CN102831405B - Method and system for outdoor large-scale object identification on basis of distributed and brute-force matching - Google Patents
Method and system for outdoor large-scale object identification on basis of distributed and brute-force matching Download PDFInfo
- Publication number
- CN102831405B CN102831405B CN201210292640.5A CN201210292640A CN102831405B CN 102831405 B CN102831405 B CN 102831405B CN 201210292640 A CN201210292640 A CN 201210292640A CN 102831405 B CN102831405 B CN 102831405B
- Authority
- CN
- China
- Prior art keywords
- image
- feature vectors
- identified
- matching
- descriptor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a system for outdoor large-scale object identification on the basis of distributed and brute-force matching. The method includes: by a terminal, acquiring images and GPS (global position system) information, to be identified, of current scenes, extracting local features of the images to be identified, converting the local features into descriptor feature vectors, packing GPS information of the images to be identified and the descriptor feature vectors into a descriptor file, and sending the descriptor file to a distributed processing system; and inquiring current processed matching tasks to judge whether identical matching tasks exist in the current processed matching tasks or not by the distributed processing system according to the GPS information, if no identical matching tasks exist, then searching sample feature vectors of sample images related to the images to be identified in a database, sharing and loading the sample feature vectors to multiple computing nodes to realize matching, and finally determining that the sample images with the highest matching rate is matching images and feeding the matching images to the terminal. By the method and the system, large-scale image identification and retrieval of the intelligent terminal can be realized, data calculation is lowered, and image identification and retrieval speed is increased.
Description
Technical field
The invention belongs to mobile augmented reality technical field, be specifically related to a kind of outdoor extensive object identification method and system based on distributed treatment and violence coupling.
Background technology
The essence of object identification is exactly to set up a computing system that can identify attention object classification in image, and the demand that has a wide range of applications in actual life, has quite high using value and Research Significance.
Along with the development of internet, the mankind are stepping into an informationalized society, and internet has become that the mankind issue, obtain, the Important Platform of exchange message.The exponential growth of internet information amount, makes how to allow user in the data of magnanimity, find rapidly and accurately its information needed to become an important problem.In the last few years, along with the progress of digital photography and memory device and universal, the growth at full speed on the internet of outdoor amount of images, had also reached thousands of.How effectively utilizing these data messages, these existing information is provided to the current user in same position, is an important research direction of business circles and academia.Yet, along with the very big growth of image library scale, guarantee the real-time of picture search, corresponding database index technology and image retrieval technologies also must be done corresponding adjustment or acceleration.
The fast development of simultaneous computer soft and hardware technology, for augmented reality is walked out indoor application and then is supported complicated analysis, decision-making and management to lay a solid foundation.The function of some mobile terminal devices (as PDA, smart mobile phone etc.) is also more and more abundanter, and has had embedded OS, touch-screen, GPS location, the first-class function of video camera, has also possessed stronger calculating and processing power simultaneously.The augmented reality system of exploitation based on mobile terminal that be integrated into of these functions laid a good foundation.According to interrelated data, by the end of China mobile phone user in 2010, can reach 7.4 hundred million, the user who wherein has smart mobile phone has accounted for suitable proportion, and smart mobile phone will have very large application potential as the application platform of augmented reality.Progressively opening, moving of 3G net, means the beginning in brand-new epoch of mobile value-added service, and augmented reality and LBS combine and can realize real-time, interactive, the dynamic three-dimensional display of information, can make man-machine interface more friendly and have intelligent.
Based on above-mentioned analysis, in conjunction with thering is the terminal of camera and the picture search of server end and matching technique, the ONLINE RECOGNITION of extensive object can be become to possibility, and can greatly shorten the ONLINE RECOGNITION time in conjunction with distributed proccessing.
Summary of the invention
In view of this, the invention provides a kind of outdoor extensive object identification scheme based on distributed treatment and violence coupling, this scheme combines distributed proccessing and makes intelligent terminal possess the recognition function of Large Scale Graphs picture with computer vision technique, realize extensive image recognition and the retrieval of intelligent terminal, and then realize the multiple application of mobile augmented reality.And the present invention utilizes GPS information to dwindle Data Matching scope, has reduced data operation quantity, thereby has further improved the speed of image recognition and retrieval, realized online real-time outdoor extensive object identification.
