CN103345625A - Method and system for analyzing three-dimensional image - Google Patents
Method and system for analyzing three-dimensional image Download PDFInfo
- Publication number
- CN103345625A CN103345625A CN2013102995236A CN201310299523A CN103345625A CN 103345625 A CN103345625 A CN 103345625A CN 2013102995236 A CN2013102995236 A CN 2013102995236A CN 201310299523 A CN201310299523 A CN 201310299523A CN 103345625 A CN103345625 A CN 103345625A
- Authority
- CN
- China
- Prior art keywords
- monitoring image
- characteristic quantity
- sample characteristics
- image
- characteristics amount
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Processing (AREA)
- Alarm Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method and system for analyzing a three-dimensional image. The method comprises the following steps of collecting a monitor image by using a camera, carrying out advanced processing on the monitor image to obtain a three-dimensional monitor image, storing the three-dimensional monitor image in a three-dimensional buffer area, extracting the three-dimensional monitor image from the three-dimensional buffer area, converting information in the three-dimensional monitor image into verification characteristic quantity by using a preset algorithm, maintaining sample characteristic quantity in inner storage in advance, reading the sample characteristic quantity from the inner storage, matching the sample characteristic quantity and the verification characteristic quantity, judging whether a human body appears in the three-dimensional monitor image or not, and changing an alarm mark if the human body appears.
Description
Technical field
The present invention relates to technical field of video monitoring, particularly a kind of 3-D view analytical approach and system.
Background technology
Safety problem concerns broad masses' life and property, always all is one of of paramount importance factor in the middle of the daily routines; No matter the particularly maintenance of public place safety still is that whole society all has great meaning for individuality, is the most important thing.And along with in recent years, social security events happens occasionally all over the world, and especially at bank, large-scale court, meeting venue, exhibition center's large-scale public place, people are more and more stronger to the demand of safety.
Intelligent security-protecting and monitoring now is applied in the middle of the various large-scale public places widely as a kind of effective security protection means.Traditional intelligent security-protecting and monitoring utilizes in the camera collection place video monitoring picture everywhere, and utilizes computer equipment that monitored picture is handled, shown and storage.That is to say functions such as traditional supervisory system possesses the catching of the video of knowing clearly, stores, distribution.
But in traditional supervisory system, the content of monitored picture can only be leaned on artificial judgment, that is to say the staff needs to pay close attention to for a long time monitored picture, observes whether the someone occurs in the monitored picture.Above-mentioned manual work has been brought serious burden for the staff, need very high requirement be proposed to the reaction capacity of staff's notice, vigilance and abnormal conditions, and manual work can cause staff's decreased attention for a long time, thereby cause the monitoring error, for safety is brought hidden danger.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of 3-D view analytical approach and system, by the collection 3-D view, and in conjunction with the 3-D view analytical technology, realize whether having human body to occur in the identification monitored picture automatically.
For achieving the above object, the present invention has following technical scheme:
A kind of 3-D view analytical approach said method comprising the steps of:
Utilize the camera collection monitoring image, monitoring image is carried out advanced treating obtain the 3 D monitoring image, and the 3 D monitoring image is kept in the three-dimensional buffer zone;
Extract the 3 D monitoring image from three-dimensional buffer zone, utilize default algorithm that information translation in the 3 D monitoring image is the checking characteristic quantity;
In advance the sample characteristics amount is remained in the internal memory, from internal memory, read the sample characteristics amount, with sample characteristics amount and described checking characteristic quantity coupling;
Judge whether have human body to occur in the 3 D monitoring image according to matching result, as human body occurs and then change alarm banner.
Described advanced treating is specially:
Zone emission infrared ray speckle to the camera collection image utilizes speckle and two-dimentional monitoring image to carry out overlay analysis and obtains the 3 D monitoring image.
Described from three-dimensional buffer zone extraction 3 D monitoring image, utilize default algorithm that information translation in the 3 D monitoring image is specially for the checking characteristic quantity:
From three-dimensional buffer zone, extract a frame 3 D monitoring image as first monitoring image, extract a back frame 3 D monitoring image of first monitoring image as second monitoring image;
Contrast first monitoring image and second monitoring image, judge whether the two exists the modified-image zone;
When there not being the modified-image zone, then utilize default algorithm that full detail in second monitoring image is converted to the first checking characteristic quantity;
When there being the modified-image zone, then utilize default algorithm just the modified-image area information be converted to the second checking characteristic quantity.
