CN202676617U - Embedded processor-based distributed paper defect detection system - Google Patents
Embedded processor-based distributed paper defect detection system Download PDFInfo
- Publication number
- CN202676617U CN202676617U CN 201220306189 CN201220306189U CN202676617U CN 202676617 U CN202676617 U CN 202676617U CN 201220306189 CN201220306189 CN 201220306189 CN 201220306189 U CN201220306189 U CN 201220306189U CN 202676617 U CN202676617 U CN 202676617U
- Authority
- CN
- China
- Prior art keywords
- paper
- paper defects
- image
- detection
- processor
- 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 - Fee Related
Links
Images
Abstract
The utility model relates to an embedded processor-based distributed paper defect detection system. The embedded processor-based distributed paper defect detection system is characterized in that multiple industrial linear array charge-coupled device (CCD) cameras which are arranged in parallel acquire image data of moving paper webs, and the image data are transmitted to each corresponding embedded paper defect detection processor through a CameraLink cable; each embedded paper defect detection processor acquires and preprocesses images through a field programmable gate array (FPGA) and then detects data of types, areas, positions and the like of paper defects existing in the images through a digital signal processor (DSP); the detection result and the corresponding paper defect image of each detection processor are transmitted to a central server through a gigabit Ethernet; and the central server updates and displays multiple detection results and the paper defect images through child windows, and stores the detection results and the paper defect images to backend database. The embedded processor-based distributed paper defect detection system has the advantages of low construction cost and easiness in realization of distributed detection, and can be applied to high-speed on-line detection of common paper defects such as black spots, holes, wrinkles, scratches and the like of paper webs which move at a speed of not less than 1 km/min, and the detection speed can be improved greatly.
Description
Technical field
The utility model relates to the paper sheet defect detection technique field based on machine vision, relates in particular to a kind of distributed web inspection system based on flush bonding processor.
Background technology
In the paper production, paper surface can produce some defectives inevitably, i.e. paper defects, such as spot, hole, gauffer, scratch etc., these paper defects are to have a strong impact on paper quality, thus paper defects to detect be a important step in the production run.Traditional paper defects detects by manually finishing, but modern paper already has wide, the fireballing characteristics of paper web, relies on manual detection page paper defects can not meet the demands.In recent years, along with the development of machine computing machine and digital image processing techniques, be widely applied at some large-scale papermaking enterprises based on the paper defects on-line monitoring system of machine vision, this system can realize fast and effectively that the paper web paper defects detects online.But in this system, paper defects detects the main image processing software that operates on the PC that relies on and realizes, this detection scheme is unfavorable for realizing large-scale distributed online detection on the one hand, on the other hand because its construction cost is generally all higher, be restricted in the application of the papermaking enterprise of middle and small scale.For these problems, in the paper defects detection field that progressively is applied to based on the machine vision technique of flush bonding processor, and some application systems have appearred, these systems adopt the high-speed dsp chip as the core processor of paper defects graphical analysis mostly, overcome to a certain extent the shortcoming based on the paper defects detection of PC, but because the collection of paper defects image, pre-service, all processing operations such as paper defects detection are all finished by DSP, the computational load of DSP own is just very large, and dsp chip also needs frequent and External memory equipment carries out exchanges data, therefore detect for real-time online, argin is inadequate, adopts the speed of this technical scheme to promote limited.
The utility model content
The purpose of this utility model is exactly the defective that exists in order to overcome existing paper defects detection technique, a kind of distributed web inspection system based on flush bonding processor is provided, it consists of that cost is low, the speed room for promotion large, is easy to realize Distributed Detection can realize the online detection of the common paper defects such as blackspot, hole, fold, scratch.
To achieve these goals, the utility model adopts following technical scheme:
A kind of distributed web inspection system based on flush bonding processor, it comprises at least one group of high speed camera, high speed camera is installed in motion paper web top, then is provided with led light source below the motion paper web; High speed camera is connected with embedded paper defects measurement processor, and embedded paper defects measurement processor is communicated by letter with central server by Ethernet switch; Wherein, embedded paper defects measurement processor is realized by FPGA and DSP cooperation, comprise the image acquisition units in the FPGA, it is communicated by letter with transceiving chip successively with dual port RAM, paper defects image pretreatment unit, Ethernet interface control module and the FPGA kilomega network control outward in the FPGA; Paper defects image pretreatment unit detects DSP with paper defects, high-speed SRAM is communicated by letter; Image acquisition units is by LVDS/ LVTTL conversion chip acquisition of image data, by LVTTL/ LVDS conversion chip control high speed camera; Paper defects image pretreatment unit receives the view data of image acquisition units, and it is carried out pre-service, and paper defects detects the view data that DSP receives paper defects image pretreatment unit, and carries out paper defects and detect; The Ethernet interface control module is controlled kilomega network control and the camera parameters of transceiving chip reception from Ethernet on the one hand, and be sent to image acquisition units, receive on the other hand the paper defects testing result data from the image pretreatment unit, and the control of control kilomega network is sent to Ethernet with transceiving chip.
Described high speed camera is comprised of multichannel Industry thread array CCD camera, laterally is installed in side by side the top of motion paper web; Every road camera is connected to separately embedded paper defects measurement processor by Camera Link cable.
A kind of detection method that adopts based on the distributed web inspection system of flush bonding processor, its step is:
(1) parameter setting
Software operation is detected on foreground by central server, and the parameters such as line frequency, time shutter and picture size of each road high speed camera are arranged;
(2) detect startup
After the parameter setting is finished, detect software by the foreground of central server and send the paper defects sense command, this order is sent to the embedded paper defects measurement processor in each road by Ethernet, starts online paper defects and detects;
(3) page image acquisition
The embedded paper defects measurement processor in each road is saved in the page view data of collected by camera in the dual port RAM in the FPGA by the image acquisition units in the FPGA and Camera Link interface;
(4) image pre-service
Embedded paper defects measurement processor is passed through the image pretreatment unit reads image data from dual port RAM in the FPGA, and it is carried out medium filtering and the linear preconditioning of segmentation gray scale;
(5) paper defects detects and the location
Pretreated view data input paper defects detects DSP, and paper defects detects DSP and utilizes thresholding method that speck, scratch class high brightness paper defects zone and blackspot, fold class low-light level paper defects zone in the image are separated from background respectively; Then calculate the circularity in each paper defects zone, and distinguish hole and scratch, blackspot and fold according to circularity;
(6) have paper defects if paper defects detection DSP detects in the page image, then type, area, location parameter and the corresponding paper defects image with each paper defects is sent to central server by gigabit Ethernet;
(7) central server carries out update displayed by a plurality of subwindows to multichannel testing result and paper defects image, simultaneously it is saved in the background data base.
In the described step (1), detection software on the central server is directly called in internal memory with parameters such as the line frequency of each road high speed camera, time shutter, image sizes from hard disk, then be sent to the image acquisition units in the FPGA in the embedded paper defects measurement processor by Ethernet, image acquisition units produces camera control signal thus, and is sent to high speed camera by LVTTL/LVDS conversion chip and Camera Link cable.
In the described step (3), the effective pixel data of high speed camera output image is sent to the image acquisition units in the FPGA through Camera Link stube cable and LVDS/LVTTL interface conversion chip; Image acquisition units, is written in the dual port RAM as writing the view data of clock with high speed camera with pixel clock as with the effect signal with line synchronizing signal.
In the described step (4), after collecting a frame image data, image pretreatment unit in the FPGA reads the view data in the dual port RAM, according to picture size image being carried out the border cutting first processes, then carry out medium filtering, disturb with Removing Random No, carry out again afterwards piecewise linear gray transformation, strengthen the paper defects zone in the image, suppress its background area; The pre-service result is saved in the outer high-speed SRAM of FPGA; After pre-service finished, the image pretreatment unit detected DSP to paper defects and sends look-at-me.
In the described step (5), after paper defects detection DSP receives look-at-me, read view data among the SRAM; Utilize first Threshold segmentation publish picture picture Hole and scratch class high brightness paper defects zone and spot, fold class low-light level paper defects zone; After utilizing opening operation to remove noise, the recycling labelling method is determined position and the area in each paper defects zone in the image; Then calculate each paper defects area circumference square and obtain circularity with Area Ratio, the zone that circularity is larger is fold or scratch, and circularity is blackspot and speck than the zonule.
In the described step (6), after paper defects detection DSP finishes image detection, if the number of paper defects is non-vanishing, then the parameters such as the type of each paper defects, area, position are sent to the Ethernet interface control module of FPGA, paper defects parameter, paper defects image and affiliated camera number are packed in this unit, and the control kilomega network controls and transceiving chip is sent to Ethernet with data, is sent to central server by Ethernet again.