This scheme is achieved in that
First a kind of outdoor extensive object identification method based on distributed treatment and violence coupling is provided, obtain in advance the sample image with GPS information, extract the local feature of every width sample image and be converted to descriptor proper vector, be called sampling feature vectors, by GPS information, sample image information and sampling feature vectors corresponding stored in sample file system;
Described recognition methods comprises the steps:
Step 1: terminal gathers image to be identified and the GPS information of current scene;
Step 2: extract the local feature of described image to be identified, and be converted into descriptor proper vector;
Step 3: the GPS information of image to be identified and descriptor proper vector are packaged into a descriptor file, send to distributed processing system(DPS);
Step 4: be provided with dispatch deal cluster, a plurality of computing node and described sample file system in distributed processing system(DPS); Dispatch deal cluster receives after described descriptor file, from this descriptor file, extract GPS information, in dispatch deal cluster, inquire about and in the current matching task of processing, whether have the matching task identical with the GPS information of extracting, if, illustrate the sampling feature vectors with image correlation to be identified is loaded in each computing node, descriptor file is sent to each computing node, then perform step 6; Otherwise, execution step 5;
The query criteria of described matching task is: if the GPS information of GPS information corresponding to the matching task processed current and image to be identified is consistent or differ a predetermined threshold value, thinks and has identical matching task;
Step 5: dispatch deal cluster is that image to be identified loads relevant sampling feature vectors:
The GPS information of image to be identified of take is the center of circle, according to predefined screening radius, determine a border circular areas, from sample file system, filter out the sampling feature vectors of GPS information in described border circular areas, the sampling feature vectors filtering out is shared and is loaded in each computing node; Meanwhile, dispatch deal cluster also sends to descriptor file each computing node;
Step 6: each computing node parses descriptor proper vector to be matched from descriptor file, by the descriptor proper vector to be matched of image to be identified, be violence coupling with loaded coupling one by one with sampling feature vectors this image correlation to be identified, matching result is aggregated into dispatch deal cluster;
Step 7: dispatch deal cluster is added up the matching result of each computing node, selects the highest sample image of coupling ratio as matching image, and the information of matching image is returned to described terminal.
The present invention also provides a kind of outdoor extensive object identification system based on distributed treatment and violence coupling, comprises distributed processing system(DPS), wireless network and has image acquisition and the terminal of GPS positioning function; Distributed processing system(DPS) comprises switching equipment, dispatch deal cluster, at least 2 computing nodes and sample file system; Dispatch deal cluster is by switching equipment access of radio network, and is connected with sample file system, all computing node;
Described sample file system, the pre-stored descriptor proper vector that has sample image and every width sample image of all kinds of scenes, is called sampling feature vectors, also has the GPS information of sample image; The GPS information of sample image, sample image information and sampling feature vectors corresponding stored;
Described terminal, for gathering image to be identified and the GPS information of current scene, extracts the local feature of described image to be identified, and is converted to descriptor proper vector; The GPS information of image to be identified and descriptor proper vector are packaged into a descriptor file, send to distributed processing system(DPS);
Described dispatch deal cluster, for after receiving described descriptor file, from this descriptor file, extract GPS information, inquire about and in the current matching task of processing, whether have the matching task identical with the GPS information of extracting, if, illustrate the sampling feature vectors with image correlation to be identified is loaded in each computing node, descriptor file is sent to each computing node; Otherwise, for image to be identified loads relevant sampling feature vectors;
Wherein, the query criteria of described matching task is: if the GPS information of GPS information corresponding to the matching task processed current and image to be identified is consistent or differ a predetermined threshold value, thinks and has identical matching task;
Describedly load relevant sampling feature vectors and be for image to be identified: the GPS information of image to be identified of take is the center of circle, according to predefined screening radius, determine a border circular areas, from sample file system, filter out the sampling feature vectors of GPS information in described border circular areas, the sampling feature vectors filtering out is shared and is loaded in each computing node; Meanwhile, dispatch deal cluster also sends to descriptor file each computing node;
Described computing node, for parsing descriptor proper vector to be matched from descriptor file, by the descriptor proper vector to be matched of image to be identified, be violence coupling with loaded coupling one by one with sampling feature vectors this image correlation to be identified, matching result is aggregated into dispatch deal cluster;
Described dispatch deal cluster is further used for, and adds up the matching result of each computing node, selects the highest sample image of coupling ratio as matching image, and the information of matching image is returned to described terminal.
Beneficial effect:
The present invention is based on the large-scale object identification method of distributed system, use ripe distributed structure/architecture, the augmented reality of realization based on vision, can use the image of smart machine Real-time Obtaining current scene, by image being extracted to local feature and carrying out violence coupling, identify real-time and accurately current scene, and then for providing further additional information to lay a good foundation, can expand the interactive application of intelligent terminal, met tourism, navigation, traffic, the expanded application on intelligent terminal such as hotel service, make Virtual network operator and content supplier can utilize its abundant server resource and superior server performance to develop its business.
The present invention, when loading sampling feature vectors for current task, judges whether, by similar or identical task, if had, not need again to load according to GPS information, and the sampling feature vectors that directly employing has loaded, has saved the time of loading data.
And when loading sampling feature vectors, only load the sampling feature vectors in certain limit, can further dwindle operand like this.
Accompanying drawing explanation
Fig. 1 the present invention is based on the extensive object identification schematic diagram that distributed system is mated with violence;
Fig. 2 (a) is the schematic diagram of destructuring storage;
The schematic diagram that Fig. 2 (b) is structured storage;
Fig. 3 is the composition structural representation of recognition system of the present invention.
Embodiment
Below in conjunction with accompanying drawing, object identification analytical approach of the present invention is described in detail.
Fig. 1 shows the schematic diagram of augmented reality (AR) object identification method based on distributed system and violence coupling of the present invention.As shown in Figure 1, the system that realizes the method comprises distributed processing system(DPS), wireless network, has the terminal of image acquisition and GPS positioning function.Terminal and distributed processing system(DPS) are by wireless network intercommunication.Distributed processing system(DPS) specifically comprises switching equipment (gateway), dispatch deal cluster, a plurality of computing node and sample file system.Dispatch deal cluster is by gateway accessing wireless network, and is connected with sample file system, all computing node.
Based on said system, specific implementation flow process of the present invention is as follows:
Preliminary work before identification: obtain in advance the sample image with GPS information, in general can obtain image by scene, for example, from network or shooting on the spot, each scene is obtained a few width sample images from different perspectives, and the GPS information of scene is exactly the GPS information of sample image.Extract the local feature of every width sample image and be converted to descriptor proper vector, being called sampling feature vectors, by GPS information, sample image information and sampling feature vectors corresponding stored in sample file system.
Identifying is as follows:
Step 1: the capture apparatus that user opens a terminal, terminal gathers the image to be identified of current scene; Call again the GPS sensor interface of terminal, obtain the GPS information of current location.
Step 2: in order to reduce operand, terminal is carried out down-sampled processing (reducing resolution) to the image to be identified collecting, to reduce image resolution ratio; And then adopt local feature detection algorithm (SIFT, SURF or ORB) to carry out feature point detection, and to extract the local feature of described scene image, then show by the form of proper vector, be called descriptor proper vector.Piece image may have up to a hundred descriptor proper vectors.
Step 3: by the GPS information of image to be identified and all descriptor proper vector and proper vector number are packaged into a descriptor file, GPS information and proper vector number are placed on the beginning of file, and proper vector number is convenient to receiving end and is judged whether a descriptor file receives.Then descriptor file is sent to distributed processing system(DPS) by wireless network.
Step 4: the dispatch deal cluster in distributed processing system(DPS) receives after descriptor file by gateway, from this descriptor file, extract GPS information, in dispatch deal cluster, inquire about and in the current matching task of processing, whether have the matching task identical with the GPS information of extracting, if, illustrate the sampling feature vectors with image correlation to be identified is loaded in each computing node, descriptor file is sent to each computing node, then perform step 6; Otherwise, execution step 5.
The query criteria of matching task is: if the GPS information of image GPS information corresponding to the matching task processed current and image to be identified is consistent or differ a predetermined threshold value, thinks and has identical matching task.Wherein, the current matching task of processing can be buffered in a dispatch list so that inquiry.
Step 5: dispatch deal cluster is that image to be identified loads relevant sampling feature vectors.Be specially:
The GPS information of image to be identified of take is the center of circle, according to predefined screening radius, determine a border circular areas, from sample file system, filter out the sampling feature vectors of GPS information in described border circular areas, the sampling feature vectors filtering out is shared and is loaded in each computing node; Meanwhile, dispatch deal cluster also sends to descriptor file each computing node.Wherein, screening scope can be 10 meters, 20 meters, 30 meters, 40 meters or 50 meters.
Step 6: each computing node parses descriptor proper vector to be matched from descriptor file, it is violence coupling that the descriptor proper vector to be matched of image to be identified and the sampling feature vectors of loading of this image correlation to be identified are mated one by one, and matching result is aggregated into dispatch deal cluster.
