuously for 24 hours all day, supervises the safety of the tunnel road all the time, protects the safety of lives and properties of people and has great social benefit. At present, the development of China in the field of intelligent detection of spilled objects on roads and tunnels is immature, detection equipment and technology have a plurality of defects, such as poor real-time performance, false alarm and missing report of spilled objects, poor environmental adaptability and the like.
Disclosure of Invention
In order to solve the above problems, the present invention provides an intelligent projectile detection device and method. The device has the advantages of high processing speed, good real-time performance, good control on false alarm and missed alarm phenomena, high detection precision and good adaptability to emergency situations in the environment.
The technical scheme adopted by the invention is as follows:
the intelligent detection device and method for the sprinkled objects comprise an intelligent detection module, a display module and an alarm module.
The intelligent detection module mainly comprises a color camera, a protection device, an infrared illuminating lamp and a DSP processing chip. The color camera collects color video images, and is connected with the DSP processing chip through a network cable to transmit data. The color camera transmits the image to the DSP processing chip, the DSP processing chip is mainly an algorithm unit and provides an intelligent detection scheme for the sprinkled objects, and the color camera and the DSP processing chip are packaged in the shell. The shell is by the cylindrical cavity encapsulation shell that stainless steel 316L made, and transparent glass is installed to the inside front end of shell, and transparent glass is used for protecting color camera and DSP and handles the chip, and color camera installs in the one side that is close to front end transparent glass, and DSP handles the chip and installs in the one side of keeping away from front end transparent glass, and color camera can see through front end glass simultaneously and gather the condition image all around. The rear end of the shell is provided with two wiring holes, one hole is used for connecting a power line to supply power to the color camera and the DSP processing chip, and the other hole is used for connecting an optical fiber cable and transmitting a video image signal acquired by the color camera and an alarm signal obtained by processing of the DSP processing chip to a monitoring room which is hundreds of meters or kilometers away. The protection device is a rain shade, is in an arc sheet shape, is installed on the shell of the intelligent detection module, can shade rain, and prolongs the service life of the device. The infrared illuminating lamp provides night illumination, the detection distance is 150 meters, and the color camera can normally acquire images under the condition of no background light source. The infrared illuminating lamp is packaged in the cylindrical stainless steel protective shell, two shells are rigidly connected under the shell of the intelligent detection module, and the power supply line is connected to the opening at the tail part of the stainless steel protective shell and supplies power to the infrared lamp.
The display module is a display screen, is arranged in the monitoring room, receives the video information transmitted by the color camera and displays the video information in real time. And meanwhile, the throwing object information obtained by the processing of the DSP chip is received and displayed.
The alarm module is an alarm and is used for receiving the alarm signal of the sprinkled object obtained by the processing of the DSP processing chip and prompting monitoring personnel to process the sprinkled objects on the tunnel and the highway in time.
The intelligent detection device and method for the sprinkled objects comprise the following monitoring steps:
(1) the color image information collected by the color camera is transmitted to the display module and the DSP processing chip simultaneously, the display module displays the tunnel and road condition information in real time, and the DSP processing chip receives the color video image information transmitted by the color camera for further intelligent identification processing.
(2) And the DSP processing chip carries out background modeling, foreground detection, protected area setting, foreground matching updating and sprinkle judgment processing procedures on the received color video image information to obtain the sprinkle information.
(3) The DSP processing chip transmits the processing result to the alarm module, and transmits the object throwing information to the display module, so that the object throwing information is displayed, and monitoring personnel are prompted to perform next processing.
The intelligent detection method for the sprinkled objects adopted by the device mainly comprises background modeling, foreground detection, setting of a protection area, foreground matching updating and sprinkled object judgment.
The background modeling is mainly to perform pattern recognition according to color video images input by a video acquisition module and intelligently establish a self-adaptive learning background model. And establishing a background model by using the brightness and the chrominance information of the video image. Because the algorithm performs operation in the embedded equipment, all data information adopted by the algorithm is integer parameters. And sequentially traversing pixel points in the image, searching eight neighborhood pixel points for each pixel point, and establishing a Gaussian model according to the seed points and the eight neighborhood information, wherein parameters of the Gaussian model comprise brightness, chroma, saturation, weight and variance information. A gaussian model is built for each seed point.