Described default algorithm comprises model algorithm and characteristics algorithm, and then described is that the checking characteristic quantity is specially with information translation in the 3 D monitoring image:
Utilize model algorithm that the information that comprises in the 3 D monitoring image is set up object model;
Utilize characteristics algorithm that described object model is converted to the checking characteristic quantity.
Described method also comprises:
Human body do not occur when the matching result according to sample characteristics amount and the first checking characteristic quantity is judged as, then the 3 D monitoring image that does not change for later relative second monitoring image of second monitoring image directly is judged as and does not change alarm banner.
Described in advance the sample characteristics amount being remained in the internal memory is specially:
Obtain the sample characteristics amount by sample training, in advance the sample characteristics amount is remained in the internal memory.
The described sample characteristics measuring that reads from internal memory is:
Sample characteristics amount in the internal memory is set up index, utilize search index and read the sample characteristics amount.
Described sample characteristics measuring is:
Human sample characteristic quantity, people's face sample characteristics amount, human eye sample characteristics amount, mouth sample characteristics amount and people's nose sample characteristics amount.
A kind of 3-D view analytic system, described system specifically comprises following:
Depth module is used for that the monitoring image that utilizes camera collection is carried out advanced treating and obtains the 3 D monitoring image, and the 3 D monitoring image is kept in the three-dimensional buffer zone;
The characteristic quantity modular converter is used for extracting the 3 D monitoring image from three-dimensional buffer zone, utilizes default algorithm that information translation in the 3 D monitoring image is the checking characteristic quantity;
Stay memory modules, be used for the sample characteristics amount is remained on internal memory;
Matching module is used for reading the sample characteristics amount from internal memory, with sample characteristics amount and described checking characteristic quantity coupling;
Judge the change module, be used for judging according to matching result whether the 3 D monitoring image has the human body appearance, as human body occurs and then change alarm banner.
Described characteristic quantity modular converter comprises:
Model unit is used for utilizing model algorithm that the information that the 3 D monitoring image comprises is set up object model;
Feature unit is used for utilizing characteristics algorithm that described object model is converted to the checking characteristic quantity.
The described memory modules of staying comprises:
Training unit is used for obtaining the sample characteristics amount and remaining on internal memory by sample training;
Indexing units is used for the sample characteristics amount of internal memory is set up index.
Described matching module comprises:
The inquiry extraction unit is used for utilizing search index and reading the sample characteristics amount;
Matching unit is used for sample characteristics amount and described checking characteristic quantity coupling.
Described characteristic quantity modular converter comprises:
The frame extraction unit is used for extracting a frame 3 D monitoring image as first monitoring image from three-dimensional buffer zone, extracts a back frame 3 D monitoring image of first monitoring image as second monitoring image;
The contrast unit is used for contrast first monitoring image and second monitoring image, judges whether the two exists the modified-image zone;
Converting unit is used for not having the modified-image zone, then utilizes default algorithm that full detail in second monitoring image is converted to the first checking characteristic quantity; Or have the modified-image zone, then utilize default algorithm just the modified-image area information be converted to the second checking characteristic quantity.
As seen through the above technical solutions, the beneficial effect that the present invention exists is: the monitoring image of gathering is carried out advanced treating obtain the 3 D monitoring image, by setting up the coupling of object model and characteristic quantity, realize human body whether occurring in the automatic analyzing three-dimensional monitoring image, replace the process of artificial surveillance monitor image with this, the labor-saving while has also been improved accuracy rate; In addition by in advance the sample characteristics amount being remained in the internal memory, make in the image analysis process process that reads of sample characteristics amount direct fast, can satisfy the requirement of safety monitoring real-time; Also increase the sample characteristics amount in the described method as required, in graphical analysis, in conjunction with human body identification and the identification of people's face face, improved the accuracy rate of graphical analysis; Method described in the present embodiment is handled static state and dynamic monitored picture respectively, and gets rid of the replicate analysis for static 3 D monitoring image, has avoided unnecessary image analysis process, has alleviated the workload of system's operation.
