CN207133982U - A kind of radar velocity measurement device based on deep learning GPU with video analysis - Google Patents
A kind of radar velocity measurement device based on deep learning GPU with video analysis Download PDFInfo
- Publication number
- CN207133982U CN207133982U CN201720965511.6U CN201720965511U CN207133982U CN 207133982 U CN207133982 U CN 207133982U CN 201720965511 U CN201720965511 U CN 201720965511U CN 207133982 U CN207133982 U CN 207133982U
- Authority
- CN
- China
- Prior art keywords
- gpu
- velocity measurement
- radar velocity
- network signal
- deep learning
- 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
- Traffic Control Systems (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The utility model discloses a kind of radar velocity measurement device based on deep learning GPU with video analysis, including：Head end video collecting unit, radar velocity measurement unit, GPU deep learning processors, photo grasp shoot device, memory cell, traffic control monitor supervision platform, transmission unit, database, the head end video collecting unit, radar velocity measurement unit is electrically connected by network signal with GPU deep learning processors respectively, the GPU deep learnings processor divides two-way to be electrically connected respectively with photo grasp shoot device and transmission unit by network signal, the photo grasp shoot device electrically connects a memory cell by network signal, the traffic control monitor supervision platform of the memory cell docking location, the transmission unit electrically connects a database by network signal.The utility model has enhancing video analysis resolution ratio, precisely identification 3-D view and recognition speed is fast, improves supervisory efficiency, reduces the advantages of accident rate.
Description
Technical field
A kind of radar velocity measurement device based on deep learning GPU with video analysis is the utility model is related to, applied to friendship
Logical regulation technique field.
Background technology
With the continuous improvement of people's living standards, the probability that people need to drive vehicle is consequently increased, but mostly
The sense of traffic that number driver's famine is fastened the safety belt, and like out exceed the speed limit car, car of being angry, add the danger of traffic accident
Property, this requires that traffic control department is monitored and managed to it, the traffic behavior analyzer installed on conventional road using
DSP Processor, only two mega pixels, video analysis amount and resolution ratio are all relatively small and to the recognition capability of 3-D view
Also it is relatively weak, therefore the low DSP Processor of efficiency is difficult under conditions of vehicle flowrate increases severely, effectively to car steering row
To be analyzed.
Utility model content
To solve the defects of prior art, the utility model discloses one kind enhancing video analysis resolution ratio, precisely
Identify 3-D view and recognition speed is fast, effectively judge, collect evidence, improve efficiency, improve supervisory efficiency and reduce vehicle indirectly
Accident rate.
The utility model discloses a kind of radar velocity measurement device based on deep learning GPU with video analysis, including：Before
Hold video acquisition unit, radar velocity measurement unit, GPU deep learnings processor, photo grasp shoot device, memory cell, traffic control monitoring
Platform, transmission unit, database, the head end video collecting unit, radar velocity measurement unit are deep by network signal and GPU respectively
Degree study processor electrical connection, the GPU deep learnings processor by network signal divide two-way respectively with photo grasp shoot device
And transmission unit electrical connection, the photo grasp shoot device electrically connect a memory cell, the memory cell by network signal
The traffic control monitor supervision platform of location is docked, the transmission unit electrically connects a database by network signal.
The height of the head end video collecting unit installation is road surface above 3-4 meter Chu positions.
The radar velocity measurement unit is radar meter.
The GPU deep learnings processor is using CEVA-XM4 vision processors.
Because using above-mentioned technical proposal, the utility model has following advantageous benefits：
1st, strengthen video analysis resolution ratio, precisely identification 3-D view and recognition speed is fast；
2nd, effectively judge, collect evidence, improve efficiency；
3rd, improve supervisory efficiency, reduce accident rate.
Brief description of the drawings
Fig. 1 is a kind of outer map interlinking of the radar velocity measurement device based on deep learning GPU with video analysis of the utility model.
Wherein：1- head end video collecting units；2- radar velocity measurement units；3-GPU deep learning processors；4- photos are captured
Device；5- memory cell；6- traffic control monitor supervision platforms；7- transmission units；8- databases.