The brightness and the chroma of the model are obtained through the seed points and the eight neighborhood related information of the seed points. Such as a matrixThe positions are shown, the middle 1 is seed point pixel information, and the eight neighborhoods are represented by 0. Marking the eight neighborhood pixel point information as P from left to right and from top to bottom in sequence1，P2，P3，P4，P5，P6，P7，P8 Seed point is marked as P0. Establishing a Gaussian model for the seed points according to the weight parameters,
wherein,in order to establish the brightness after the gaussian model,、、、、、、、、respectively the seed point and the brightness of its eight neighborhoods.
In the same way, the Gaussian model of the seed point chroma and saturation is established by the same method,
wherein,in order to establish the chromaticity after the gaussian model,、、、、、、、、respectively, the seed point and its eight neighborhood.
Wherein,in order to establish the saturation after the gaussian model is established,、、、、、、、、the seed point and its saturation of the eight neighborhoods, respectively.
The model variance learning process is as follows:
wherein,represents the result of the variance learning of the image of the new frame,the variance of the current background is represented as,a model variance learning factor is represented as a function of the model variance,、、respectively representing the brightness, the chroma and the saturation information of the Gaussian model at the seed point of the background image.、、Respectively representing the brightness, the chroma and the saturation information of the Gaussian model at the seed point at the corresponding position of the current frame image and the background image.
The model weight records the number of successful modeling of the background seed points. When the total number of background seed point modeling reaches 1/3 of the total number of seed points, the background modeling is successful.
The foreground detection is to detect a new person or object in the background. The foreground detection method adopted by the invention is a method for comparing the Gaussian models of the foreground and the background. And similarly, establishing the brightness, the chroma, the saturation, the weight and the variance information of the Gaussian model parameters of the pixel seed points of the current frame image. And respectively carrying out difference on brightness, chroma, saturation and variance parameters corresponding to the Gaussian models of the foreground and the background. And if the difference value is smaller than a certain threshold value, the seed point is a suspected foreground point. Otherwise, it is a background point. And after comparing all the seed points of the foreground points, carrying out communicated expansion processing on the seed points, so that all the seed points of the same foreground object are communicated, calculating the size, the perimeter, the area, the gravity center and the height-width ratio information of all the communicated areas, eliminating the influence of small false foreground interference points, and then respectively marking masks of different foregrounds with different serial numbers. And storing effective foreground data in the first frame of image after background learning is finished into a foreground historical information base so as to facilitate operations such as foreground matching updating and projectile judgment in the subsequent detection process.
The protection area is set as an area where vehicles run on a tunnel or a road. The protection area information is marked by drawing two lines along the edge of the tunnel or road in the image. And meanwhile, the alarm sensitivity of the throwing object, the monitoring limit of the area of the throwing object and the limit of the height-width ratio are set, so that the influence of too small stones, fruit stones and the like is prevented.
The foreground matching updating process mainly refers to the steps of using newly detected foreground information to match and update a foreground historical information base. The method comprises the steps of detecting emerging foreground information of a current frame image, comparing foreground historical data according to foreground size, perimeter, area, gravity center and aspect ratio information, if the change is smaller than a threshold value, determining that the foreground historical data belong to the same foreground information, updating the foreground historical data, and if no matched historical information is found, storing the emerging foreground into the historical data.
The projectile event judgment mainly judges whether the foreground is the projectile or not, whether the projectile is in a protection area or not and whether the alarm sensitivity meets the requirements or not. The judgment process is to traverse the foreground historical information base, judge whether the area and the height-width ratio of the foreground meet the threshold setting, and if so, judge that the object is a projectile. Whether the protection area is set is further judged according to the foreground gravity center information, whether the alarm sensitivity meets the requirement is judged if the protection area is set, and if the alarm sensitivity meets the requirement, the object is thrown, alarm information is sent out simultaneously, and the foreground information is marked and displayed on a screen.
The invention has the beneficial effects that: the device can work for 24 hours without interruption, is simple and flexible, is suitable for on-site variable factors, can effectively identify the objects thrown in the road tunnel, can give an alarm in time, avoids traffic accidents, protects personal and property safety, and is beneficial to maintaining traffic safety and order. The intelligent detection device for the sprinkled objects is simple and easy to install and convenient to operate. The intelligent detection algorithm for the sprinkled objects has the advantages of intelligent learning and updating strategies for the foreground and the background, strong adaptability and higher intelligent degree.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the example serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a schematic view of the apparatus of the present invention.
Fig. 2 is a flow chart of the projectile detection algorithm of the present invention.
Fig. 3 is an illustration of the projectile alarm condition of the present invention.
Detailed Description
The following describes various details of the technical solution of the present invention with reference to specific examples. It should be noted that the described examples are only intended to facilitate the understanding of the invention and do not therefore limit the scope of protection of the invention.
Fig. 1 shows a diagram of the intelligent projectile detection device, which comprises an intelligent detection module 9, a display module 7 and an alarm module 8.
The intelligent detection module 9 mainly comprises a color camera 2, a protection device 1, an infrared illuminating lamp 6 and a DSP processing chip 4. The color camera 2 collects color video images, and the color camera 2 is connected with the DSP processing chip 4 through a network cable to transmit data. The color camera 2 transmits images to the DSP processing chip 4, the DSP processing chip 4 is mainly an algorithm unit and provides an intelligent detection scheme for the sprinkled objects, and the color camera 2 and the DSP processing chip 4 are packaged in the shell 3. The cylindrical cavity encapsulation shell that shell 3 was made by stainless steel 316L, and transparent glass is installed to the inside front end of shell 3, and transparent glass is used for protecting color camera 2 and DSP processing chip 4, and color camera 2 installs in the one side that is close to front end transparent glass, and DSP processing chip 4 installs in the one side of keeping away from front end transparent glass, and color camera 2 can see through front end glass simultaneously and gather the condition image all around. The rear end of the shell 3 is provided with two wiring holes, one hole is used for connecting a power line to supply power to the color camera 2 and the DSP processing chip 4, and the other hole is used for connecting an optical fiber cable to transmit video image signals collected by the color camera 2 and alarm signals obtained by processing of the DSP processing chip 4 to a monitoring room with the length of hundreds of meters or kilometers away. Protection device 1 is for keeping out the rain cover, for the arc sheet form, installs on intelligent detection module's shell 3, can keep out rain, extension device life age. The infrared illuminating lamp 6 provides night illumination, the detection distance is 150 meters, and the color camera 2 can normally acquire images under the condition of no background light source. The encapsulation of infrared light 6 is in cylindrical stainless steel protective housing 5, under intelligent detection module 9's shell 3, two shells 5, 3 rigid connection, and the power cord is connected for the infrared lamp power supply in the 5 afterbody trompils of stainless steel protective housing.
The display module is a display screen 7, is arranged in the monitoring room, and receives and displays the video information transmitted by the color camera 2 in real time. And meanwhile, the throwing object information obtained by the processing of the DSP chip 4 is received and displayed.
The alarm module is an alarm 8 and is used for receiving the projectile alarm signal obtained by the DSP processing chip 4 and prompting monitoring personnel to process projectiles on the tunnel and the highway in time.
The intelligent detection device and method for the sprinkled objects comprise the following monitoring steps:
(4) the color image information collected by the color camera 2 is transmitted to the display module 7 and the DSP processing chip 4 at the same time, the display module 7 displays the tunnel and road condition information in real time, and the DSP processing chip 4 receives the color video image information transmitted by the color camera for further intelligent identification processing.
(5) The DSP processing chip 4 carries out background modeling, foreground detection, protected area setting, foreground matching updating and projectile judgment processing procedures on the received color video image information to obtain projectile information.
(6) The DSP processing chip 4 transmits the processing result to the alarm module 8, and simultaneously transmits the projectile information to the display module 7, so that the projectile information is displayed, and the monitoring personnel is prompted to perform the next processing.
As shown in fig. 2, the intelligent detection method for the projectile mainly includes background modeling, foreground detection, setting of a protection area, foreground matching update, and projectile judgment.
The background modeling is mainly to perform pattern recognition according to color video images input by the color camera 2 and intelligently establish a self-adaptive learning background model. And establishing a background model by using the brightness and the chrominance information of the video image. Because the algorithm performs operation in the embedded equipment, all data information adopted by the algorithm is integer parameters. And sequentially traversing pixel points in the image, searching eight neighborhood pixel points for each pixel point, and establishing a Gaussian model according to the seed points and the eight neighborhood information, wherein parameters of the Gaussian model comprise brightness, chroma, saturation, weight and variance information. A gaussian model is built for each seed point.