In the described step (7), detect on the software interface on the foreground of central server, each camera has the subwindow of a correspondence, is used for showing this road camera calibration result; After software receives the data that Ethernet sends, according to the phase plane No. under it with the paper defects image that gathers and corresponding paper defects parameter display in the subwindow of correspondence, and it is saved in the background data base.
The utility model has adopted multichannel Industry thread array CCD camera side by side to gather motion paper web view data, view data is sent to each self-corresponding embedded paper defects measurement processor through Camera Link interface, because the transfer rate of Camera Link interface reaches as high as 2.1Gbps, can guarantee the high speed of data acquisition.
Embedded paper defects measurement processor realizes that by the paper defects image being carried out analyzing and processing paper defects detects, and is the core of whole detection system, also is the bottleneck of the whole detection system speed of restriction.In order to guarantee the high speed of image analysis processing, this unit is realized by FPGA and DSP cooperation, is taken full advantage of the high-speed data processing power of DSP and the complex logic processing power of FPGA.After embedded paper defects measurement processor was finished the paper defects detection, testing result and corresponding paper defects image were sent to central server by gigabit Ethernet.Central server is by computing machine and software is detected on operation foreground on computers and background data base forms.Server receives the testing result of each road detecting unit, and shows and preserve.
The beneficial effects of the utility model are: the Quick Acquisition of having realized the page image by Camera Link interface line array CCD camera; Utilize FPGA and DSP flush bonding processor to realize that quick paper defects detects and the location; Utilize gigabit Ethernet to guarantee the fast transport of testing result and paper defects image.The paper defects on-line detecting system that makes up has that cost is low, the speed room for promotion large, be easy to realize the Distributed Detection advantage, can be applicable at a high speed online detection of common paper defects such as blackspot, hole, fold, scratch that the speed of a motor vehicle is not less than the motion paper web of 1km/min.
Description of drawings:
Fig. 1 is based on the distributed web inspection system structural representation of flush bonding processor;
The embedded paper defects measurement processor of Fig. 2 theory diagram;
Fig. 3 paper defects detection algorithm overview flow chart;
Fig. 4 determines paper defects area, position and separates the algorithm flow chart of hole, scratch;
Fig. 5 a-g example paper defects image and paper defects type identification result.
Wherein, 1. high speed camera, 2. motion paper web, 3.LED light source, 4. embedded paper defects measurement processor, 5. Ethernet switch, 6. central server, 7. LVDS/ LVTTL conversion chip, 8. LVTTL/LVDS conversion chip, 9. image acquisition units, 10. dual port RAM, 11. image pretreatment units, 12. paper defects detects DSP, 13. high-speed SRAM, 14. Ethernet interface control modules, the control of 15. kilomega networks and transceiving chip.
Embodiment
The utility model is described in further detail below in conjunction with drawings and Examples.
As shown in Figure 1, the distributed web inspection system based on flush bonding processor mainly comprises led light source 3, motion paper web 2, high speed camera 1, embedded paper defects measurement processor 4, Ethernet switch 5 and central server 6.Led light source 3 is installed in the below of motion paper web 2, and high speed camera 1 is comprised of the multi-route array CCD camera, laterally is installed in side by side the top of motion paper web 2.Every road camera is connected to separately embedded paper defects measurement processor 4 by Camera Link interface.As shown in Figure 2, embedded paper defects measurement processor 4 mainly detects DSP12 by the image acquisition units 9 in the FPGA, paper defects image pretreatment unit 11, Ethernet interface control module 14, dual port RAM 10 and paper defects and forms, and comprises in addition the external interface chips such as the outer high-speed SRAM 13 of sheet and LVDS/LVTTL conversion chip 7, LVTTL/LVDS conversion chip 8, kilomega network control and transceiving chip 15.Ethernet interface control module 14 in the embedded paper defects measurement processor 4 in each road is connected to central server 6 by Ethernet switch 5 again by being connected to Ethernet switch 5 with kilomega network control with transceiving chip 15.Central server 6 is by computing machine and software is detected on operation foreground on computers and background data base forms.
Paper defects detection method of the present utility model is:
(1) parameter setting.Software is called in each road high speed camera 1 from hard disk the parameters such as line frequency, time shutter, image size are detected on the foreground of central server 6, and parameters is sent to embedded paper defects measurement processor 4 by Ethernet.As shown in Figure 2, the image acquisition units 9 in the FPGA produces the camera control signal of TTL form, and this control signal converts the LVDS form to by LVTTL/ LVDS conversion chip 8, and is sent to each road high speed camera 1 by Camera Link stube cable.
(2) detect startup.Send the paper defects sense command by software interface, this order is sent to embedded paper defects measurement processor 4 by Ethernet, starts online paper defects and detects.
(3) image data acquiring.As shown in Figure 2, the view data of high speed camera 1 output is sent to LVDS to LVTTL conversion chip 7 through Camera Link stube cable, and this chip converts 4 pairs of LVDS data-signals and 1 pair of LVDS clock signal to TTL form 28 bit data and 1 tunnel clock signal.Comprise frame synchronizing signal (FVAL, invalid in the native system) in 28 bit data, row has a synchronizing signal (LVAL).Image acquisition units 9 in the FPGA, writes in the dual port RAM 10 in the FPGA as writing the view data of clock with high speed camera 1 with pixel clock as with imitating signal with LVAL.Being stored in the dual port RAM 10 is 8 gray level images, totally 256 gray shade scales.
(4) image pre-service.After collecting a frame image data, image pretreatment unit 9 in the FPGA reads the view data of dual port RAM 10, as shown in Figure 3, according to picture size image being carried out the border cutting first processes, the Crisscross template of recycling 3 * 3 carries out medium filtering, then utilize piecewise linear gray transformation that the paper defects zone is strengthened processing, Fig. 5 a is that a width of cloth is through paper defects image (512 * 512) after the pre-service.Pretreated view data is deposited in the high-speed SRAM 13.After pre-service finished, the image pretreatment unit 9 in the FPGA detected DSP12 to paper defects and sends interrupt request.
(5) paper defects detects.After paper defects detection DSP12 receives interrupt request, read view data in the high-speed SRAM 13, and carry out paper defects and detect and the location.Shown in Fig. 5 a, because led light source 3 is below motion paper web 2, hole and scratch belong to the high-brightness region in the image, and spot and fold are the low brightness area that belongs in the image.As shown in Figure 3, utilize less empirical value
T 1Be partitioned into the low-light level paper defects zones such as blackspot, fold, segmentation result is shown in Fig. 5 b.The recycling opening operation is removed noise, and the mark connected domain is determined paper defects position, area, then calculates the circularity of regional, and what circularity was less is blackspot, and what circularity was larger is fold, as shown in Figure 4.The blackspot recognition result is shown in Fig. 5 c, and the fold recognition result is shown in Fig. 5 d.As shown in Figure 3, utilize larger empirical value
T 2Be partitioned into the high brightness paper defects zones such as image Hole, scratch, segmentation result is shown in Fig. 5 e.Use the same method again and distinguish hole and scratch and determine its position, area parameters.The hole recognition result is shown in Fig. 5 f, and the scratch recognition result is shown in Fig. 5 g.
(6) testing result transmits.After paper defects detection DSP12 finishes image detection, if the number of paper defects is non-vanishing, then the data such as the type of each paper defects, area, position are reached the Ethernet interface control module 14 in the FPGA, number pack paper defects parameter, paper defects image and affiliated high speed camera in this unit, by kilomega network control and transceiving chip 15 data are sent to Ethernet again, are sent to central server 6 by Ethernet again.
(7) testing result is upgraded, is shown.Foreground at central server 6 is detected on the software interface, and every road high speed camera 1 has the subwindow of a correspondence, is used for showing this road high speed camera 1 testing result.After software receives the data that Ethernet sends, on the one hand according to the high speed camera under it number with paper defects image and corresponding paper defects parameter display in the subwindow of correspondence, it is saved in the background data base of central server 6 on the other hand.
Claims (2)
1. the distributed web inspection system based on flush bonding processor is characterized in that it comprises at least one group of high speed camera, and high speed camera is installed in motion paper web top, then is provided with led light source below the motion paper web; High speed camera is connected with embedded paper defects measurement processor, and embedded paper defects measurement processor is communicated by letter with central server by Ethernet switch; Wherein, embedded paper defects measurement processor is realized by FPGA and DSP cooperation, comprise the image acquisition units in the FPGA, it is communicated by letter with transceiving chip successively with dual port RAM, paper defects image pretreatment unit, Ethernet interface control module and the FPGA kilomega network control outward in the FPGA; Paper defects image pretreatment unit detects DSP with paper defects, high-speed SRAM is communicated by letter; Image acquisition units is by LVDS/ LVTTL conversion chip acquisition of image data, by LVTTL/ LVDS conversion chip control high speed camera; Paper defects image pretreatment unit receives the view data of image acquisition units it is carried out pre-service, and paper defects detects the view data that DSP receives paper defects image pretreatment unit, and carries out paper defects and detect; The Ethernet interface control module is controlled kilomega network control and the camera parameters of transceiving chip reception from Ethernet on the one hand, and be sent to image acquisition units, receive on the other hand the paper defects testing result data from the image pretreatment unit, and the control of control kilomega network is sent to Ethernet with transceiving chip.
2. the distributed web inspection system based on flush bonding processor as claimed in claim 1 is characterized in that, described high speed camera is comprised of multichannel Industry thread array CCD camera, laterally is installed in side by side the top of motion paper web; Every road camera is connected to separately embedded paper defects measurement processor by Camera Link cable.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201220306189 CN202676617U (en) | 2012-06-27 | 2012-06-27 | Embedded processor-based distributed paper defect detection system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201220306189 CN202676617U (en) | 2012-06-27 | 2012-06-27 | Embedded processor-based distributed paper defect detection system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN202676617U true CN202676617U (en) | 2013-01-16 |
Family
ID=47497516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201220306189 Expired - Fee Related CN202676617U (en) | 2012-06-27 | 2012-06-27 | Embedded processor-based distributed paper defect detection system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN202676617U (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102721702A (en) * | 2012-06-27 | 2012-10-10 | 山东轻工业学院 | Distributed paper defect detection system and method based on embedded processor |
| CN103592312A (en) * | 2013-11-14 | 2014-02-19 | 无锡港湾网络科技有限公司 | On-line detection method for cables on basis of machine vision |
| CN113504241A (en) * | 2021-07-13 | 2021-10-15 | 陕西科技大学 | Design method of paper defect detection system |
-
2012
- 2012-06-27 CN CN 201220306189 patent/CN202676617U/en not_active Expired - Fee Related
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102721702A (en) * | 2012-06-27 | 2012-10-10 | 山东轻工业学院 | Distributed paper defect detection system and method based on embedded processor |
| CN103592312A (en) * | 2013-11-14 | 2014-02-19 | 无锡港湾网络科技有限公司 | On-line detection method for cables on basis of machine vision |
| CN113504241A (en) * | 2021-07-13 | 2021-10-15 | 陕西科技大学 | Design method of paper defect detection system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102721702A (en) | Distributed paper defect detection system and method based on embedded processor | |
| CN103399016A (en) | Online detection system for surface defects of coldly-rolled aluminum plate and detection method of online detection system | |
| CN107014819A (en) | A kind of solar panel surface defects detection system and method | |
| CN105388162A (en) | Raw material silicon wafer surface scratch detection method based on machine vision | |
| CN203231981U (en) | Online paper defect detecting system | |
| CN105044126A (en) | Visual detection system for large-width continuous surface defects | |
| CN103872983A (en) | Device and method for detecting defects on surface of solar cell | |
| CN102020036A (en) | Visual detection method for transparent paper defect of outer package of strip cigarette | |
| CN101482518A (en) | On-line quality detection system for movable band-shaped material | |
| CN202676617U (en) | Embedded processor-based distributed paper defect detection system | |
| CN105424709A (en) | Fruit surface defect detection method based on image marking | |
| CN102608130A (en) | Smart card stain detecting system based on image feature matching technology and detection and detecting method | |
| CN102692216A (en) | Real-time optical fiber winding defect detection method based on machine vision technology | |
| CN102323272A (en) | Filter paper defect detecting system and detection method thereof based on machine vision technique | |
| CN103473778A (en) | Detecting algorithm for eccentrically-inserting defect of LED luminous chip | |
| CN102879404B (en) | System for automatically detecting medical capsule defects in industrial structure scene | |
| CN103822926B (en) | Holographic shading radium-shine wrappage press quality amount detection systems and detection method | |
| CN205721986U (en) | A kind of cable visual defects detection and character recognition device | |
| CN202177587U (en) | Filter paper defect detecting system based on machine vision technology | |
| CN204882402U (en) | Detection apparatus for chip resistor and carrier band | |
| CN203337559U (en) | Digitized full-automatic detection system for quality of presswork | |
| CN103091330A (en) | Copper strip surface defect identifying device for simulating human visual perception mechanism and method thereof | |
| CN103175839A (en) | Processing method and system for detection of offset plate surface | |
| CN102854195B (en) | Method for detecting defect coordinates on color filter | |
| CN202676612U (en) | Device used for testing LED circuit chip |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130116 Termination date: 20130627 |