Embodiment
As shown in figure 1, the utility model discloses a kind of radar velocity measurement dress based on deep learning GPU with video analysis
Put, including：Head end video collecting unit 1, radar velocity measurement unit 2, GPU deep learnings processor 3, photo grasp shoot device 4, storage
Unit 5, traffic control monitor supervision platform 6, transmission unit 7, database 8, the head end video collecting unit 1, radar velocity measurement unit 2 are distinguished
Electrically connected by network signal with GPU deep learnings processor 3, the GPU deep learnings processor 3 is divided to two by network signal
Road electrically connects with photo grasp shoot device 4 and transmission unit 7 respectively, and the photo grasp shoot device 4 electrically connects one by network signal
Individual memory cell 5, the memory cell 5 dock the traffic control monitor supervision platform 6 of location, and the transmission unit 7 is believed by network
Number electrical connection one database 8.
The height that the head end video collecting unit 1 is installed is road surface above 3-4 meter Chu positions.
The radar velocity measurement unit 2 is radar meter.
The GPU deep learnings processor 3 is using CEVA-XM4 vision processors, CEVA deep-neural-network frameworks
Provided for the algorithm based on convolutional neural networks from off-line training to the quick smooth path detected in real time, can be in short number
The unique target detection network implementations by optimization are obtained in it, and the other platforms of power dissipation ratio significantly reduce.
The utility model is implemented：In use, come head-on in the road pavement of head end video collecting unit 1
Vehicle carry out video acquisition, radar velocity measurement unit 2 measures to Vehicle Speed, and then both believe video and measurement
Number GPU deep learnings processor 3 is uploaded to by network signal, using GPU processors GPU deep learnings processor 3 will
The 3-D view arrived carries out rapid identifying processing, and whether the shoulder and waist for passing through shot the video middle driver all have safety belt
Imaging, so as to judge whether the driver in driving vehicle does not fasten the safety belt, while can also be according to the signal of radar velocity measurement unit 2
Judge whether vehicle exceeds the speed limit, meet one if GPU deep learnings processor 3 judges that vehicle is not fastened the safety belt or exceeded the speed limit, in both
Item or both is satisfied by, and now GPU deep learnings processor 3 transmits rapidly a network signal to photo grasp shoot device 4, photo
Grasp shoot device 4 carries out high definition snapshot to front vehicles immediately after network signal is received, and photo grasp shoot device 4 will after candid photograph
Gained image and data upload to memory cell 5, and the data timing in memory cell 4 is uploaded to traffic control monitor supervision platform 6, then by handing over
The staff of pipe monitor supervision platform 6 is classified, and is then punished；If driver's driving behavior meets the requirements, GPU
The video collected is sent to transmission unit 7 by deep learning processor 3, then by the network of transmission unit 7 to be uploaded to database 8 right
It is retained, and using the technical program, is constrained driver to a certain extent, is reduced accident rate.
Finally it should be noted that：Above example only not limits the utility model and retouched to illustrate the utility model
The technical scheme stated；Therefore, although this specification the utility model is had been carried out with reference to above-mentioned each embodiment it is detailed
Illustrate, still, it will be understood by those within the art that, still the utility model can be modified or equally replaced
Change；And all do not depart from technical scheme and its improvement of spirit and scope of the present utility model, it all should cover new in this practicality
In the right of type.
Claims (4)
1. a kind of radar velocity measurement device based on deep learning GPU with video analysis, including：Head end video collecting unit, thunder
Up to the unit that tests the speed, GPU deep learnings processor, photo grasp shoot device, memory cell, traffic control monitor supervision platform, transmission unit, data
Storehouse, it is characterised in that：The head end video collecting unit, radar velocity measurement unit pass through network signal and GPU deep learnings respectively
Processor electrically connect, the GPU deep learnings processor by network signal divide two-way respectively with photo grasp shoot device and transmission
Unit electrically connects, and the photo grasp shoot device electrically connects a memory cell by network signal, and the memory cell docks institute
Traffic control monitor supervision platform in area, the transmission unit electrically connect a database by network signal.