Wherein, " with the sampling feature vectors of loading of this image correlation to be identified " described here is exactly the sampling feature vectors of GPS information in border circular areas, and they participate in the violence coupling of image to be identified.For the image to be identified that finds identical match task in step 4, which sampling feature vectors participates in its violence coupling and can according to the GPS information that loads sample vector, again be judged by computing node, and its judgement calculated amount is very little, can not affect computing velocity; Certainly, also can inform which sampling feature vectors of computing node participates in violence and mates by dispatch deal cluster.
Step 7: dispatch deal cluster is added up the matching result of each computing node, selects the highest sample image of coupling ratio as matching image, and the information of matching image is returned to described terminal.After this, terminal can show above-mentioned recognition result, and user can click above-mentioned classification results, checks details.
Violence coupling is exactly by exhaustive method, and all descriptors of proper vector to be matched and sample vector are contrasted one by one, and the distance of compute vector, asks for nearest proper vector.Conventionally judging whether two unique points mate, is by setting an overall threshold value, and when feature, describing apart from two unique points that are less than or equal to this threshold value is match point, and experiment shows that this method can produce a large amount of mistake couplings.A kind of better method is to calculate arest neighbors to determine with time neighbour's ratio whether these two unique points mate, if this ratio is less than certain threshold value, represent two Feature Points Matching, and the method that experimental results show that is very effective.The present invention also can this kind of characteristic point matching method.Can be described in detail respectively below.
So far, this flow process finishes.
In above-mentioned flow process, the matching image information of returning to terminal can be image I D, can be also image itself or other information relevant with matching image.If what return is image itself, also need to be in sample file system storing sample image.So as shown in Figure 1, in sample file system, set up image library and feature database.For the sample image with GPS information of each scene collection can be stored in image library, image library is recorded the GPS information of sample image, sample image ID and sample image collecting location.And be stored in feature database according to the sampling feature vectors that sample image extracts, feature database is recorded sampling feature vectors that sample image ID, sample image are corresponding and the GPS information of sample image collecting location.
In practice, the feature database in sample file system can adopt structuring or destructuring mode to store.Introduce this two kinds of storage modes below, and the preferably proper vector matching scheme designing for every kind of storage mode.
(1) adopt destructuring mode storing sample proper vector:
As shown in Figure 2 (a) shows, adopt blocks of files organising data, the corresponding scene of each description document piece, the sampling feature vectors relevant to this scene is all stored in same description document piece.Referring to accompanying drawing, description document piece is stored the GPS information of corresponding scene, the sample image quantity of obtaining for this scene, the ID of each sample image, sampling feature vectors and sampling feature vectors quantity that each sample image is corresponding.
Due to description document Kuai Wei unit storage proper vector, and the read operation of description document piece needs designed, designed, and speed is slower, is therefore written into a description document piece at every turn, mated that to be written into another description document piece proper again.Step 6 so of the present invention and 7 preferred process mode are:
1. the descriptor file that the description document piece that each computing node reception is written into and terminal are sent parses descriptor proper vector to be matched from descriptor file; Take sample image as unit, by descriptor proper vector to be matched, mate with all sampling feature vectors of single sample image one by one, statistical match rate, the proper vector quantity that the match is successful accounts for the ratio of all sampling feature vectors of single sample image, and matching rate is aggregated into dispatch deal cluster; And then be written into a description document piece, matching rate is aggregated into dispatch deal cluster, until handle all associated documents pieces.
2. the maximum matching rate of searching the matching rate that dispatch deal cluster feeds back from all computing nodes, the sample image that maximum matching rate is corresponding is exactly matching image;
3. the information of matching image is returned to described terminal.
(2) adopt structured way storing sample proper vector:
As shown in Fig. 2 (b), sampling feature vectors is with the form storage of record.A sampling feature vectors of each record storage; The form of each record is identical.For structured storage, preferably load mode is for to carry out rapid loading according to index.Therefore, the content of every record comprises: draw ID, sample image ID, GPS information, sampling feature vectors sequence number, sampling feature vectors.
Wherein the building mode of index ID is routine techniques means, the preparatory stage before identification, in sample file system, store concordance list, and this concordance list has been recorded GPS information and has been recorded the index relative of memory location.According to GPS information searching concordance list, just GPS information can be loaded in each computing node all contents that record in border circular areas so.When setting up index, also can adopt a plurality of computing nodes to carry out Distributed Calculation, to save computing time; In like manner the image in image library being carried out to feature extraction also can adopt the mode of Distributed Calculation to realize.
Owing to storing proper vector with the unit of being recorded as, and set up concordance list, therefore can adopt existing database technology to read fast and record and be loaded in computing node, speed, therefore can directly all be loaded into the record of all and image correlation to be matched in computing node.Step 6 so of the present invention and 7 preferred process mode are:
1. each computing node parses descriptor proper vector to be matched from descriptor file, look for a descriptor proper vector x to be matched to mate one by one with all sampling feature vectors that are loaded into self, find the sampling feature vectors A nearest with descriptor proper vector x and time near sampling feature vectors B, and be aggregated into dispatch deal cluster and carry out integral body sequence, find the sampling feature vectors A ' nearest with descriptor proper vector x and time near sampling feature vectors B ', calculate the ratio of A ' and B ', if this ratio is less than predetermined threshold value, determine that sampling feature vectors A ' is the match point of descriptor proper vector x,
2. for each descriptor proper vector to be matched, all carry out aforesaid operations, then dispatch deal cluster, for each sample image, calculates the ratio that match point accounts for this sample image, and the sample image that ratio is the highest is exactly matching image;
3. the information of matching image is returned to described terminal.
In fact, be no matter structured storage or destructuring storage, above two kinds of matching processs are all applicable, just comparatively speaking, the preferred the latter of structured storage, destructuring storage preferably the former.
In distributed system the inside, unexpected except carrying out Distributed Calculation between each computing unit, parallel computation can also be carried out in computing unit the inside.The descriptor of such 1000 width sample images, for example use 10 computing units, each computing unit is born the matching task of 100 samples, each computing unit is opened up the thread of 10 parallel computations, each thread is born the matching task of 10 sample images, the time that only needs 10 graphic violence couplings when like this 1000 sample images are mated, conventionally at the CPU2.8G of four cores, inside save as on the machine of 8G, image resolution ratio is 320 * 240, with SIFT descriptor, 10 sample images are mated, elapsed time is 126ms left and right.So, by distributed system, Large Scale Graphs looking like to mate and can reach good real-time, the mode of coupling is carried out unique point arest neighbors and is searched by force simultaneously, can realize higher matching precision.
In above-mentioned flow process, relate to intelligent terminal capture video image, process and show returning results of sending from the distributed system server of network-side, these steps are all extensively implemented on current intelligent terminal, these existing embodiments all can be used for the present invention, and video image of the present invention is taken, processed and show that the data from network server end are not limited to existing mode.
Known based on said method, the function that respectively forms module in outdoor extensive object identification system provided by the invention is:
Described sample file system, the pre-stored descriptor proper vector that has sample image and every width sample image of all kinds of scenes, is called sampling feature vectors, also has the GPS information of sample image; The GPS information of sample image, sample image information and sampling feature vectors corresponding stored;
Described terminal, for gathering image to be identified and the GPS information of current scene, extracts the local feature of described image to be identified, and is converted to descriptor proper vector; The GPS information of image to be identified and descriptor proper vector are packaged into a descriptor file, send to distributed processing system(DPS);
Dispatch deal cluster, for after receiving described descriptor file, from this descriptor file, extract GPS information, according to whether having the matching task identical with the GPS information of extracting in the current matching task of processing of aforementioned query criteria inquiry, if, illustrate the sampling feature vectors with image correlation to be identified is loaded in each computing node, descriptor file is sent to each computing node; Otherwise, for image to be identified loads relevant sampling feature vectors; Wherein, the concrete mode that loads relevant sampling feature vectors for image to be identified is identical with the description in method flow, does not repeat here.
Computing node, for parsing descriptor proper vector to be matched from descriptor file, it is violence coupling that the descriptor proper vector to be matched of image to be identified is mated one by one with the sampling feature vectors loading for this image to be identified, and matching result is aggregated into dispatch deal cluster;
Dispatch deal cluster is further used for, and adds up the matching result of each computing node, selects the highest sample image of coupling ratio as matching image, and the information of matching image is returned to described terminal.
Preferably, terminal, before extracting the local feature of described image to be identified, is further carried out down-sampled processing to described image to be identified, to reduce image resolution ratio.
In sum, these are only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any modification of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.
Claims (8)
1. the outdoor extensive object identification method based on distributed treatment and violence coupling, it is characterized in that, obtain in advance the sample image with GPS information, extract the local feature of every width sample image and be converted to descriptor proper vector, be called sampling feature vectors, by GPS information, sample image information and sampling feature vectors corresponding stored in sample file system;
Described recognition methods comprises the steps:
Step 1: terminal gathers image to be identified and the GPS information of current scene;
Step 2: extract the local feature of described image to be identified, and be converted into descriptor proper vector;
Step 3: the GPS information of image to be identified and descriptor proper vector are packaged into a descriptor file, send to distributed processing system(DPS);
Step 4: be provided with dispatch deal cluster, a plurality of computing node and described sample file system in distributed processing system(DPS); Dispatch deal cluster receives after described descriptor file, from this descriptor file, extract GPS information, in dispatch deal cluster, inquire about and in the current matching task of processing, whether have the matching task identical with the GPS information of extracting, if, illustrate the sampling feature vectors with image correlation to be identified is loaded in each computing node, descriptor file is sent to each computing node, then perform step 6; Otherwise, execution step 5;
The query criteria of described matching task is: if the GPS information of GPS information corresponding to the matching task processed current and image to be identified is consistent or differ a predetermined threshold value, thinks and has identical matching task;
Step 5: dispatch deal cluster is that image to be identified loads relevant sampling feature vectors:
The GPS information of image to be identified of take is the center of circle, according to predefined screening radius, determine a border circular areas, from sample file system, filter out the sampling feature vectors of GPS information in described border circular areas, the sampling feature vectors filtering out is shared and is loaded in each computing node; Meanwhile, dispatch deal cluster also sends to descriptor file each computing node;
Step 6: each computing node parses descriptor proper vector to be matched from descriptor file, by the descriptor proper vector to be matched of image to be identified, be violence coupling with loaded coupling one by one with sampling feature vectors this image correlation to be identified, matching result is aggregated into dispatch deal cluster;
Step 7: dispatch deal cluster is added up the matching result of each computing node, selects the highest sample image of coupling ratio as matching image, and the information of matching image is returned to described terminal.
2. the method for claim 1, is characterized in that, in step 2, before extracting the local feature of described image to be identified, further described image to be identified is carried out to down-sampled processing, to reduce image resolution ratio.
3. the method for claim 1, is characterized in that, described step 6 and step 7 are specially:
1. each computing node parses descriptor proper vector to be matched from descriptor file, look for a descriptor proper vector x to be matched to mate one by one with all sampling feature vectors that are loaded into self, find the sampling feature vectors A nearest with descriptor proper vector x and time near sampling feature vectors B, and be aggregated into dispatch deal cluster and carry out integral body sequence, find the sampling feature vectors A ' nearest with descriptor proper vector x and time near sampling feature vectors B ', calculate the ratio of A ' and B ', if this ratio is less than predetermined threshold value, determine that sampling feature vectors A ' is the match point of descriptor proper vector x,
2. for each descriptor proper vector to be matched, all carry out aforesaid operations, then dispatch deal cluster, for each sample image, calculates the ratio that match point accounts for the sample bit string of this sample image, and the sample image that ratio is the highest is exactly matching image;
3. the information of matching image is returned to described terminal.
4. the method for claim 1, is characterized in that, described step 6 and step 7 are specially:
1. each computing node parses descriptor proper vector to be matched from descriptor file, take sample image as unit, by descriptor proper vector to be matched, mate with all sampling feature vectors of single sample image one by one, statistical match rate, the proper vector quantity that the match is successful accounts for the ratio of all sampling feature vectors of single sample image, and matching rate is aggregated into dispatch deal cluster;
2. dispatch deal cluster is found maximum matching rate from the matching rate of all computing node feedbacks, and the sample image that maximum matching rate is corresponding is exactly matching image;
3. the information of matching image is returned to described terminal.
5. the method as described in claim 1 or 3, is characterized in that, adopts structured storage mode storing sample proper vector, the corresponding record of each sampling feature vectors; The field of each record comprises index ID, sample image ID, GPS information, sampling feature vectors sequence number, sampling feature vectors;
Further in sample file system, store concordance list, this concordance list has been recorded GPS information and has been recorded the index relative of memory location;
While loading relevant sampling feature vectors for image to be identified in described step 5, according to GPS information searching concordance list, the content that records by GPS information in described border circular areas is loaded in each computing node.
6. the method for claim 1, is characterized in that, opens up a plurality of parallel computation threads in each computing node, and each thread is born identical task amount.
7. the outdoor extensive object identification system based on distributed treatment and violence coupling, is characterized in that, comprises distributed processing system(DPS), wireless network and has image acquisition and the terminal of GPS positioning function; Distributed processing system(DPS) comprises switching equipment, dispatch deal cluster, at least 2 computing nodes and sample file system; Dispatch deal cluster is by switching equipment access of radio network, and is connected with sample file system, all computing node;
Described sample file system, the pre-stored descriptor proper vector that has sample image and every width sample image of all kinds of scenes, is called sampling feature vectors, also has the GPS information of sample image; The GPS information of sample image, sample image information and sampling feature vectors corresponding stored;
Described terminal, for gathering image to be identified and the GPS information of current scene, extracts the local feature of described image to be identified, and is converted to descriptor proper vector; The GPS information of image to be identified and descriptor proper vector are packaged into a descriptor file, send to distributed processing system(DPS);
Described dispatch deal cluster, for after receiving described descriptor file, from this descriptor file, extract GPS information, inquire about and in the current matching task of processing, whether have the matching task identical with the GPS information of extracting, if, illustrate the sampling feature vectors with image correlation to be identified is loaded in each computing node, descriptor file is sent to each computing node; Otherwise, for image to be identified loads relevant sampling feature vectors;
Wherein, the query criteria of described matching task is: if the GPS information of GPS information corresponding to the matching task processed current and image to be identified is consistent or differ a predetermined threshold value, thinks and has identical matching task;
Describedly load relevant sampling feature vectors and be for image to be identified: the GPS information of image to be identified of take is the center of circle, according to predefined screening radius, determine a border circular areas, from sample file system, filter out the sampling feature vectors of GPS information in described border circular areas, the sampling feature vectors filtering out is shared and is loaded in each computing node; Meanwhile, dispatch deal cluster also sends to descriptor file each computing node;
Described computing node, for parsing descriptor proper vector to be matched from descriptor file, by the descriptor proper vector to be matched of image to be identified, be violence coupling with loaded coupling one by one with sampling feature vectors this image correlation to be identified, matching result is aggregated into dispatch deal cluster;
Described dispatch deal cluster is further used for, and adds up the matching result of each computing node, selects the highest sample image of coupling ratio as matching image, and the information of matching image is returned to described terminal.
8. object identification system as claimed in claim 7, is characterized in that, described terminal is further used for, and before extracting the local feature of described image to be identified, further described image to be identified is carried out to down-sampled processing, to reduce image resolution ratio.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210292640.5A CN102831405B (en) | 2012-08-16 | 2012-08-16 | Method and system for outdoor large-scale object identification on basis of distributed and brute-force matching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210292640.5A CN102831405B (en) | 2012-08-16 | 2012-08-16 | Method and system for outdoor large-scale object identification on basis of distributed and brute-force matching |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102831405A CN102831405A (en) | 2012-12-19 |
| CN102831405B true CN102831405B (en) | 2014-11-26 |
Family
ID=47334533
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210292640.5A Active CN102831405B (en) | 2012-08-16 | 2012-08-16 | Method and system for outdoor large-scale object identification on basis of distributed and brute-force matching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102831405B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103514446B (en) * | 2013-10-16 | 2016-08-24 | 北京理工大学 | A kind of outdoor scene recognition methods merging sensor information |
| CN103530377B (en) * | 2013-10-16 | 2016-08-24 | 北京理工大学 | A kind of scene information searching method based on binary features code |
| CN103678610A (en) * | 2013-12-16 | 2014-03-26 | 浙江大学 | Method for recognizing object based on intelligent mobile phone sensor |
| CN106295657A (en) * | 2016-08-04 | 2017-01-04 | 天津汉光祥云信息科技有限公司 | A kind of method extracting human height's feature during video data structure |
| CN106874445A (en) * | 2017-02-10 | 2017-06-20 | 深圳前海大造科技有限公司 | High in the clouds image-recognizing method based on words tree retrieval with similarity checking |
| CN106844725A (en) * | 2017-02-10 | 2017-06-13 | 深圳前海大造科技有限公司 | A kind of high in the clouds image data base generation and recognition methods |
| CN106844726B (en) * | 2017-02-10 | 2020-11-10 | 成都弥知科技有限公司 | Image recognition method based on vocabulary tree retrieval and violence matching |
| CN107133260A (en) * | 2017-03-22 | 2017-09-05 | 新奥特（北京）视频技术有限公司 | The matching and recognition method and device of a kind of landmark image |
| CN107133261A (en) * | 2017-03-22 | 2017-09-05 | 新奥特（北京）视频技术有限公司 | The input method and device of a kind of landmark information |
| CN108509870B (en) * | 2018-03-14 | 2019-07-12 | 安徽工大信息技术有限公司 | A kind of Eriocheir sinensis uniqueness recognition methods based on images match |
| CN108512586B (en) * | 2018-04-17 | 2020-08-04 | Oppo广东移动通信有限公司 | Signal processing method and device, electronic equipment and computer readable storage medium |
| CN111046879B (en) * | 2019-10-15 | 2023-09-29 | 平安科技（深圳）有限公司 | Certificate image classification method, device, computer equipment and readable storage medium |
| CN111401482B (en) * | 2020-04-29 | 2024-03-19 | Oppo广东移动通信有限公司 | Feature point matching method and device, equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009146719A1 (en) * | 2008-06-03 | 2009-12-10 | Eth Zurich | Method and system for generating a pictorial reference database using geographical information |
| CN102254345A (en) * | 2011-06-30 | 2011-11-23 | 上海大学 | Method for registering natural characteristic based on cloud computation |
| WO2012053118A1 (en) * | 2010-10-22 | 2012-04-26 | Telefonaktiebolaget L M Ericsson (Publ) | Image matching apparatus and image matching method |
| CN102625993A (en) * | 2009-07-30 | 2012-08-01 | Sk普兰尼特有限公司 | Method for providing augmented reality, server for same, and portable terminal |
-
2012
- 2012-08-16 CN CN201210292640.5A patent/CN102831405B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009146719A1 (en) * | 2008-06-03 | 2009-12-10 | Eth Zurich | Method and system for generating a pictorial reference database using geographical information |
| CN102625993A (en) * | 2009-07-30 | 2012-08-01 | Sk普兰尼特有限公司 | Method for providing augmented reality, server for same, and portable terminal |
| WO2012053118A1 (en) * | 2010-10-22 | 2012-04-26 | Telefonaktiebolaget L M Ericsson (Publ) | Image matching apparatus and image matching method |
| CN102254345A (en) * | 2011-06-30 | 2011-11-23 | 上海大学 | Method for registering natural characteristic based on cloud computation |
Non-Patent Citations (4)
| Title |
|---|
| city-scale landmark identification on mobile devices;David M. Chen et al.;《2011 IEEE conference on computer vision and pattern recognition 》;20111231;737-744 * |
| David M. Chen et al..city-scale landmark identification on mobile devices.《2011 IEEE conference on computer vision and pattern recognition 》.2011,737-744. * |
| Eduardo Todt et al..outdoor landmark-view recognition based on bipartite-graph matching and logistic regression.《proceedings of the 2007 ieee international conference on robotics and automation》.2007, * |
| outdoor landmark-view recognition based on bipartite-graph matching and logistic regression;Eduardo Todt et al.;《proceedings of the 2007 ieee international conference on robotics and automation》;20070414;4289-4294 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102831405A (en) | 2012-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102831405B (en) | Method and system for outdoor large-scale object identification on basis of distributed and brute-force matching | |
| CN102880854B (en) | Distributed processing and Hash mapping-based outdoor massive object identification method and system | |
| CN102880879B (en) | Distributed processing and support vector machine (SVM) classifier-based outdoor massive object recognition method and system | |
| US10482122B2 (en) | System and method for multiresolution and multitemporal image search | |
| CN105760469B (en) | Higher-dimension approximation image retrieval method based on the row of falling LSH under cloud computing environment | |
| CN111586091B (en) | Edge computing gateway system for realizing computing power assembly | |
| CN105760468B (en) | The extensive image query system of sensitive hash index is set under mobile environment based on ranking | |
| CN102819752B (en) | System and method for outdoor large-scale object recognition based on distributed inverted files | |
| CN103530649A (en) | Visual searching method applicable mobile terminal | |
| CN102236714A (en) | Extensible markup language (XML)-based interactive application multimedia information retrieval method | |
| CN101790064A (en) | Hard disk video equipment with video structural description function and method thereof | |
| US9519543B2 (en) | Method and apparatus for image search using feature point | |
| CN103530377B (en) | A kind of scene information searching method based on binary features code | |
| CN100462978C (en) | Image searching method and system | |
| CN104486585A (en) | Method and system for managing urban mass surveillance video based on GIS | |
| CN103999097B (en) | System and method for compact descriptor for visual search | |
| CN106407268A (en) | Method and system for content retrieval based on rate-coverage optimization | |
| CN102542041A (en) | Method and system for processing raster data | |
| CN101599077B (en) | Method for retrieving three-dimensional object | |
| CN114637886A (en) | Machine vision system based on multiple protocols | |
| CN102932421A (en) | Cloud back-up method and device | |
| CN105183737A (en) | Industrial Design instant searching system and searching method thereof | |
| CN110633379B (en) | Graphics Processing Unit (GPU) -parallel-operation-based graphics searching system and method | |
| CN112115281A (en) | Data retrieval method, device and storage medium | |
| CN106257447A (en) | The video storage of cloud storage server and search method, video cloud storage 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 |