CN203217590U - Video intelligent gesture recognition system - Google Patents
Video intelligent gesture recognition system Download PDFInfo
- Publication number
- CN203217590U CN203217590U CN 201220331184 CN201220331184U CN203217590U CN 203217590 U CN203217590 U CN 203217590U CN 201220331184 CN201220331184 CN 201220331184 CN 201220331184 U CN201220331184 U CN 201220331184U CN 203217590 U CN203217590 U CN 203217590U
- Authority
- CN
- China
- Prior art keywords
- video
- gesture identification
- video intelligent
- hardware
- intelligent gesture
- 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.)
- Expired - Lifetime
Links
Images
Landscapes
- User Interface Of Digital Computer (AREA)
Abstract
The utility model provides a video intelligent gesture recognition system which comprises the following components: a video acquiring device which is used for acquiring dynamic images in a speed above 25 frames per seconds and converting the dynamic images to a digital video signal; a hardware video data accelerating preprocessing device which is used for performing characteristic data extraction on the digital video signal and is connected with the video acquiring device; and a video intelligent gesture recognition core device which is used for performing gesture motion track analysis according to the characteristic data and obtaining a gesture recognition result and is connected with the hardware video data accelerating preprocessing device. According to the video intelligent gesture recognition system, on condition that 50cm is used as the radius around the detected and tracked human face without increase of any CPU hardware physical resource, gesture recognition precision reaches above 98% when external light is not lower than 5LM and gesture recognition distance is smaller than 4.5 meters, and video intelligent gesture recognition requirement is satisfied.
Description
Technical field
The utility model relates to electronic information field, relates in particular to a kind of video intelligent gesture identification system.
Background technology
For operations such as easy input, controls, be extensive use of touch screen technology in Digital Media and the mobile device at present.Owing to touch-screen production technology and cost reason, very high based on the cost of the host computer system of touch screen technology, particularly the large scale touch-screen system especially a thousand pieces of gold difficulty ask.And the core technology of high-performance capacitor touch-screen is all controlled by overseas enterprise basically, as Samsung, and Sony etc., the cost that this has raised national system merchant manufacturing more and has managed Related product, it is synchronous, fast-developing to have restricted domestic industries.More novel man-machine interaction mode will promote technological innovation and industrial upgrading, also will increase the new chance that China's related industry is obtained the industry development right of speech.
Video intelligent identification field has become the hot market of international each major company's competition, Microsoft releases the system Kinect based on the identification of XBOX360 main body sense video intelligent, its birth meaning fully can with Microsoft(Microsoft) the mentioning in the same breath of the Windows released then.This product has changed traditional human-computer interaction interface fully, has promoted the widespread use of video intelligent recognition technology, has promoted nature, the fast development of this interactive class of people technology and growth.May progressively replace existing expensive capacitance touch screen technology after this type of technology maturation, become the standard input device of man-machine interactions such as digital intelligent TV, set-top box, mobile hand-held device, auto-navigation system, mobile personal medium, intelligentized Furniture and intelligent appliance.
The utility model content
Fundamental purpose of the present utility model is to provide a kind of video intelligent gesture identification system, is not increasing any CPU(central processing unit) situation of hardware physical resource is issued to the requirement of video intelligent gesture identification.
In order to achieve the above object, the utility model provides a kind of video intelligent gesture identification system, comprising:
Be used for the continuous dynamic image of the above speed acquisition of per second 25 frames and be converted into the video acquisition device of digital video signal;
Be used for this digital video signal is carried out the hardware video data acceleration pretreatment unit that characteristic is extracted, be connected with described video acquisition device;
Be used for carrying out the gesture motion trajectory analysis and obtaining gesture identification result's video intelligent gesture identification core apparatus according to this characteristic, accelerate pretreatment unit with described hardware video data and be connected.
During enforcement, video intelligent gesture identification described in the utility model system also comprises the data communication bus device;
Described video intelligent gesture identification core apparatus accelerates pretreatment unit by described data communication bus device and described hardware video data and is connected.
During enforcement, described data communication bus device is the high-speed data communication bus unit.