2. a kind of radar velocity measurement device based on deep learning GPU with video analysis, its feature are existed according to claim 1
In：The height of the head end video collecting unit installation is road surface above 3-4 meter Chu positions.
3. a kind of radar velocity measurement device based on deep learning GPU with video analysis, its feature are existed according to claim 1
In：The radar velocity measurement unit is radar meter.
4. a kind of radar velocity measurement device based on deep learning GPU with video analysis, its feature are existed according to claim 1
In：The GPU deep learnings processor is using CEVA-XM4 vision processors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201720965511.6U CN207133982U (en) | 2017-08-04 | 2017-08-04 | A kind of radar velocity measurement device based on deep learning GPU with video analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201720965511.6U CN207133982U (en) | 2017-08-04 | 2017-08-04 | A kind of radar velocity measurement device based on deep learning GPU with video analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN207133982U true CN207133982U (en) | 2018-03-23 |
Family
ID=61640635
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201720965511.6U Active CN207133982U (en) | 2017-08-04 | 2017-08-04 | A kind of radar velocity measurement device based on deep learning GPU with video analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN207133982U (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111490491A (en) * | 2020-04-30 | 2020-08-04 | 国网上海市电力公司 | Ultra-high voltage transmission line inspection unmanned aerial vehicle based on deep learning |
-
2017
- 2017-08-04 CN CN201720965511.6U patent/CN207133982U/en active Active
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111490491A (en) * | 2020-04-30 | 2020-08-04 | 国网上海市电力公司 | Ultra-high voltage transmission line inspection unmanned aerial vehicle based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103593678B (en) | A kind of long-span bridge vehicle dynamic load distribution detection method | |
| CN104599249B (en) | Cableway platform bridge floor car load is distributed real-time detection method | |
| CN103057470B (en) | A kind of vehicle peccancy line ball pre-tip device and reminding method thereof | |
| CN202793389U (en) | On-line routing inspection device for high-speed rail overhead line system | |
| CN108197610A (en) | A kind of track foreign matter detection system based on deep learning | |
| CN104573646A (en) | Detection method and system, based on laser radar and binocular camera, for pedestrian in front of vehicle | |
| CN106530831A (en) | System and method for monitoring and early warning of high-threat vehicles | |
| CN103021059A (en) | Video-monitoring-based public transport passenger flow counting method | |
| CN102765365A (en) | Pedestrian detection method based on machine vision and pedestrian anti-collision warning system based on machine vision | |
| CN102509418A (en) | Fatigue driving estimation and early-warning method and device of multi-sensor information fusion | |
| CN102252859A (en) | Road train straight-line running transverse stability automatic identification system | |
| CN206074832U (en) | A kind of railcar roof pantograph foreign matter detection system | |
| CN105632180B (en) | A kind of bridge tunnel entrance model recognition system and method based on ARM | |
| CN101308606A (en) | Traffic information collecting apparatus and method thereof | |
| CN104899880A (en) | Method for automatically detecting open and closed states of vehicle door of public traffic vehicle | |
| CN104809470A (en) | Vehicle converse running detection device and method based on SVM | |
| CN206961331U (en) | A kind of high threat vehicle monitoring early warning system | |
| CN103236158A (en) | Method for warning traffic accidents in real time on basis of videos | |
| CN109987098A (en) | The method and device of safe driving control is carried out based on Pavement analysis | |
| CN107516423A (en) | A kind of vehicle heading detection method based on video | |
| CN207133982U (en) | A kind of radar velocity measurement device based on deep learning GPU with video analysis | |
| CN204613981U (en) | A kind of vehicle based on SVM drives in the wrong direction pick-up unit | |
| CN106960193A (en) | A kind of lane detection apparatus and method | |
| CN110738852A (en) | intersection steering overflow detection method based on vehicle track and long and short memory neural network | |
| KR20140146837A (en) | Smart road monitoring system using sensor and camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |