CN105787463A - Wi-Fi-based remote target tracking system and method - Google Patents

Abstract

The invention provides a Wi-Fi-based remote target tracking system and method, relates to the field of target tracking and intelligent monitoring and is to realize the purpose of carrying out stable tracking on a specific target within a regional Wi-Fi network coverage range and is to improve network bandwidth occupancy and reaction sensitivity. The system comprises a target tracking robot and a remote server. The robot has a customized Linux system and realizes real-time on-site video acquisition through a USB camera, carries out hard-coding to enable the on-site video collected in real time to be converted from a YUV format to a H.264 format, and transmits the converted video to the remote server through Wi-Fi; and the remote server is established in a Windows system in a computer, carries out FFmpeg library decoding on the received H.264 video stream and controls the mobile robot to track the target after carrying out STC target tracking algorithm processing on the specific target.

Description

Remote object based on Wi-Fi follows the tracks of system and method

Technical field

The present invention relates to target following and field of intelligent monitoring, be specifically related to a kind of remote object tracking technique.

Background technology

Building smart city, under the big tide of safe city, the figure of intelligent monitor system occurs in increasing field. Traditional telesecurity monitoring system is often based upon single camera or camera array, intelligent not, monitors mesh at some The modern technologies such as the occasion of mark behavior act rule, can pass through machine learning completely, the automatic detect and track of target reduce The consumption of manpower.

Remote object follows the tracks of the concept of system, it is simply that be tracked remote object with the mobile machine artificially basis of specific customization And by target information Real-time Feedback to the whole set of system of server.Remote object follows the tracks of system to being not easy to the danger that manually participates in Dangerous section makees, and saves unnecessary human cost, and improving the aspects such as system operating efficiency has important theory significance and application valency Value.Along with the development of the rapid technological improvements such as video compress, the video monitoring system of a new generation is able to be passed in real time by network Defeated on-the-spot HD video, carries out various process for rear end to real-time video and lays a good foundation；Target tracking algorism is current calculating In the numerous correlation technique of machine visual field important one, it applies attention degree also to grow with each passing day in actual life； On the other hand, the use moving robot the most ripe also provides reliable and stable for various relevant intelligent use Carrier.

Summary of the invention

The present invention is to realize in the range of region Wi-Fi network covers specifying target to carry out tenacious tracking, Yi Jiti High network bandwidth occupation rate and reaction sensitivity, thus provide a kind of remote object based on Wi-Fi to follow the tracks of system and method.

Remote object based on Wi-Fi follows the tracks of system, and it includes target following robot and Windows operating system environment Under streaming media server；

Described target following robot includes that the video flowing under motion planning and robot control module and Linux environment controls client Module；

Streaming Media under embedded Windows operating system environment in streaming media server under Windows operating system environment Service module；

Video flowing under Linux environment controls client modules and includes rs 232 serial interface signal sending module, based on V4L2 framework regards Frequently acquisition module, S5PV210MFC hardware encoding block and video sending module based on jrtlib storehouse；

Streaming media service module under Windows operating system environment includes STC target tracking algorism based on OpenCV Module, video decompression module based on FFmpeg storehouse and video reception module based on jrtlib storehouse；

Video acquisition module based on V4L2 framework: for setting the collection resolution ratio of image, frame per second and mass parameter；Also For by camera collection to the video flowing of yuv420p form be converted to NV12 format video stream, and be sent to S5PV210MFC hardware encoding block；

S5PV210MFC hardware encoding block: for the video flowing of NV12 form is utilized hardware compression module boil down to H.264 the video flowing of form, and it is sent to video transmission module based on jrtlib storehouse；

Video sending module based on jrtlib storehouse: for the NALU unit of H.264 video flowing is split according to rtp agreement, NALU unit after being split by the api function of jrtplib again is encapsulated as RTP network packet and is sent to Windows operation Streaming media server under system environments；

Video reception module based on jrtlib storehouse: for receiving the RTP net from video sending module based on jrtlib storehouse Network bag, is reassembled as complete NALU unit H.264, and is sent to video compression based on FFmpeg storehouse after verification Module；

The video decompression module in FFmpeg storehouse: being used for the H.264 decoding video stream received is BGR888 video flowing, And it is sent to STC target tracking algorism module based on OpenCV；

STC target tracking algorism module based on OpenCV: be used for utilizing OpenCV storehouse to write STC target following and calculate Method also carries out computing to video flowing, is sent to rs 232 serial interface signal sending module after generating robot control signal；

Rs 232 serial interface signal sending module: for receiving the robot from STC target tracking algorism module based on OpenCV Control signal, and send control signals to motion planning and robot control module；

Motion planning and robot control module: be used for controlling target following robotic tracking's target.

Remote object tracking based on Wi-Fi, it includes:

For setting the collection resolution ratio of image, frame per second and mass parameter；It is additionally operable to the yuv420p arrived by camera collection The video flowing of form is converted to NV12 format video stream, and is sent to the step of S5PV210MFC hardware encoding block；

S5PV210MFC hardware encoding block: for the video flowing of NV12 form is utilized hardware compression module boil down to H.264 the video flowing of form, and it is sent to the step of video transmission module based on jrtlib storehouse；

Video sending module based on jrtlib storehouse: for the NALU unit of H.264 video flowing is split according to rtp agreement, NALU unit after being split by the api function of jrtplib again is encapsulated as RTP network packet and is sent to Windows operation The step of the streaming media server under system environments；

Video reception module based on jrtlib storehouse: for receiving the RTP net from video sending module based on jrtlib storehouse Network bag, is reassembled as complete NALU unit H.264, and is sent to video compression based on FFmpeg storehouse after verification The step of module；

The video decompression module in FFmpeg storehouse: being used for the H.264 decoding video stream received is BGR888 video flowing, And it is sent to the step of STC target tracking algorism module based on OpenCV；

STC target tracking algorism module based on OpenCV: be used for utilizing OpenCV storehouse to write STC target following and calculate Method also carries out computing to video flowing, is sent to the step of rs 232 serial interface signal sending module after generating robot control signal；

Rs 232 serial interface signal sending module: for receiving the robot from STC target tracking algorism module based on OpenCV Control signal, and send control signals to the step of motion planning and robot control module；

Motion planning and robot control module: for controlling the step of target following robotic tracking's target.

The present invention is in the range of region WiFi network covers, it is achieved that carry out tenacious tracking, the network bandwidth to specifying target Occupation rate is low, is quick on the draw.And the present invention has a following advantage:

1, from network environment requires, what the present invention was transmitted is the video of the H.264 form of high compression, only takes up pole Little network speed, either LAN or school net can meet the communication environment of the present invention；

2, multi-thread mechanism makes the several functions of calling program can separately reduce the degree of coupling between each function so that program module Change, and be greatly improved the operating efficiency of program, more reasonably use system resource；

3, the arithmetic core of the present invention is the arm of S5PV210, has been internally integrated multimedia driver element MFC, can adopt It is H.264 form by the mode of hardware encoding by video flowing compression coding, is greatly accelerated program operation speed, alleviates The computational burden of arm；

4, data send and all use third party library design, the most multi-platform expansion with receiving, decode and showing.

5, from the processing speed of server software, when bright and clear, the processing speed of server has reached 19.6 Frame is per second, and when insufficient light, still to reach 15 frames per second for processing speed.The present invention has RMON and moving target Following the tracks of two big Core Features, extensibility is strong, and tradition safety defense monitoring system cooperation can apply to more occasion.

Accompanying drawing explanation

Fig. 1 is the function framework schematic diagram of remote object based on the Wi-Fi tracking system of the present invention；

Fig. 2 is video acquisition working cycles schematic diagram；

Fig. 3 is multimedia application software architecture diagram based on MFC；

Fig. 4 is streaming media server software interface figure；

Tu5Shi robot work demonstration graph；

Fig. 6 is that robot view field divides schematic diagram；

Fig. 7 is STC algorithm spatial context learning process schematic diagram.

Fig. 8 is STC algorithm target location testing process schematic diagram.

Detailed description of the invention

Detailed description of the invention one, combining Fig. 1 this detailed description of the invention is described, remote object based on Wi-Fi follows the tracks of system, It includes the streaming media server under target following robot and Windows operating system environment；

Described target following robot includes that the video flowing under motion planning and robot control module and Linux environment controls client Module；Video flowing under Linux environment controls client modules and utilizes third-party jrtplib storehouse, FFmpeg storehouse, OpenCV Storehouse, achieves video reception, H.264 decoding and the tracking of STC target tracking algorism respectively.Whole remote object The software systems following the tracks of systems soft ware use C/S framework, are divided into: video flowing under Linux environment control client modules, Streaming media service module under motion planning and robot control module and Windows operating system environment.

Motion planning and robot control module uses the minimum system plate circuit of Freescale single-chip microcomputer MC9S12XS128.Video flowing Control platform to be made up of video flowing Control function circuit, USB2.0 camera, TP-LINK Wireless USB network interface card；Select Samsung arm processor S5PV210 controls the master controller of platform as video flowing；Select OK210 development board as regarding Frequently the carrier of flow control platform.

Streaming Media under embedded Windows operating system environment in streaming media server under Windows operating system environment Service module；

Video flowing under Linux environment controls client modules and includes rs 232 serial interface signal sending module, based on V4L2 framework regards Frequently acquisition module, S5PV210MFC hardware encoding block and video sending module based on jrtlib storehouse；

Streaming media service module under Windows operating system environment includes STC target tracking algorism based on OpenCV Module, video decompression module based on FFmpeg storehouse and video reception module based on jrtlib storehouse；

Video acquisition module based on V4L2 framework: for setting the parameters such as the resolution ratio gathering image, frame per second, quality. The video flowing of the yuv420p form collected is converted to the MFC hardware volume that S5PV210 processor carries by coding NV12 form required for code module；

S5PV210MFC hardware encoding block: for NV12 video being fed for the hardware-compressed mould being integrated within arm Block, is H.264 form by video compress；

Video sending module based on jrtlib storehouse: for the NALU unit of H.264 video flowing is split according to rtp agreement, NALU unit after being split by the api function of jrtplib again is encapsulated as RTP network packet and is sent to Windows operation Streaming media server under system environments；

Video reception module based on jrtlib storehouse: for receiving the RTP net from video sending module based on jrtlib storehouse Network bag, is reassembled as complete NALU unit H.264, and is sent to video compression based on FFmpeg storehouse after verification Module；

The video decompression module in FFmpeg storehouse: for H.264 the api function utilizing FFmpeg storehouse is regarded receive Frequently stream is decoded as the BGR888 video flowing that OpenCV needs；

STC target tracking algorism module based on OpenCV: utilize OpenCV storehouse to write STC target tracking algorism journey Sequence, carries out operational analysis to video flowing, according to Tracking Control Strategy, is sent to robot after drawing the control signal of robot Wireless module；

Rs 232 serial interface signal sending module: the robot control signal received is sent to robot by serial ports and transports by linux system Autocontrol system；

Motion planning and robot control module: be used for receiving serial port control signal and control robot tracking target.

Principle: building of S5PV210 development board platform Linux environment:

Here directly select the kernel source code of development board producer for Speeding up development speed, source code is operated, to add kernel Support to TP-LINK TL-WN725N v2 version Wireless USB interface module.After it should be noted that to be introduced The source code flow that MFC hardware encoding H.264 function needs is NV12 type, and NV12 has two kinds of storages in internal memory Form linear mode and tile mode, supports tile mode type NV12 code stream, for the ease of journey under kernel default situations Writing of sequence, needs MFC multimedia to drive module support to receive the NV12 code stream of linear mode type.Need to revise phase Close kernel file, open kernel file ./drivers/media/video/samsung/mfc50/mfc_opr, then MEM_STRUCT_TILE_ENC preserves after making MEM_STRUCT_LINER into.

The process of compiling kernel:

1), in the position of console entrance kernel source code, execution order #make distclen, removes and compiles generation in the past All Files.

2), official's configuration file of using Fei Ling embedded technology Co., Ltd to provide simplify configuration process, hold at console Line command #cp ok210p_linux_config.config.

3), perform order #make menuconfig, enter graphical kernel setup interface, verification required function in adding Support, the support of verification idle work energy in deleting.The path, place to the support that rtl8188eu drives added is: Device Drivers/Networkdevice support/Wireless LAN/RTL8188EU。

4), perform order #make zImage compile and generate kernel mirror image.

Here repeat no more the process of compiling u-boot-1.3.4 and yaffs2 file system image, by generating file above Complete the installation of development board (SuSE) Linux OS.By revising the configuration of wpa_spplicant after installation so that open Send out plate and can connect WiFi network by TP-LINK TL-WN725N v2 version Wireless USB interface module.So far, soft Linux environment needed for part system is built complete.

2, Software for Design based on V4L2 framework

1), equipment is opened

In general, the node title of V4L2 equipment is /dev/dev/videoX.The design opens readable with non-blocking fashion Writeable equipment.

2), relevant parameter is set

A), picture format is set, is mainly used in determining the color space of image and resolution ratio；

B), arranging video flowing parameter, be mainly used in arranging the resolution ratio of video, resolution ratio is set to 25 frames per second by program, Best through testing the operational effect of this frame per second.

3) by the way of internal memory maps, read video flowing design

Being the process of an operation mass data to continuous videos collection, the mode using the equipment that directly reads is not wise choosing Selecting the present invention here uses internal memory to map, as in figure 2 it is shown, read video flowing by the way of internal memory maps to need following steps:

A), to driving application video flowing frame buffer zone to deposit image for temporarily；

B), program sets up internal memory mapping relations between the address and user's space of frame buffer zone.

3, hardware encoding block based on MFC multimedia driver element design

Samsung official provides one group of api function being user-friendly to and controls MFC equipment, and these api functions are directly adjusted With MFC drive input and output control function, multimedia application software configuration based on MFC as it is shown on figure 3, User uses these API, API Calls MFC to drive the output control function L provided, and output control function controls MFC, Complete the encoding and decoding of video image.

4, TP-LINK TL-WN725N v2 version Wireless USB trawl performance is transplanted

Linux 2.6.35.7 is per se with the driving of the wireless chip rtl8188cus of realtek Realtek, but can buy on the market To model mostly be the TL-WN725N v2 version Wireless USB network interface card with rtl8188eu chip, so need by The driving of rtl8188eu chip is transplanted in Linux, and the interface configurations of makefile and meunconfig to Linux File is modified.Required following step:

1), download from the Internet 8188ue file is placed under ./drivers/net/wireless.

2) the Makefile file, under amendment wireless adds obj-$ (CONFIG_RTL8188EU) +=rtl8188eu/, It is revised as CONFIG_POWER_SAVING=n, closes battery saving mode.

3), before code endif#WLAN, add source " drivers/net/wireless/rtl8188eu/Kconfig ", give figure Menu is added at shape kernel setup interface.

So far, the program portable success of TP-LINK TL-WN725N v2 version Wireless USB trawl performance.

5, RTP packet stream journey is transmitted based on jrtplib storehouse

Owing to the design is multiplatform environments at client-side and server end and it is directed to the transmission work of RTP bag, institute To use jrtp3.9.1 to carry out the exploitation of correlation function.

It is common that the transmission work of RTP bag is received two parts by the RTP bag transmission of robot client and the RTP of server end Complete, wherein the program frame of robot client mainly include detecting receive packet, obtain and be currently received data The relevant information of bag, RTP bag transferring H.264 code stream to, decompression also transcoding is the BGR888 form needed for OpenCV After be stored in chained list buffer area.

6, the design of the streaming media server module under Windows environment

The streaming media server of the present invention operates under windows7 environment, and program is write also in visual studio2010 Debugging, the human-computer interaction interface development of server utilizes MFC to complete.The exploitation of server module framework uses multi-thread Journey programming idea.

The present invention opens main thread respectively at server end as required, Video_rec_thread (video flowing receiving thread), Video_show_thread (video flowing display thread), Video_proc_thread (video flow processing thread), work concomitantly Make.Server workflow is as shown in Figure 4.Wherein, main thread be responsible for operate interface display, the reception of user command, The establishment of worker thread and destruction；Video flowing receiving thread is responsible for receiving the RTP packet of robot client, and decoding is also Transcoding is BGR form, is stored in by BGR frame in buffering chained list；Video flowing display thread is responsible for from BGR buffering chained list Extract present frame, the video flowing real-time display in server display, and present frame is passed to video flow processing thread； Video flow processing thread is responsible for being tracked present frame algorithm process, controls robot client and moves, release present frame Data space.

7, the Tracking Control Strategy of robot

Owing to the robot in the present invention uses camera collection external information, therefore the tracking strategy used here is view-based access control model Tracking strategy.Robot work demonstration is as shown in Figure 5.Implement process as follows:

The visual field of mobile robot is rectangle, this rectangle is divided into 3 parts from top to bottom, is from left to right divided into 3 Part, rectangular field of view so can be divided into 9 rectangular areas, as shown in Figure 6.In figure, red expression follows the tracks of in target area The heart, the coordinate at this center be (x, y), mobile robot is according to the relative position of this coordinate with central region coordinate, it may be assumed that Distribution situation in 9 rectangular areas moves.Target is upwards looked up by mobile robot, 9 be divided in the visual field Individual region by from left to right, order arrangement from top to bottom, when target's center occurs in that region robot make with Lower reaction: turn left and retreat, retreating, turning right and retreating, flicker, stopping, flicker, turn left and advance, front Enter, turn right and advance.

8, the processing procedure of STC target tracking algorism

The core missions of Target Tracking Problem are that calculate can the confidence map of future position.

C (x)=P (x | o), (1)

Wherein x ∈ R²Representing the position that target is occurred in, o represents that this event occurs in target.Subsequently, space-time context Information is used as estimating.

In the current frame, we learn target location x^*(i.e. the coordinate at tracked target center).Contextual feature is defined For X^c=c (z)=(I (z), z) | z ∈ Ω_c(x^*), wherein I (z) represents the image intensity value at z, Ω_c(x^*) represent x^*Attached Near position.By marginalisation joint probability P (x, c (z) | o), the possibility of cursor position can be calculated by minor function:

$\begin{array}{c}c\left(x\right)=P\left(x|o\right)\\ ={\Σ}_{c\left(z\right)\∈x^c}P\left(x,c\left(z\right)|o\right)\\ ={\Σ}_{c\left(z\right)\∈x^c}P\left(x,c\left(z\right),o\right)P\left(c\left(z\right)|o\right)\end{array},---\left(2\right)$

Here, and conditional probability P (x | c (z), o) represent target location and the time-space relationship of its contextual information, context Information can solve the different produced ambiguity due to measurement result.P (x, c (z) | o) it is used to adjacent context to carry out The prior probability of abstract.The groundwork of algorithm learns P exactly, and (x, c (z), o), because it is target location and at that time The tie of empty context.

Probability function P in (2) formula (x, c (z) | o) can be expressed as,

P (x, c (z) | o)=h^sc(x-z), (3)

Wherein, h^sc(x-z) used target location x and his adjacent contextual location z, so it is to be appreciated that target and it Context space-time between relation.Context prior probability can be reduced to further,

P (c (z) | o)=I (z) w_σ(z-x^*) (4)

(4) I (.) in formula represents the gradation of image of context, and w σ (.) can be further represented as,

$w_{\mathrm{\σ}}\left(z\right)={\mathrm{ae}}^{-\frac{|z{|}^2}{{\mathrm{\σ}}^2}}---\left(5\right)$

Wherein a be P (c (z) | o) is restricted to 0 to 1 normaliztion constant, σ is scale parameter.

Describe target location confidence map can may be further described into

$c\left(x\right)=P\left(x\right|o)={\mathrm{be}}^{-|\frac{x-x^{*}}{\mathrm{\α}}{|}^{\mathrm{\β}}}---\left(6\right)$

Wherein, b normaliztion constant, α is scale parameter, and β is profile parameters.

According to confidence map function (6), context prior probability (4), in order to calculate space-time context model (3), by above-mentioned Take (3) to obtain

$\begin{array}{c}c\left(x\right)={\mathrm{be}}^{-{\left|\frac{x-x^{*}}{\mathrm{\α}}\right|}^{\mathrm{\β}}}\\ ={\Σ}_{z\∈{\mathrm{\Ω}}_c\left(x^{*}\right)}{h}^{sc}\left(x-z\right)I\left(z\right)w_{\mathrm{\σ}}\left(z-x^{*}\right)\\ =h^{sc}\left(x\right)\⊗\left(I\left(x\right)w_{\mathrm{\σ}}\left(x-x^{*}\right)\right)\end{array},---\left(7\right)$

WhereinRepresent convolution algorithm.

Use Fast Fourier Transform (FFT) (FFT) to carry out convolutional calculation, obtain frequency-domain expression:

The basic procedure that STC algorithm performs is as shown in Figure 7 and Figure 8.Assuming that primary condition: pass through when the first frame Operating personnel are manually or some detection algorithm have chosen the tracking target specified.When t frame, sky now can be obtained Between context model h^sc(x), and use h^scX () detects the position of target in t+1 frame,

$h^{sc}\left(x\right)=F^{-1}\left(\frac{F\left({\mathrm{be}}^{-|\frac{x-x^{*}}{\mathrm{\α}}{|}^{\mathrm{\β}}}\right)}{F\left(I\right(x)w_{\mathrm{\σ}}(x-x^{*}))}\right)---\left(9\right)$

Wherein, F^-1Represent reverse FFT.Subsequently in space-time context modelBe updated, concrete more New formula is:

$H_{t+1}^{stc}\left(x\right)=(1-\mathrm{\ρ})H_t^{stc}+{\mathrm{\ρh}}_t^{sc}---\left(10\right)$

Wherein, ρ is learning parameter,Being to calculate the space-time context model when t frame, x is target location, I (x) For the gray scale of an x, w is weighting function.This is a time-domain filtering process, after Fourier transformation it appeared that on space-time Hereafter model can filter out picture noise, the most reinflated.

When t+1 frame, it is cut out regional area Ω according to target location during t frame_c(x^*), and build target signature Collection x^c=c (z)=(I (z), z) | z ∈ Ω_c(x^*)), now can in the hope of the maximum likelihood probability of confidence map during t+1 frame also Obtain the new position that target occurs,

$x_{t+1}^{*}=\arg \underset{x\∈{\mathrm{\Ω}}_c\left(x_t^{*}\right)}{max}{c}_{t+1\left(x\right)},---\left(11\right)$

By (8) Shi Ke get,

Although the position of target can be obtained, the problem the most not stopping change in order to solve target scale, need to use Following formula (13) real-time update^wσ in σ.

$\left\{\begin{array}{c}{s}^{\′}=\sqrt{\frac{c_t\left(x_t^{*}\right)}{c_{t-}\left(x_{t-1}^{*}\right)}},\\ \stackrel{\‾}{s}=\frac{1}{n}{\mathrm{\Σ}}_{i=1}^n{s^{\′}}_{t-i},\\ s_{t+1}=(1-\mathrm{\λ})s_t+\mathrm{\λ}{\stackrel{\‾}{s}}_t,\\ {\mathrm{\σ}}_{t+1}=s_t{\mathrm{\σ}}_t,\end{array}\right.---\left(13\right).$

Claims (6)

1. remote object based on Wi-Fi follows the tracks of system, it is characterized in that: it includes target following robot and Windows Streaming media server under operating system environment；

Described target following robot includes that the video flowing under motion planning and robot control module and Linux environment controls client mould Block；

Streaming Media under embedded Windows operating system environment clothes in streaming media server under Windows operating system environment Business module；

Video flowing under Linux environment controls client modules and includes rs 232 serial interface signal sending module, video based on V4L2 framework Acquisition module, S5PV210MFC hardware encoding block and video sending module based on jrtlib storehouse；

Streaming media service module under Windows operating system environment includes STC target tracking algorism mould based on OpenCV Block, video decompression module based on FFmpeg storehouse and video reception module based on jrtlib storehouse；

Video acquisition module based on V4L2 framework: for setting the collection resolution ratio of image, frame per second and mass parameter；Also For by camera collection to the video flowing of yuv420p form be converted to NV12 format video stream, and be sent to S5PV210 MFC hardware encoding block；

S5PV210MFC hardware encoding block: for the video flowing of NV12 form is utilized hardware compression module boil down to H.264 the video flowing of form, and it is sent to video transmission module based on jrtlib storehouse；

Video sending module based on jrtlib storehouse: for the NALU unit of H.264 video flowing is split according to rtp agreement, NALU unit after being split by the api function of jrtplib again is encapsulated as RTP network packet and is sent to Windows operation Streaming media server under system environments；

Video reception module based on jrtlib storehouse: for receiving the RTP network from video sending module based on jrtlib storehouse Bag, is reassembled as complete NALU unit H.264, and is sent to video decompression module based on FFmpeg storehouse after verification；

The video decompression module in FFmpeg storehouse: being used for the H.264 decoding video stream received is BGR888 video flowing, and It is sent to STC target tracking algorism module based on OpenCV；

STC target tracking algorism module based on OpenCV: be used for utilizing OpenCV storehouse to write STC target following and calculate Method also carries out computing to video flowing, is sent to rs 232 serial interface signal sending module after generating robot control signal；

Rs 232 serial interface signal sending module: for receiving the robot control from STC target tracking algorism module based on OpenCV Signal processed, and send control signals to motion planning and robot control module；

Motion planning and robot control module: be used for controlling target following robotic tracking's target.

Remote object based on Wi-Fi the most according to claim 1 follows the tracks of system, it is characterised in that based on jrtlib storehouse Video sending module and video reception module based on jrtlib storehouse based on jrtplib storehouse realize.

Remote object based on Wi-Fi the most according to claim 1 follows the tracks of system, it is characterised in that based on FFmpeg The video decompression module in storehouse uses the api function in FFmpeg storehouse to realize.

Remote object based on Wi-Fi the most according to claim 1 follows the tracks of system, it is characterised in that robot motion is controlled Molding block uses the minimum system plate circuit realiration of MC9S12XS128 single-chip microcomputer.

Remote object based on Wi-Fi the most according to claim 1 follows the tracks of system, it is characterised in that S5PV210MFC Hardware encoding block uses S5PV210ARM processor to control the master controller of platform as video flowing.

6. remote object tracking based on Wi-Fi, is characterized in that: it includes:

For setting the collection resolution ratio of image, frame per second and mass parameter；It is additionally operable to the yuv420p lattice arrived by camera collection The video flowing of formula is converted to NV12 format video stream, and is sent to the step of S5PV210MFC hardware encoding block；

S5PV210MFC hardware encoding block: for the video flowing of NV12 form is utilized hardware compression module boil down to H.264 the video flowing of form, and it is sent to the step of video transmission module based on jrtlib storehouse；

Video sending module based on jrtlib storehouse: for the NALU unit of H.264 video flowing is split according to rtp agreement, NALU unit after being split by the api function of jrtplib again is encapsulated as RTP network packet and is sent to Windows operation The step of the streaming media server under system environments；

Video reception module based on jrtlib storehouse: for receiving the RTP network from video sending module based on jrtlib storehouse Bag, is reassembled as complete NALU unit H.264, and is sent to video decompression module based on FFmpeg storehouse after verification Step；

The video decompression module in FFmpeg storehouse: being used for the H.264 decoding video stream received is BGR888 video flowing, and It is sent to the step of STC target tracking algorism module based on OpenCV；

STC target tracking algorism module based on OpenCV: be used for utilizing OpenCV storehouse to write STC target following and calculate Method also carries out computing to video flowing, is sent to the step of rs 232 serial interface signal sending module after generating robot control signal；

Rs 232 serial interface signal sending module: for receiving the robot control from STC target tracking algorism module based on OpenCV Signal processed, and send control signals to the step of motion planning and robot control module；

Motion planning and robot control module: for controlling the step of target following robotic tracking's target.