During enforcement, video intelligent gesture identification described in the utility model system also comprises gesture identification output unit as a result, is connected with described video intelligent gesture identification core apparatus.
During enforcement, described video acquisition device comprises optical filter, lens and the cmos image sensor that is arranged in order, and the signal processor that is connected with this cmos image sensor.
During enforcement, described hardware video data are accelerated pretreatment unit and are adopted hardware-accelerated unit.
During enforcement, described video intelligent gesture identification core apparatus comprises ASIC class chip.
During enforcement, described gesture identification output unit as a result comprises USB interface, SPI interface or SDIO interface.
Compared with prior art, video intelligent gesture identification described in the utility model system centered by the people's face that detects and track with 50CM(centimetre) be radius, under the situation that does not increase any CPU hardware physical resource, be not less than the 5LM(lumen at extraneous light), the gesture identification distance reaches more than 98% less than gesture accuracy of identification under 4.5 meters the situation, has reached the requirement of video intelligent gesture identification.
Description of drawings
Fig. 1 is the structured flowchart of the described video intelligent gesture identification of the utility model first embodiment system;
Fig. 2 is the structured flowchart of the described video intelligent gesture identification of the utility model second embodiment system;
Fig. 2 A is the structured flowchart that the hardware video data of the described video intelligent gesture identification of the utility model second embodiment system are accelerated pretreatment unit 22;
Fig. 2 B is the gesture identification structured flowchart of output unit 25 as a result of the described video intelligent gesture identification of the utility model second embodiment system;
Fig. 3 is the inner structure block diagram for video intelligent gesture identification described in the utility model system specific embodiment.
Embodiment
At first, the related technical term of the utility model is described:
CMOS Image Sensor:CMOS (complementary metal oxide semiconductor (CMOS)) imageing sensor;
ITU601: standard image data transmission data layout;
ISP:Input Signal Processor, signal processor;
BT656: standard serial image data transmission data layout;
ASIC: special IC;
USB: USB (universal serial bus);
SPI: serial bus interface;
SDIO: serial data IO interface;
IP ACC: hardware-accelerated unit;
RAM: random access memory;
RTL: RTL;
FIFO: First Input First Output.
For the purpose, technical scheme and the advantage that make the utility model embodiment is clearer, below in conjunction with embodiment and accompanying drawing, the present utility model is done explanation in further detail.At this, the utility model be used for is explained in schematic enforcement of the present utility model and explanation, but not as to restriction of the present utility model.
It centered by the people's face that detects and track is the high-precision video intelligent gesture identification system of radius with 50CM that the utility model provides a kind of, cooperates high-level efficiency Intelligent Recognition algorithm to reach intelligent gesture identification purpose down at specialized hardware (IP ACC accelerator module).
Can identify gesture at present comprises wave (Wave), a left side (Lift), right (Right), Back stroke (Up), waves (Down), both hands be in step with (Double Wave) down.
As shown in Figure 1, the utility model provides a kind of video intelligent gesture identification system, comprising:
Be used for the continuous dynamic image of the above speed acquisition of per second 25 frames and be converted into the video acquisition device 11 of digital video signal;
Be used for this digital video signal is carried out the hardware video data acceleration pretreatment unit 12 that characteristic is extracted, be connected with described video acquisition device 11;
Be used for carrying out the gesture motion trajectory analysis and obtaining gesture identification result's video intelligent gesture identification core apparatus 13 according to this characteristic, accelerate pretreatment unit 12 with described hardware video data and be connected.
During enforcement, described video acquisition device 11 adopts front end CMOS Image Sensor(imageing sensor) video acquisition device.
Front end CMOS Image Sensor(imageing sensor) research and the design of video acquisition device mainly concentrate on special software and drive exploitation, filter the noise jamming of CMOS Image Sensor output to guarantee software; Filter is selected, and guarantees that images of gestures information is accurate; The design of data-switching form and physical interface guarantees the data transmission minimum delay to greatest extent.
The hardware video data are accelerated special-purpose Hardware I P ACC accelerator module design in the pretreatment unit, this Hardware I P ACC accelerator module is the kernal hardware unit of whole video intelligence gesture identification system, has directly determined gesture identification precision, retardance and the gesture identification dependence to external environment condition.Double Data BUFFER(buffer memory is adopted in the design of specialized hardware IPACC accelerator module) and special-purpose 32 4 passage DMA(direct memory accesss) framework, adopt standard A HB(Advanced High-performance Bus) bus interface standards, greatly increase data and handle and data throughput capabilities, images of gestures sharpening processing, colourity change process, the processing of image convergent-divergent and image feature data abstraction function are mainly finished in the design of specialized hardware IPACC accelerator module.
Gesture Intelligent Recognition algorithm efficiently in the video intelligent gesture identification core apparatus, adopt top-down modularization development model, the moving object dynamic mathematical models that adopt the utilization of blurred picture analytical technology to set up realize the analysis of intelligent gesture motion track, thereby reach the purpose of intelligent gesture identification, the blurred picture analytical technology of gesture Intelligent Recognition algorithm is the foundation of dynamic mathematical models that radius makes up with 50CM centered by the detected central point of people's face.This hardware algorithm dependence is little, carries out the efficient height, can finish the intellectual analysis of dynamic gesture under the cooperation of IP ACC accelerator module fast.Because hardware is not had too high requirement, the feasible cost of the product of this utility model technology that uses is well controlled, and greatly reduces cost, and the industrialization space is very big.
The hardware video data accelerate in the pretreatment unit in the special-purpose Hardware I P ACC accelerator module and video intelligent gesture identification core apparatus efficiently that gesture Intelligent Recognition algorithm cooperates in the utility model, comprise that dynamic image pre-service, dynamic image Algorithm Analysis and communication speed coupling adopt the work of pipelined burst line mode, just can guarantee to be not less than 5LM at extraneous light, the gesture identification distance reaches more than 98% less than gesture accuracy of identification under 4.5 meters the situation.
Referring to Fig. 2, the described video intelligent gesture identification of the utility model second embodiment system comprises: the video acquisition device 21 of Lian Jieing, hardware video data are accelerated pretreatment unit 22, high-speed data communication bus unit 23, video intelligent gesture identification core apparatus 24 and gesture identification output unit 25 as a result successively, wherein:
Described video acquisition device 21, by the requirement of 24 pairs of continuous dynamic images of video of video intelligent gesture identification core apparatus to gather continuous dynamic image more than per second 25 frames and image information changed into ITU601 data layout or BT656 data layout and output;
Described video acquisition device 21 comprises optical filter 211, lens 212 and the cmos image sensor 213 that is arranged in order, and the signal processor 214 that is connected with this cmos image sensor 213;
Described hardware video data are accelerated pretreatment unit 22, adopt hardware mode to realize that the image information data that video acquisition device 21 is exported is carried out the needed characteristic of video intelligent gesture identification core apparatus 24 identifications to be extracted, and the output characteristic data;
Described high-speed data communication bus unit 23 is used for connecting hardware video data acceleration pretreatment unit 22 and video intelligent gesture identification core apparatus 24 to carry out high-speed data communication;
Described video intelligent gesture identification core apparatus 24 carries out the gesture motion trajectory analysis to reach the gesture identification purpose according to the characteristic that hardware video data acceleration pretreatment unit 22 obtains;
Described gesture identification is output unit 25 as a result, exports the result of gesture identification according to the recognition result of video intelligent gesture identification core apparatus 24 to reach the purpose of other device of control.
As shown from the above technical solution, the utility model can be good at being implemented in extraneous light and is not less than 5LM under the situation that does not increase any extra hardware expense, and the gesture identification distance reaches more than 98% less than gesture accuracy of identification under 4.5 meters the situation.
Described video acquisition device 21, the data layout that can use but not limit to output must adopt ITU601 data layout or BT656 data layout, as long as can be to the requirement of the continuous dynamic image of video to gather the output after the line data format conversion of going forward side by side of continuous dynamic image more than per second 25 frames.
Described hardware video data are accelerated pretreatment unit 22, can use but do not limit to realization and view data is carried out characteristic extract, as long as can satisfy the needed characteristic of video intelligent gesture identification core apparatus 24 identifications.
Described high-speed data communication bus unit 23 can use but is not limited to the technical indicator of form, type, bit wide and the bandwidth of data communication bus.The hardware video data are accelerated pretreatment unit 22 and video intelligent gesture identification core apparatus 24 carries out the high-speed data communication demand as long as can satisfy.
Described video intelligent gesture identification core apparatus 24 can use but not limit to the chip type of realizing, comprises ASIC class chip, as long as can satisfy the minimum resource requirement of video intelligent gesture identification core apparatus 24 operations.
Described gesture identification output unit 25 as a result can be used physical hardware interface architectures such as USB, SPI, SDIO.
Fig. 2 A is the structured flowchart that the hardware video data of the described video intelligent gesture identification of the utility model second embodiment system are accelerated pretreatment unit 22;
Fig. 2 B is the gesture identification structured flowchart of output unit 25 as a result of the described video intelligent gesture identification of the utility model second embodiment system;
In Fig. 2 A, RAM FIFO refers to the random access memory First Input First Output, and the RTL kernel refers to the RTL kernel; In Fig. 2 B, FIFO refers to First Input First Output.
Referring to Fig. 3, be the inner structure block diagram of video intelligent gesture identification described in the utility model system specific embodiment, this figure has clearly show the relation between each functional module and Data Stream Processing flow process among the present invention.
Below be the respective explanations of the english abbreviation among Fig. 3:
CPU, central processing unit is mainly finished the operation of Intelligent Recognition software algorithm and computing function;
DDR2, memory interface mainly provides outside DDR2 storer to connect, and result data or other need the data storage function of temporary transient storage when offering system software computing zero;
Video Capture ITU601(Video Capture ITU601), CMOS Sensor(CMOS sensor) interface, the CMOS Sensor that meets the ITU601 standard interface in order to connection are with the image of catching the gesture continuous motion and change into YUV422 standard image data form and offer systematic analysis;
IPACC, hardware-accelerated unit cooperates the software recognizer to reach the purpose of quick identification user gesture, mainly improves system's gesture identification response speed;
USB2.0Device(USB2.0 equipment), standard USB2.0 communication interface provides the external USB communication function;
EDMA, enhancement mode direct memory addressed location strengthens the design of system CPU utilization factor;
Interrupt Controller, interruptable controller, the response interruption is also carried out the interrupt schedule function;
Ethernet MAC, Ethernet media interviews layer protocol;
AHB/APB, the AHB high-speed bus is to the APB(peripheral bus) low speed internal bus conversion bridge unit;
NAND Flash Controller, outside nand flash memory controller provides the outside nand flash memory visit that connects;
Static Memory Controller, the static memory controller unit;
UART, the general serial standard interface;
SDIO, the serial data IO interface provides communication function;
SPI, serial communication interface;
I2S, the universal audio communication interface.
The utility model can be good at being implemented in extraneous light and is not less than 5LM under the situation that does not increase any extra hardware expense, the gesture identification distance reaches more than 98% less than gesture accuracy of identification under 4.5 meters the situation.
More than explanation is just illustrative for the utility model; and it is nonrestrictive; those of ordinary skills understand; under the situation that does not break away from the spirit and scope that claims limit; can make many modifications, variation or equivalence, but all will fall in the protection domain of the present utility model.
Claims (8)
1. a video intelligent gesture identification system is characterized in that, comprising:
Be used for the continuous dynamic image of the above speed acquisition of per second 25 frames and be converted into the video acquisition device of digital video signal;
Be used for this digital video signal is carried out the hardware video data acceleration pretreatment unit that characteristic is extracted, be connected with described video acquisition device;
Be used for carrying out the gesture motion trajectory analysis and obtaining gesture identification result's video intelligent gesture identification core apparatus according to this characteristic, accelerate pretreatment unit with described hardware video data and be connected.
2. video intelligent gesture identification as claimed in claim 1 system is characterized in that, also comprises the data communication bus device;
Described video intelligent gesture identification core apparatus accelerates pretreatment unit by described data communication bus device and described hardware video data and is connected.
3. video intelligent gesture identification as claimed in claim 2 system is characterized in that described data communication bus device is the high-speed data communication bus unit.
4. as the described video intelligent gesture identification of arbitrary claim system in the claim 1 to 3, it is characterized in that, also comprise gesture identification output unit as a result, be connected with described video intelligent gesture identification core apparatus.
5. video intelligent gesture identification as claimed in claim 1 system is characterized in that described video acquisition device comprises optical filter, lens and the cmos image sensor that is arranged in order, and the signal processor that is connected with this cmos image sensor.
6. video intelligent gesture identification as claimed in claim 1 system is characterized in that, described hardware video data are accelerated pretreatment unit and adopted hardware-accelerated unit.
7. video intelligent gesture identification as claimed in claim 1 system is characterized in that described video intelligent gesture identification core apparatus comprises ASIC class chip.
8. video intelligent gesture identification as claimed in claim 1 system is characterized in that described gesture identification output unit as a result comprises USB interface, SPI interface or SDIO interface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201220331184 CN203217590U (en) | 2012-07-10 | 2012-07-10 | Video intelligent gesture recognition system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201220331184 CN203217590U (en) | 2012-07-10 | 2012-07-10 | Video intelligent gesture recognition system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN203217590U true CN203217590U (en) | 2013-09-25 |
Family
ID=49207107
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201220331184 Expired - Lifetime CN203217590U (en) | 2012-07-10 | 2012-07-10 | Video intelligent gesture recognition system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN203217590U (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861640A (en) * | 2021-01-15 | 2021-05-28 | 复旦大学 | Dynamic gesture recognition hardware accelerator for intelligent terminal field |
-
2012
- 2012-07-10 CN CN 201220331184 patent/CN203217590U/en not_active Expired - Lifetime
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861640A (en) * | 2021-01-15 | 2021-05-28 | 复旦大学 | Dynamic gesture recognition hardware accelerator for intelligent terminal field |
| CN112861640B (en) * | 2021-01-15 | 2022-07-22 | 复旦大学 | Dynamic gesture recognition hardware accelerator for intelligent terminal field |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103914149B (en) | Gesture interaction method and gesture interaction system for interactive television | |
| WO2020078017A1 (en) | Method and apparatus for recognizing handwriting in air, and device and computer-readable storage medium | |
| US20140033239A1 (en) | Next generation television with content shifting and interactive selectability | |
| CN105353876A (en) | Multi-point light sensation based spatial gesture identification control system and method | |
| CN102200834A (en) | television control-oriented finger-mouse interaction method | |
| CN112002311A (en) | Text error correction method and device, computer readable storage medium and terminal equipment | |
| CN104267834A (en) | Air mouse system and control method thereof | |
| CN109189219A (en) | The implementation method of contactless virtual mouse based on gesture identification | |
| CN104469265A (en) | Glass defect image acquiring and processing system | |
| CN203217590U (en) | Video intelligent gesture recognition system | |
| CN104143104A (en) | Image recognizing method and device, terminal device and server | |
| CN113656546A (en) | Multimodal search method, apparatus, device, storage medium, and program product | |
| CN101620723A (en) | Offline intelligent image information processing method and intelligent image information processing system | |
| Shi et al. | GestureCam: a smart camera for gesture recognition and gesture-controlled web navigation | |
| CN211293894U (en) | Hand-written interaction device in air | |
| Nam et al. | Smartphone accelerometer-based gesture recognition and its robotic application | |
| CN202720661U (en) | Gesture recognition system | |
| Zhu et al. | Hand gesture recognition based on structure analysis | |
| CN112580355B (en) | News information topic detection and real-time aggregation method | |
| CN111444771B (en) | Gesture preposing real-time identification method based on recurrent neural network | |
| CN205388775U (en) | Adopt parallel data processing's figure processing system | |
| Nikhil et al. | Design and development of a DSP processor based reconfigurable hand Gesture Recognition System for real time applications | |
| Zhao | A review on machine learning and gesture recognition | |
| CN201548972U (en) | Fingerprint recognition system for credit cards | |
| CN111079716A (en) | FPGA-based video gesture recognition intelligent home control system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CX01 | Expiry of patent term |
Granted publication date: 20130925 |
|
| CX01 | Expiry of patent term |