The brightness and the chroma of the model are obtained through the seed points and the eight neighborhood related information of the seed points. Such as a matrixThe positions are shown, the middle 1 is seed point pixel information, and the eight neighborhoods are represented by 0. Marking the eight neighborhood pixel point information as P from left to right and from top to bottom in sequence1，P2，P3，P4，P5，P6，P7，P8 Seed point is marked as P0. Establishing a Gaussian model for the seed points according to the weight parameters,
wherein,in order to establish the brightness after the gaussian model,、、、、、、、、respectively the seed point and the brightness of its eight neighborhoods.
In the same way, the Gaussian model of the seed point chroma and saturation is established by the same method,
wherein,in order to establish the chromaticity after the gaussian model,、、、、、、、、respectively, the seed point and its eight neighborhood.
Wherein,in order to establish the saturation after the gaussian model is established,、、、、、、、、the seed point and its saturation of the eight neighborhoods, respectively.
The model variance learning process is as follows:
wherein,represents the result of the variance learning of the image of the new frame,the variance of the current background is represented as,a model variance learning factor is represented as a function of the model variance,、、respectively representing the brightness, the chroma and the saturation information of the Gaussian model at the seed point of the background image.、、Respectively representing the brightness, the chroma and the saturation information of the Gaussian model at the seed point at the corresponding position of the current frame image and the background image.
The model weight records the number of successful modeling of the background seed points. When the total number of background seed point modeling reaches 1/3 of the total number of seed points, the background modeling is successful.
The foreground detection is to detect a new person or object in the background. The foreground detection method adopted by the invention is a method for comparing the Gaussian models of the foreground and the background. And similarly, establishing the brightness, the chroma, the saturation, the weight and the variance information of the Gaussian model parameters of the pixel seed points of the current frame image. And respectively carrying out difference on brightness, chroma, saturation and variance parameters corresponding to the Gaussian models of the foreground and the background. And if the difference value is smaller than a certain threshold value, the seed point is a suspected foreground point. Otherwise, it is a background point. And after comparing all the seed points of the foreground points, carrying out communicated expansion processing on the seed points, so that all the seed points of the same foreground object are communicated, calculating the size, the perimeter, the area, the gravity center and the height-width ratio information of all the communicated areas, eliminating the influence of small false foreground interference points, and then respectively marking masks of different foregrounds with different serial numbers. And storing effective foreground data in the first frame of image after background learning is finished into a foreground historical information base so as to facilitate operations such as foreground matching updating and projectile judgment in the subsequent detection process.
The protection area is set as an area where vehicles run on a tunnel or a road. The protection area information is marked by drawing two lines along the edge of the tunnel or road in the image. And meanwhile, the alarm sensitivity of the throwing object, the monitoring limit of the area of the throwing object and the limit of the height-width ratio are set, so that the influence of too small stones, fruit stones and the like is prevented.
The foreground matching updating process mainly refers to updating a foreground historical information base by using newly detected foreground information. The method comprises the steps of detecting emerging foreground information of a current frame image, comparing foreground historical data according to foreground size, perimeter, area, gravity center and aspect ratio information, if the change is smaller than a threshold value, determining that the foreground historical data belong to the same foreground information, updating the foreground historical data, and if no matched historical information is found, storing the emerging foreground into the historical data.
The projectile event judgment mainly judges whether the foreground is the projectile or not, whether the projectile is in a protection area or not and whether the alarm sensitivity meets the requirements or not. And in the judging process, traversing the foreground historical information base, judging whether the area and the height-width ratio information of the foreground meet threshold setting, and if so, judging that the object is a projectile. Whether the protection area is set is further judged according to the foreground gravity center information, whether the alarm sensitivity meets the requirement is judged if the protection area is set, and if the alarm sensitivity meets the requirement, the object is thrown, alarm information is sent out simultaneously, and the foreground information is marked and displayed on a screen.
Fig. 3 is a diagram showing the intelligent detection alarm condition of the projectile. As shown in fig. 3, the diagonal information on the left and right sides is road or tunnel direction edge information. The upper and lower parts are limited far and near ranges of the monitoring area position, and form a closed area A. The closed area A is a protection area, and only when the closed area has the throwing object information, the alarm can be given. If the projectile 11 is in the protection area A, an alarm can be given out in the detection process; the projectile 10 is in the unprotected area B and no alarm will occur during the detection process.
Any person skilled in the art may, using the teachings disclosed above, change or modify the equivalent embodiments with equivalent changes. However, any simple modification, equivalent change and modification of the above embodiments according to the technical essence of the present invention are within the protection scope of the technical solution of the present invention.
Claims (7)
1. Intelligent detection device and method of shed thing, its characterized in that: the intelligent detection device for the sprinkled objects comprises an intelligent detection module, a display module and an alarm module.
2. The apparatus and method of claim 1, wherein: the intelligent detection module mainly comprises a color camera, a protection device, an infrared illuminating lamp and a DSP processing chip, wherein the color camera acquires color video images, the color camera is connected with the DSP processing chip through a network cable to transmit data, the color camera transmits the images to the DSP processing chip, and the DSP processing chip is mainly an algorithm unit and provides an intelligent detection scheme for the sprinkled objects.
3. The apparatus and method of claim 1, wherein: the intelligent detection device and the intelligent detection method for the sprinkled objects are characterized in that a color camera and a DSP processing chip are packaged in a shell, the shell is a cylindrical hollow packaging shell made of stainless steel 316L, transparent glass is arranged at the front end inside the shell and used for protecting the color camera and the DSP processing chip, the color camera is arranged on one side close to the front end transparent glass, the DSP processing chip is arranged on one side far away from the front end transparent glass, and meanwhile, the color camera can acquire surrounding situation images through the front end glass; the rear end of the shell is provided with two wiring holes, one hole is used for connecting a power line to supply power to the color camera and the DSP processing chip, and the other hole is used for connecting an optical fiber cable and transmitting a video image signal acquired by the color camera and an alarm signal obtained by processing of the DSP processing chip to a monitoring room which is hundreds of meters or kilometers away.
4. The apparatus and method of claim 1, wherein: the intelligent detection device and method for the sprinkled objects comprise the following steps:
color image information acquired by the color camera is simultaneously transmitted to the display module and the DSP processing chip, the display module displays tunnel and road real-time information, and the DSP processing chip receives the color video image information transmitted by the color camera for further intelligent identification processing;
the DSP processing chip carries out background modeling, foreground detection, protected area setting, foreground matching updating and projectile judgment processing procedures on the received color video image information to obtain projectile information;
the DSP processing chip transmits the processing result to the alarm module, and transmits the object throwing information to the display module, so that the object throwing information is displayed, and monitoring personnel are prompted to perform next processing.
5. The apparatus and method of claim 1, wherein: the intelligent detection device and method for the sprinkled objects are characterized in that the background modeling is mainly to perform mode recognition according to color video images input by a color camera, intelligently establish a self-adaptive learning background model, and establish a background model by utilizing brightness and chrominance information of the video images; sequentially traversing pixel points in the image, searching eight neighborhood pixel points for each pixel point, and establishing a Gaussian model according to the seed points and the eight neighborhood information, wherein parameters of the Gaussian model comprise brightness, chroma, saturation, weight and variance information; a Gaussian model is established for each seed point:
the brightness and chroma of the model are obtained by the seed points and the eight neighborhood related information thereof, such as a matrixThe positions are shown, the middle 1 is seed point pixel information, eight neighborhoods are represented by 0, and the eight neighborhood pixel point information is marked as P sequentially from left to right and from top to bottom1，P2，P3，P4，P5，P6，P7，P8 Seed point is marked as P0Establishing a Gaussian model for the seed points according to the weight parameters,
wherein,in order to establish the brightness after the gaussian model,、、、、、、、、respectively the brightness of the seed point and the eight neighborhoods thereof;
in the same way, the Gaussian model of the seed point chroma and saturation is established by the same method,
wherein,in order to establish the chromaticity after the gaussian model,、、、、、、、、respectively the seed point and the chromaticity of eight neighborhoods thereof;
wherein,in order to establish the saturation after the gaussian model is established,、、、、、、、、respectively the saturation of the seed point and the eight neighborhoods thereof;
the model variance learning process is as follows:
wherein,represents the result of the variance learning of the image of the new frame,the variance of the current background is represented as,represents a model variance learning factor,、Respectively representing brightness, chroma and saturation information of the Gaussian model at the seed point of the background image;、、respectively representing the brightness, the chroma and the saturation information of the Gaussian model at the seed point at the corresponding position of the current frame image and the background image.
6. The apparatus and method of claim 1, wherein: the foreground detection of the intelligent detection device and the intelligent detection method for the sprinkled objects is to detect newly appeared people or objects in the background; the method for comparing the Gaussian models of the foreground and the background is adopted, the Gaussian model parameters of the pixel seed point of the current frame image such as brightness, chroma, saturation, weight and variance information are established in the same way, the brightness, the chroma, the saturation and the variance parameters corresponding to the Gaussian models of the foreground and the background are respectively subjected to difference, if the difference value is smaller than a certain threshold value, the seed point is a suspected foreground point, otherwise, the seed point is a background point; and after comparing all the seed points of the foreground points, carrying out communicated expansion processing on the seed points, so that all the seed points of the same foreground object are communicated, calculating the size, the perimeter, the area, the gravity center and the height-width ratio information of all the communicated areas, eliminating the influence of small false foreground interference points, and then respectively marking masks of different foregrounds with different serial numbers.
7. The apparatus and method of claim 1, wherein: according to the intelligent detection device and method for the sprinkled objects, the protection area is set to be a tunnel or a road, and the area where vehicles run is marked by drawing two lines along the edge of the tunnel or the road in an image to mark the information of the protection area, and meanwhile, the alarm sensitivity of the sprinkled objects, the monitoring limit of the area of the sprinkled objects and the limit of the height-width ratio are set, so that the influence of undersized stones, fruit stones and the like is prevented.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410688800.7A CN104392630A (en) | 2014-11-26 | 2014-11-26 | Throw-out intelligent detection device and method |
| CN201510595135.1A CN105245831B (en) | 2014-11-26 | 2015-09-18 | Shed object intelligent detection device |
| CN201510598595.XA CN105187800B (en) | 2014-11-26 | 2015-09-18 | Detector with high definition modular structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410688800.7A CN104392630A (en) | 2014-11-26 | 2014-11-26 | Throw-out intelligent detection device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104392630A true CN104392630A (en) | 2015-03-04 |
Family
ID=52610523
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410688800.7A Pending CN104392630A (en) | 2014-11-26 | 2014-11-26 | Throw-out intelligent detection device and method |
| CN201510598595.XA Active CN105187800B (en) | 2014-11-26 | 2015-09-18 | Detector with high definition modular structure |
| CN201510595135.1A Active CN105245831B (en) | 2014-11-26 | 2015-09-18 | Shed object intelligent detection device |
Family Applications After (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510598595.XA Active CN105187800B (en) | 2014-11-26 | 2015-09-18 | Detector with high definition modular structure |
| CN201510595135.1A Active CN105245831B (en) | 2014-11-26 | 2015-09-18 | Shed object intelligent detection device |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN104392630A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105141925A (en) * | 2015-09-18 | 2015-12-09 | 天津艾思科尔科技有限公司 | Detector with transmission line equipped with network connector |
| CN105187800A (en) * | 2014-11-26 | 2015-12-23 | 天津艾思科尔科技有限公司 | Detector with high definition module structure |
| CN105227909A (en) * | 2015-09-18 | 2016-01-06 | 天津艾思科尔科技有限公司 | Anti-static type detector |
| CN107067725A (en) * | 2017-05-26 | 2017-08-18 | 安徽皖通科技股份有限公司 | Tunnel road conditions dynamic early-warning and linkage method of disposal |
| CN107396043A (en) * | 2017-07-19 | 2017-11-24 | 天津市广通信息技术工程股份有限公司 | Muck truck side sprinkling monitoring system based on wireless communication |
| CN107918762A (en) * | 2017-10-24 | 2018-04-17 | 江西省高速公路投资集团有限责任公司 | A kind of highway drops thing rapid detection system and method |
| CN111127507A (en) * | 2019-12-18 | 2020-05-08 | 成都通甲优博科技有限责任公司 | Method and system for determining throwing object |
| CN111163285A (en) * | 2018-11-08 | 2020-05-15 | 佳维技术有限公司 | High-altitude falling object monitoring method and system and computer readable storage medium |
| CN111274982A (en) * | 2020-02-04 | 2020-06-12 | 浙江大华技术股份有限公司 | Method and device for identifying projectile and storage medium |
| CN111654664A (en) * | 2020-05-08 | 2020-09-11 | 浙江大华技术股份有限公司 | High-altitude parabolic detection method and system, computer equipment and storage medium |
| CN111814668A (en) * | 2020-07-08 | 2020-10-23 | 北京百度网讯科技有限公司 | Method and device for detecting road sprinklers |
| CN111951573A (en) * | 2020-07-21 | 2020-11-17 | 华设设计集团股份有限公司 | Intelligent public transportation system and method based on vehicle-road cooperation technology |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106442528A (en) * | 2016-09-09 | 2017-02-22 | 上海新纤仪器有限公司 | Dual camera microscope testing device and testing method for contents of fiber components |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3372140B2 (en) * | 1995-06-27 | 2003-01-27 | 松下電工株式会社 | Combined smoke and heat detector |
| CN201203929Y (en) * | 2008-03-24 | 2009-03-04 | 武光杰 | Smog induction module |
| CN202043404U (en) * | 2011-04-29 | 2011-11-16 | 深圳市豪恩安全科技有限公司 | Infrared detector and security prevention system |
| CN102509075B (en) * | 2011-10-19 | 2013-07-24 | 北京国铁华晨通信信息技术有限公司 | Remnant object detection method and device |
| CN102945603B (en) * | 2012-10-26 | 2015-06-03 | 青岛海信网络科技股份有限公司 | Method for detecting traffic event and electronic police device |
| WO2014178382A1 (en) * | 2013-04-30 | 2014-11-06 | 株式会社コベルコ科研 | Li-CONTAINING OXIDE TARGET ASSEMBLY |
| CN103685895B (en) * | 2013-12-30 | 2017-04-26 | 博康智能网络科技有限公司 | Surveillance camera housing |
| CN204143578U (en) * | 2014-09-27 | 2015-02-04 | 北京宏大京电电子技术有限公司 | A kind of infrared laser correlative detector |
| CN104394361A (en) * | 2014-11-20 | 2015-03-04 | 天津艾思科尔科技有限公司 | Pedestrian crossing intelligent monitoring device and detection method |
| CN104392630A (en) * | 2014-11-26 | 2015-03-04 | 天津艾思科尔科技有限公司 | Throw-out intelligent detection device and method |
| CN204496642U (en) * | 2015-04-08 | 2015-07-22 | 天津艾思科尔科技有限公司 | A kind of dual-band image type fire detecting arrangement |
| CN204637385U (en) * | 2015-05-04 | 2015-09-16 | 上海赛复安防科技有限公司 | A kind of track and localization jet extinguishing device automatically |
| CN205378015U (en) * | 2015-09-18 | 2016-07-06 | 天津艾思科尔科技有限公司 | Detecting device with two cameras |
-
2014
- 2014-11-26 CN CN201410688800.7A patent/CN104392630A/en active Pending
-
2015
- 2015-09-18 CN CN201510598595.XA patent/CN105187800B/en active Active
- 2015-09-18 CN CN201510595135.1A patent/CN105245831B/en active Active
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105187800B (en) * | 2014-11-26 | 2019-01-08 | 天津艾思科尔科技有限公司 | Detector with high definition modular structure |
| CN105187800A (en) * | 2014-11-26 | 2015-12-23 | 天津艾思科尔科技有限公司 | Detector with high definition module structure |
| CN105245831A (en) * | 2014-11-26 | 2016-01-13 | 天津艾思科尔科技有限公司 | Detector based on binocular vision |
| CN105245831B (en) * | 2014-11-26 | 2019-10-08 | 天津艾思科尔科技有限公司 | Shed object intelligent detection device |
| CN105227909A (en) * | 2015-09-18 | 2016-01-06 | 天津艾思科尔科技有限公司 | Anti-static type detector |
| CN105141925A (en) * | 2015-09-18 | 2015-12-09 | 天津艾思科尔科技有限公司 | Detector with transmission line equipped with network connector |
| CN107067725A (en) * | 2017-05-26 | 2017-08-18 | 安徽皖通科技股份有限公司 | Tunnel road conditions dynamic early-warning and linkage method of disposal |
| CN107396043A (en) * | 2017-07-19 | 2017-11-24 | 天津市广通信息技术工程股份有限公司 | Muck truck side sprinkling monitoring system based on wireless communication |
| CN107918762A (en) * | 2017-10-24 | 2018-04-17 | 江西省高速公路投资集团有限责任公司 | A kind of highway drops thing rapid detection system and method |
| CN107918762B (en) * | 2017-10-24 | 2022-01-14 | 江西省高速公路投资集团有限责任公司 | Rapid detection system and method for road scattered objects |
| CN111163285A (en) * | 2018-11-08 | 2020-05-15 | 佳维技术有限公司 | High-altitude falling object monitoring method and system and computer readable storage medium |
| CN111127507A (en) * | 2019-12-18 | 2020-05-08 | 成都通甲优博科技有限责任公司 | Method and system for determining throwing object |
| CN111274982A (en) * | 2020-02-04 | 2020-06-12 | 浙江大华技术股份有限公司 | Method and device for identifying projectile and storage medium |
| CN111274982B (en) * | 2020-02-04 | 2023-04-07 | 浙江大华技术股份有限公司 | Method and device for identifying projectile and storage medium |
| CN111654664A (en) * | 2020-05-08 | 2020-09-11 | 浙江大华技术股份有限公司 | High-altitude parabolic detection method and system, computer equipment and storage medium |
| CN111814668A (en) * | 2020-07-08 | 2020-10-23 | 北京百度网讯科技有限公司 | Method and device for detecting road sprinklers |
| CN111814668B (en) * | 2020-07-08 | 2024-05-10 | 北京百度网讯科技有限公司 | Method and device for detecting road sprinklers |
| CN111951573A (en) * | 2020-07-21 | 2020-11-17 | 华设设计集团股份有限公司 | Intelligent public transportation system and method based on vehicle-road cooperation technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105245831A (en) | 2016-01-13 |
| CN105245831B (en) | 2019-10-08 |
| CN105187800B (en) | 2019-01-08 |
| CN105187800A (en) | 2015-12-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104392630A (en) | Throw-out intelligent detection device and method | |
| CN104036575B (en) | Working-yard safety helmet wear condition monitoring method | |
| CN101872526B (en) | Smoke and fire intelligent identification method based on programmable photographing technology | |
| CN107331097A (en) | The periphery intrusion preventing apparatus and method merged based on target position information | |
| CN104394361A (en) | Pedestrian crossing intelligent monitoring device and detection method | |
| CN104091354A (en) | Fire detection method based on video images and fire detection device thereof | |
| CN109323132B (en) | Long-distance pipeline unmanned aerial vehicle detection system based on optical fiber early warning technology | |
| CN204990743U (en) | Superelevation vehicle early warning system | |
| CN201890600U (en) | Machine vision belt tearing detecting device | |
| CN104469309A (en) | Tunnel pedestrian intrusion detection device and method | |
| CN104049281A (en) | Device and method for automatically detecting foreign matter between screen door of curve subway platform and train | |
| CN111898563A (en) | Comprehensive safety monitoring equipment and method for protected area | |
| CN105575121A (en) | Intelligent traffic-based source technology overload control and prevention data acquisition system and method | |
| CN110441320A (en) | A kind of gangue detection method, apparatus and system | |
| CN103152558B (en) | Based on the intrusion detection method of scene Recognition | |
| CN107877518A (en) | A kind of crusing robot and its fall arrest method and apparatus | |
| CN116434533A (en) | AI wisdom highway tunnel synthesizes monitoring platform based on 5G | |
| CN116453278A (en) | Intrusion monitoring method combining deep learning intelligent detection and optical fiber vibration sensing | |
| CN109345787A (en) | A kind of anti-outer damage monitoring and alarming system of the transmission line of electricity based on intelligent image identification technology | |
| CN205378015U (en) | Detecting device with two cameras | |
| CN114283544A (en) | Railway platform intrusion monitoring system and method based on artificial intelligence | |
| CN105046223A (en) | Device for detecting severity of ''black-hole effect'' at tunnel entrance and method thereof | |
| CN105262984B (en) | A kind of detector with fixing device | |
| CN117011791A (en) | Mine fire disaster identification and alarm method based on image contour area characteristics | |
| CN104417504B (en) | The security protection subsystem of battery replacement of electric automobile 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 | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150304 |