Description of drawings
In order to be illustrated more clearly in the embodiment of the invention or technical scheme of the prior art, to do to introduce simply to the accompanying drawing of required use in embodiment or the description of the Prior Art below, apparently, accompanying drawing in describing below is some embodiments of the present invention, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is the described method flow diagram of the embodiment of the invention;
Fig. 2 is the described method flow diagram of another embodiment of the present invention;
Fig. 3 is the described system architecture synoptic diagram of the embodiment of the invention.
Embodiment
For the purpose, technical scheme and the advantage that make the embodiment of the invention clearer, below in conjunction with the accompanying drawing in the embodiment of the invention, technical scheme in the embodiment of the invention is clearly and completely described, obviously, described embodiment is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.
Existing intelligent safety defense monitoring system, comprise client and main control end, the some cameras of control under the client, its operational scheme may be summarized to be following: after client is utilized the camera collection monitoring image, compression also is sent to main control end, be presented in the middle of the display after main control end decompresses monitoring image, whether the staff can a suspect occur to judge monitoring position by monitoring image in the observation display.
The method of the invention and system implement in conjunction with traditional intelligent safety defense monitoring system equally, utilize described method that the monitoring image in the intelligent safety defense monitoring system is carried out advanced treating and 3-D view analysis, judge whether monitoring position a suspect occurs, substitute the work of staff's observation display thus.Following examples are all based on above-mentioned application background.
Referring to shown in Figure 1, be a specific embodiment of 3-D view analytical approach of the present invention.Present embodiment said method comprising the steps of:
In aforesaid intelligent safety defense monitoring system, utilize camera directly to gather the two-dimensional plane monitoring image.Method described in the present embodiment is independently acquisition monitoring image and advanced treating then, obtains the 3 D monitoring image.In the present embodiment, the analysis of 3 D monitoring image and the transmission of common monitoring image and reality are two parallel processes.
The main body thinking of described advanced treating can be as following: by the zone emission infrared ray speckle of infrared light probe to the camera collection image, the recycling infrared remote receiver receives this speckle, and utilize corresponding hardware chip that speckle and two-dimentional monitoring image are carried out the overlay analysis computing, two-dimentional monitoring image is converted to the 3 D monitoring image.Need to prove that above-mentioned advanced treatment process implementation in the art is not unique, can select suitable concrete grammar according to actual operating position, do not make restriction in the present embodiment.
Meanwhile, in intelligent safety defense monitoring system, equally according to original flow process, the common monitoring image that client compression camera is directly gathered also is sent to main control end, and main control end is presented at display after monitoring image is decompressed.
Monitoring image is carried out advanced treating and be kept in the three-dimensional buffer zone, can be convenient to the extraction of 3 D monitoring image, for the subsequent analysis process.According to the workflow of traditional intelligence safety defense monitoring system, described advanced treating and analytic process occur in client, and described three-dimensional buffer zone also is arranged on client.For saving communication flows, described 3 D monitoring image is not sent to main control end.
The algorithm that described utilization is preset is with in the process of information translation in the 3 D monitoring image for the checking characteristic quantity, described algorithm comprises model algorithm and characteristics algorithm two parts, detailed process is as follows: at first utilize model algorithm that the information that comprises in the 3 D monitoring image is set up object model, the recycling characteristics algorithm is converted to the checking characteristic quantity with described object model.
The information that comprises in the monitoring image generally all is the image of actual object, comes the three-dimensional information of each pixel in the analysis monitoring image by model algorithm, can determine the position feature of object in the monitoring image, sets up object model according to these position features again.
Object model is converted to the process of checking characteristic quantity, in fact just with the process of the data exhibiting graph data feature of textual form.Finish after the above-mentioned conversion, can utilize characteristic quantity to mate and graphical analysis.The described characteristics algorithm that utilizes is not unique in the art, and those skilled in the art can select voluntarily according to actual conditions, do not do concrete restriction in the present embodiment.
After being verified characteristic quantity, also need to mate with checking characteristic quantity and sample characteristics amount, with the concrete implication of information expressed in the identification checking characteristic quantity.Described sample characteristics amount obtains through sample training in advance, is representing certain fixedly characteristic quantity of implication, if checking characteristic quantity and certain sample characteristics amount are complementary, illustrates that then this checking characteristic quantity expressed the implication that this sample characteristics scale shows equally.
In the present embodiment, described sample characteristics measuring comprises: human sample characteristic quantity, people's face sample characteristics amount, human eye sample characteristics amount, mouth sample characteristics amount and people's nose sample characteristics amount.
For example, the lip that certain sample characteristics amount A fixed table is leted others have a look at, if certain the checking characteristic quantity A ' that converts from object model mates with sample characteristics amount A so, then explanation checking characteristic quantity A ' also represents people's lip; Further as can be known, being converted to the object model institute image represented of verifying characteristic quantity A ' is exactly people's lip.
In the present embodiment, the source of sample characteristics amount is primarily aimed at the real work needs of intelligent safety defense monitoring system, carries out sample training in advance and obtains.
In addition, need to prove also that characteristic quantity is preserved with the xml form in the prior art, call characteristic quantity as need and mate, must obtain from file, speed is slower, can't satisfy the real-time requirement of safety-protection system monitoring, it directly is combined in the safety-protection system is nonsensical.In the present embodiment, all sample characteristics amounts are all remained in the middle of the internal memory of client, and set up index for the sample characteristics amount.Can be according to search index to required sample characteristics amount and obtaining immediately when mating, thus satisfied the requirement of real-time.
The checking characteristic quantity that 3 D monitoring image transitions obtains often exists a plurality of, and each checking characteristic quantity is expressed an object model in 3 D monitoring image.The purpose of overall technical architecture just is whether the someone occurs in the monitored picture in real-time judging in the present embodiment.So at first be complementary with checking characteristic quantity and the sample characteristics amount that represents human body, judge whether have human body to occur in the monitored picture by matching result.For improving the accuracy rate of graphical analysis, also will further carry out the coupling of people's face and face with some sample characteristics amounts of checking characteristic quantity coupling representative facial characteristics in addition.
Have human body to occur in the 3 D monitoring image if matching result shows, also just meaning in the zone, place of camera head monitor has the people to enter, and this moment, described method made main control end receive alarm by the change sign of reporting to the police, and the prompting staff confirms concrete situation.
According to the difference in graphical analysis place, the processing mode of described alarm banner also has corresponding variation thereupon.
For example, carry out image when client and handle, what then client also need be real-time reports alarm banner to main control end, does not have in monitored picture under the situation of human body appearance, and the described alarm banner that reports can be designated 0 always, and the main control end place does not produce alarm; And after human body occurring, alarm banner being changed to 1, main control end produces alarm.And when main control end is carried out the image processing, then producing alarm banner by main control end self, alarm banner need not communication process.
Present embodiment is a basic embodiment of described method, by above embodiment as can be known, the beneficial effect that described method exists is: the monitoring image of gathering is carried out advanced treating obtain the 3 D monitoring image, by setting up the coupling of object model and characteristic quantity, realize human body whether occurring in the automatic analyzing three-dimensional monitoring image, replace the process of artificial surveillance monitor image with this, the labor-saving while has also been improved accuracy rate; In addition by in advance the sample characteristics amount being remained in the internal memory, make in the image analysis process process that reads of sample characteristics amount direct fast, can satisfy the requirement of safety monitoring real-time; Also increase the sample characteristics amount in the described method as required, in graphical analysis, in conjunction with human body identification and the identification of people's face face, improved the accuracy rate of graphical analysis.
Referring to shown in Figure 2, be another specific embodiment of the method for the invention, present embodiment has been made disclosure and description more specifically on basis embodiment illustrated in fig. 1.In the present embodiment, said method comprising the steps of:
In the present embodiment, in the process of graphical analysis, combine the scheme at the dynamic change of 3 D monitoring image.Just extract two adjacent frame monitoring images respectively as first monitoring image and second monitoring image, contrast the two learning whether monitoring image is changing, and then handle respectively changing with constant two kinds of situations.
For the relative situation about not changing with second monitoring image of first monitoring image, then be as the criterion with second monitoring image, information whole in second monitoring image all is converted to the checking characteristic quantity, and the characteristic quantity that is converted to by second monitoring image in the present embodiment is called as the first checking characteristic quantity.
Obtaining of checking characteristic quantity comprises equally and sets up object model and object model is converted to checking characteristic quantity two steps in the present embodiment, is equal to previous embodiment.
The algorithm that step 206, utilization are preset just modified-image area information is converted to the second checking characteristic quantity.
Because the article in the guarded region in site surrounding and the place generally can not change, very likely be the activity that human body has appearred in guarded region in case occur changing.So situation about changing for first monitoring image and second monitoring image, no longer whole 3 D monitoring image is all analyzed, but only information in the image-region that changes is set up object model, and be converted to checking characteristic quantity and being analyzed, save unnecessary image analysis process thus; The checking characteristic quantity that is obtained by the modified-image area information in the present embodiment is called as the second checking characteristic quantity.
Obtaining of checking characteristic quantity comprises equally and sets up object model and object model is converted to checking characteristic quantity two steps in the present embodiment, is equal to previous embodiment.
The amount of sample characteristics described in the present embodiment and above-described embodiment are as broad as long, and the matching process of the first checking characteristic quantity or the second checking characteristic quantity and sample characteristics amount also can be with reference to the idiographic flow in the previous embodiment.The change of described alarm banner is same consistent with previous embodiment.
But, for preferably can also be in conjunction with following steps in the checking present embodiment of the first checking characteristic quantity:
Step 209, when the matching result according to sample characteristics amount and the first checking characteristic quantity is judged as and human body do not occur, then the monitoring image that does not change for later relative second monitoring image of second monitoring image directly is judged as and does not change alarm banner.
If obviously do not comprise human body in second monitoring image, if second monitoring image each frame 3 D monitoring image afterwards is consistent with second monitoring image so, then must not comprise human body yet.So only need verify the consistance of 3 D monitoring image, need not to verify again conversion and the image analysis process of characteristic quantity.Variation has taken place in a certain frame monitoring image after the second checking characteristic quantity, enters step 206 again and carries out graphical analysis.Can save unnecessary graphical analysis according to above-mentioned optimization flow process.
The beneficial effect that the described method of present embodiment exists is: method described in the present embodiment is handled static state and dynamic monitored picture respectively, and eliminating is for the replicate analysis of static 3 D monitoring image, avoid unnecessary image analysis process, alleviated the workload of system's operation; The overall technical architecture of method described in the present embodiment is more complete, and is open more abundant.
Referring to shown in Figure 3, be the structural representation of 3-D view analytic system of the present invention.Method described in system described in the present embodiment is corresponding embodiment illustrated in fig. 1, the two technical scheme are consistent in itself, and concrete, described system comprises following:
Depth module is used for that the monitoring image that utilizes camera collection is carried out advanced treating and obtains the 3 D monitoring image, and the 3 D monitoring image is kept in the three-dimensional buffer zone;
The characteristic quantity modular converter is used for extracting the 3 D monitoring image from three-dimensional buffer zone, utilizes default algorithm that information translation in the 3 D monitoring image is the checking characteristic quantity;
Stay memory modules, be used for the sample characteristics amount is remained on internal memory;
Matching module is used for reading the sample characteristics amount from internal memory, with sample characteristics amount and described checking characteristic quantity coupling;
Judge the change module, be used for judging according to matching result whether the 3 D monitoring image has the human body appearance, as human body occurs and then change alarm banner.
Present embodiment is for being a basic embodiment of described system, by above embodiment as can be known, the beneficial effect that described system exists is: the monitoring image of gathering is carried out advanced treating obtain the 3 D monitoring image, by setting up the coupling of object model and characteristic quantity, realize human body whether occurring in the automatic analyzing three-dimensional monitoring image, replace the process of artificial surveillance monitor image with this, the labor-saving while has also been improved accuracy rate; In addition by in advance the sample characteristics amount being remained in the internal memory, make in the image analysis process process that reads of sample characteristics amount direct fast, can satisfy the requirement of safety monitoring real-time; Also increase the sample characteristics amount in the described method as required, in graphical analysis, in conjunction with human body identification and the identification of people's face face, improved the accuracy rate of graphical analysis.。
In addition, the described system of present embodiment can also be specially in conjunction with following prioritization scheme:
Described characteristic quantity modular converter comprises:
Model unit is used for utilizing model algorithm that the information that the 3 D monitoring image comprises is set up object model;
Feature unit is used for utilizing characteristics algorithm that described object model is converted to the checking characteristic quantity.
The described memory modules of staying comprises:
Training unit is used for obtaining the sample characteristics amount and remaining on internal memory by sample training;
Indexing units is used for the sample characteristics amount of internal memory is set up index.
Described matching module comprises:
The inquiry extraction unit is used for utilizing search index and reading the sample characteristics amount;
Matching unit is used for sample characteristics amount and described checking characteristic quantity coupling.
Described characteristic quantity modular converter comprises:
The frame extraction unit is used for extracting a frame 3 D monitoring image as first monitoring image from three-dimensional buffer zone, extracts a back frame 3 D monitoring image of first monitoring image as second monitoring image;
The contrast unit is used for contrast first monitoring image and second monitoring image, judges whether the two exists the modified-image zone;
Converting unit is used for not having the modified-image zone, then utilizes default algorithm that full detail in second monitoring image is converted to the first checking characteristic quantity; Or have the modified-image zone, then utilize default algorithm just the modified-image area information be converted to the second checking characteristic quantity.
The beneficial effect of realizing by above prioritization scheme is:
Static state and dynamic monitored picture are handled respectively, and got rid of replicate analysis for static 3 D monitoring image, avoided unnecessary image analysis process, alleviated the workload of system's operation; The technical scheme of entire system described in the present embodiment is more complete, and is open more abundant.。
The above only is preferred implementation of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the principle of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (13)
1. a 3-D view analytical approach is characterized in that, said method comprising the steps of:
Utilize the camera collection monitoring image, monitoring image is carried out advanced treating obtain the 3 D monitoring image, and the 3 D monitoring image is kept in the three-dimensional buffer zone;
Extract the 3 D monitoring image from three-dimensional buffer zone, utilize default algorithm that information translation in the 3 D monitoring image is the checking characteristic quantity;
In advance the sample characteristics amount is remained in the internal memory, from internal memory, read the sample characteristics amount, with sample characteristics amount and described checking characteristic quantity coupling;
Judge whether have human body to occur in the 3 D monitoring image according to matching result, as human body occurs and then change alarm banner.
2. according to the described method of claim 1, it is characterized in that described advanced treating is specially:
Zone emission infrared ray speckle to the camera collection image utilizes speckle and two-dimentional monitoring image to carry out overlay analysis and obtains the 3 D monitoring image.
3. according to the described method of claim 1, it is characterized in that, described from three-dimensional buffer zone extraction 3 D monitoring image, utilize default algorithm that information translation in the 3 D monitoring image is specially for the checking characteristic quantity:
From three-dimensional buffer zone, extract a frame 3 D monitoring image as first monitoring image, extract a back frame 3 D monitoring image of first monitoring image as second monitoring image;
Contrast first monitoring image and second monitoring image, judge whether the two exists the modified-image zone;
When there not being the modified-image zone, then utilize default algorithm that full detail in second monitoring image is converted to the first checking characteristic quantity;
When there being the modified-image zone, then utilize default algorithm just the modified-image area information be converted to the second checking characteristic quantity.
4. according to claim 1 or 3 described methods, it is characterized in that described default algorithm comprises model algorithm and characteristics algorithm, then described is that the checking characteristic quantity is specially with information translation in the 3 D monitoring image:
Utilize model algorithm that the information that comprises in the 3 D monitoring image is set up object model;
Utilize characteristics algorithm that described object model is converted to the checking characteristic quantity.
5. according to the described method of claim 3, it is characterized in that described method also comprises:
Human body do not occur when the matching result according to sample characteristics amount and the first checking characteristic quantity is judged as, then the 3 D monitoring image that does not change for later relative second monitoring image of second monitoring image directly is judged as and does not change alarm banner.
6. according to the described method of claim 1, it is characterized in that described in advance the sample characteristics amount being remained in the internal memory is specially:
Obtain the sample characteristics amount by sample training, in advance the sample characteristics amount is remained in the internal memory.
7. according to the described method of claim 1, it is characterized in that the described sample characteristics measuring that reads is from internal memory:
Sample characteristics amount in the internal memory is set up index, utilize search index and read the sample characteristics amount.
8. according to the described method of claim 1, it is characterized in that described sample characteristics measuring is:
Human sample characteristic quantity, people's face sample characteristics amount, human eye sample characteristics amount, mouth sample characteristics amount and people's nose sample characteristics amount.
9. a 3-D view analytic system is characterized in that, described system specifically comprises following:
Depth module is used for that the monitoring image that utilizes camera collection is carried out advanced treating and obtains the 3 D monitoring image, and the 3 D monitoring image is kept in the three-dimensional buffer zone;
The characteristic quantity modular converter is used for extracting the 3 D monitoring image from three-dimensional buffer zone, utilizes default algorithm that information translation in the 3 D monitoring image is the checking characteristic quantity;
Stay memory modules, be used for the sample characteristics amount is remained on internal memory;
Matching module is used for reading the sample characteristics amount from internal memory, with sample characteristics amount and described checking characteristic quantity coupling;
Judge the change module, be used for judging according to matching result whether the 3 D monitoring image has the human body appearance, as human body occurs and then change alarm banner.
10. according to the described system of claim 9, it is characterized in that described characteristic quantity modular converter comprises:
Model unit is used for utilizing model algorithm that the information that the 3 D monitoring image comprises is set up object model;
Feature unit is used for utilizing characteristics algorithm that described object model is converted to the checking characteristic quantity.
11., it is characterized in that the described memory modules of staying comprises according to the described system of claim 9:
Training unit is used for obtaining the sample characteristics amount and remaining on internal memory by sample training;
Indexing units is used for the sample characteristics amount of internal memory is set up index.
12., it is characterized in that described matching module comprises according to the described system of claim 11:
The inquiry extraction unit is used for utilizing search index and reading the sample characteristics amount;
Matching unit is used for sample characteristics amount and described checking characteristic quantity coupling.
13., it is characterized in that described characteristic quantity modular converter comprises according to the described system of claim 9:
The frame extraction unit is used for extracting a frame 3 D monitoring image as first monitoring image from three-dimensional buffer zone, extracts a back frame 3 D monitoring image of first monitoring image as second monitoring image;
The contrast unit is used for contrast first monitoring image and second monitoring image, judges whether the two exists the modified-image zone;
Converting unit is used for not having the modified-image zone, then utilizes default algorithm that full detail in second monitoring image is converted to the first checking characteristic quantity; Or have the modified-image zone, then utilize default algorithm just the modified-image area information be converted to the second checking characteristic quantity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310299523.6A CN103345625B (en) | 2013-07-16 | 2013-07-16 | A kind of three-dimensional image analysis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310299523.6A CN103345625B (en) | 2013-07-16 | 2013-07-16 | A kind of three-dimensional image analysis method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103345625A true CN103345625A (en) | 2013-10-09 |
| CN103345625B CN103345625B (en) | 2016-08-17 |
Family
ID=49280420
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310299523.6A Active CN103345625B (en) | 2013-07-16 | 2013-07-16 | A kind of three-dimensional image analysis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103345625B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104657708A (en) * | 2015-02-02 | 2015-05-27 | 郑州酷派电子设备有限公司 | Novel device and method for identifying three-dimensional object |
| CN112053391A (en) * | 2020-09-11 | 2020-12-08 | 中德（珠海）人工智能研究院有限公司 | Monitoring and early warning method and system based on dynamic three-dimensional model and storage medium |
| WO2021136499A1 (en) * | 2020-01-05 | 2021-07-08 | 于毅欣 | Dynamic object sensing method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201282529Y (en) * | 2008-10-27 | 2009-07-29 | 汉达尔通信技术(北京)有限公司 | Intelligent monitoring system |
| CN102122390A (en) * | 2011-01-25 | 2011-07-13 | 于仕琪 | Method for detecting human body based on range image |
| CN102196251A (en) * | 2011-05-24 | 2011-09-21 | 中国科学院深圳先进技术研究院 | Smart-city intelligent monitoring method and system |
| CN102622591A (en) * | 2012-01-12 | 2012-08-01 | 北京理工大学 | 3D (three-dimensional) human posture capturing and simulating system |
-
2013
- 2013-07-16 CN CN201310299523.6A patent/CN103345625B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201282529Y (en) * | 2008-10-27 | 2009-07-29 | 汉达尔通信技术(北京)有限公司 | Intelligent monitoring system |
| CN102122390A (en) * | 2011-01-25 | 2011-07-13 | 于仕琪 | Method for detecting human body based on range image |
| CN102196251A (en) * | 2011-05-24 | 2011-09-21 | 中国科学院深圳先进技术研究院 | Smart-city intelligent monitoring method and system |
| CN102622591A (en) * | 2012-01-12 | 2012-08-01 | 北京理工大学 | 3D (three-dimensional) human posture capturing and simulating system |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104657708A (en) * | 2015-02-02 | 2015-05-27 | 郑州酷派电子设备有限公司 | Novel device and method for identifying three-dimensional object |
| WO2021136499A1 (en) * | 2020-01-05 | 2021-07-08 | 于毅欣 | Dynamic object sensing method and device |
| CN112053391A (en) * | 2020-09-11 | 2020-12-08 | 中德（珠海）人工智能研究院有限公司 | Monitoring and early warning method and system based on dynamic three-dimensional model and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103345625B (en) | 2016-08-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Huang et al. | Detection algorithm of safety helmet wearing based on deep learning | |
| CN111047818A (en) | Forest fire early warning system based on video image | |
| CN109829443A (en) | Video behavior recognition methods based on image enhancement Yu 3D convolutional neural networks | |
| CN105426820B (en) | More people's anomaly detection methods based on safety monitoring video data | |
| CN110516529A (en) | It is a kind of that detection method and system are fed based on deep learning image procossing | |
| Li et al. | Improved YOLOv4 network using infrared images for personnel detection in coal mines | |
| CN109543631A (en) | A kind of fire image detection alarm method based on machine learning | |
| CN102495997B (en) | Reading room intelligent management system based on video detection and GIS (geographic information system) image visualization | |
| CN109711318A (en) | A kind of plurality of human faces detection and tracking based on video flowing | |
| CN112487891B (en) | Visual intelligent dynamic identification model construction method applied to electric power operation site | |
| CN110096945B (en) | Indoor monitoring video key frame real-time extraction method based on machine learning | |
| CN113807240A (en) | Intelligent transformer substation personnel dressing monitoring method based on uncooperative face recognition | |
| CN109523041A (en) | Nuclear power station management system | |
| CN112188164A (en) | AI vision-based violation real-time monitoring system and method | |
| CN112348003A (en) | Airplane refueling scene recognition method and system based on deep convolutional neural network | |
| CN103888731A (en) | Structured description device and system for mixed video monitoring by means of gun-type camera and dome camera | |
| CN115457446A (en) | Abnormal behavior supervision system based on video analysis | |
| CN112906453A (en) | Hand washing monitoring system capable of achieving intelligent identification and real-time feedback | |
| CN107133592A (en) | Electricity substation is using infrared thermal imaging and the human body target feature detection algorithm of visual light imaging technological incorporation | |
| CN114926778A (en) | Safety helmet and personnel identity recognition system under production environment | |
| CN114863489A (en) | Movable intelligent construction site auxiliary inspection method and system based on virtual reality | |
| CN106331646A (en) | Artificial intelligence safety monitoring system | |
| CN103345625A (en) | Method and system for analyzing three-dimensional image | |
| CN114067396A (en) | Vision learning-based digital management system and method for live-in project field test | |
| CN105095891A (en) | Human face capturing method, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |