CN204330614U - A kind of on-line measuring device based on multimodal registration - Google Patents
A kind of on-line measuring device based on multimodal registration Download PDFInfo
- Publication number
- CN204330614U CN204330614U CN201420807794.8U CN201420807794U CN204330614U CN 204330614 U CN204330614 U CN 204330614U CN 201420807794 U CN201420807794 U CN 201420807794U CN 204330614 U CN204330614 U CN 204330614U
- Authority
- CN
- China
- Prior art keywords
- registration
- unit
- pattern registration
- subelement
- module
- 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
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The utility model discloses a kind of on-line measuring device based on multimodal registration, comprise and be installed on pattern registration unit above mechanical carrying platform and intelligent lighting assembly from top to bottom successively; Manual intervention cellular installation is in the right side of pattern registration unit, and pattern registration unit is made up of multiple pattern registration subelement, and each pattern registration subelement is all connected with CPU (central processing unit) by interaction data communication bus respectively with manual intervention unit.The utility model carries out registration analysis to detecting multiple characteristic informations of target simultaneously, realize the classification of multiple characteristic information, collect and comprehensively, realize efficiently, on-line checkingi identification accurately; Combine the result of multiple characteristic information registration, effectively realize the compression of multicharacteristic information, reduce redundancy feature information, greatly alleviate the pressure of hardware system processing image information, improve systems axiol-ogy precision, the analysis and decision of the Automated condtrol that the result of comprehensive multimodal registration information is production line provides effective support.
Description
Technical field
The utility model belongs to a kind of on-line measuring device, is specifically related to a kind of on-line measuring device based on multimodal registration.
Background technology
In modern industrial and agricultural production process, online measuring technique is widely used in fields such as production, quality restriction and quality controls, and one of basic technology becoming modern industry gradually.In such applications, by carrying out registration identification to certain feature of various product component, to realize the detection of the defect of product, cut, spot and shape, installation site, and then eliminate or reduce defect ware.Comparatively speaking, this is a kind of traditional image matching method, needs the information such as gradation of image to detecting target to carry out comparatively strict process, and by the optimization process of a series of vision algorithm, just can reach satisfied accuracy of identification requirement.
But inevitably, all there is the phenomenons such as various tension fluctuation, noise in most production line, the problems such as the stretcher strain that all there are different situations more or less of the actual detection target image got, background complexity.Along with the complicated of on-line checkingi application scenario and improving constantly testing requirement, single characteristic information is relied on to be difficult to describe the full detail detected in scene, and traditional on-line detecting system utilizes single characteristics of image pattern to carry out detection identification, under the conditions such as specific light source, can only obtain the characteristic information detecting target, and often final process is rear as detecting this characteristic information the requirement not meeting expection of identifying coupling at every turn.
Simultaneously, the uncertainty of Test Field brings many difficulties to the operational efficiency of on-line detecting system, hardware cost, algorithm complex, be among the occasion of change owing to detecting the condition such as target location, light source, many times utilizing single feature to carry out recognition detection target, be difficult to reach desirable effect.
At many scenes needing on-line checkingi, many times need to detect multiple features of product, as production line detects the surface condition of engineering goods, inner structure damage, installation site etc., single its efficiency of pattern method for registering and accuracy is now utilized obviously to be difficult to meet the demands, for quick and realization accurately detect the on-line checkingi of target, be necessary that design is a kind of simply, reliable, the technical solution of low cost.
Utility model content
The utility model proposes to overcome the shortcoming existed in prior art, its objective is and provides a kind of on-line measuring device based on multimodal registration.
The technical solution of the utility model is:
Based on an on-line measuring device for multimodal registration, described pick-up unit comprises mechanical carrying platform, intelligent lighting assembly, pattern registration unit, manual intervention unit, interaction data communication bus and CPU (central processing unit); Pattern registration unit and intelligent lighting assembly are installed on above mechanical carrying platform from top to bottom successively; Manual intervention cellular installation is in the right side of pattern registration unit, and pattern registration unit is made up of multiple pattern registration subelement, and each pattern registration subelement is all connected with CPU (central processing unit) by interaction data communication bus respectively with manual intervention unit.
Described pattern registration subelement is made up of camera lens, sensor, driving circuit, STM32 processor; Driving circuit is made up of signal acquisition module and high-speed pulse module, STM32 processor produces signal acquisition module and the time sequential pulse needed for high-speed pulse module work, high-speed pulse module produces the high speed time series pulse signals needed for working sensor, the signal of signal acquisition module pick-up transducers, and signal is conveyed into the process of STM32 processor.
Described intelligent lighting assembly is made up of multiple lighting source and light source communication module, each lighting source is installed successively with the direction of motion detecting target, and each lighting source and each pattern registration subelement one_to_one corresponding are arranged, each lighting source is by lamp body, multiple light emitting source, optics auxiliary accessories and multiple drive control module composition, lamp body is made up of the arc that two are oppositely arranged, light emitting source is distributed in lamp body inner periphery, optics auxiliary accessories is installed on below lamp body, each light emitting source and each drive control module connect one to one, each drive control module is all connected with light source communication module, light source communication module is connected with CPU (central processing unit).
Described CPU (central processing unit) is made up of FPGA, DSP, SDRAM and data storage cell, and FPGA connects interaction data communication bus, and connect between FPGA and DSP, DSP is connected with data storage cell, DSP with SDRAM is connected.
The beneficial effects of the utility model are:
The utility model achieves multiple pattern registration unit in a detection system, and multiple characteristic informations that can simultaneously realize detecting target carry out registration analysis; The final Output rusults of detection system draws after all registration unit results carry out comprehensive analysis processing, the simultaneously real-time participation of manual intervention unit in testing process, have modified the detection inefficacy or mistake that bring because field working conditions is complicated, the intellectuality being conducive to Expert Rules is perfect; On-line checkingi based on multimodal registration achieve multiple characteristic information classification, to collect and comprehensively, contribute to the sample classification statistics realizing detecting sample, realize efficiently, on-line checkingi identification accurately, improve the automated production level of whole production line; Combine the result of multiple characteristic information registration, effectively achieve the compression of multicharacteristic information, decrease redundancy feature information, alleviate the pressure of hardware system processing feature information, improve systems axiol-ogy precision; The result combining multimodal registration information is that the analysis and decision that production line automation controls provides effective support.
Accompanying drawing explanation
Fig. 1 is the structural representation of a kind of on-line measuring device based on multimodal registration of the utility model;
Fig. 2 is the structural representation of pattern registration unit in the utility model;
Fig. 3 is the structural representation of pattern registration subelement in the utility model;
Fig. 4 is the structural representation of the lighting source of intelligent lighting assembly in the utility model;
Fig. 5 is the structural representation of CPU (central processing unit) in the utility model.
Wherein:
1 mechanical carrying platform 2 detects target
3 intelligent lighting assembly 4 pattern registration unit
5 manual intervention unit 6 interaction data communication buses
7 CPU (central processing unit) 8 pattern registration subelements
9 STM32 processor 10 driving circuits
11 sensor 12 camera lenses
13 signal acquisition module 14 high-speed pulse modules
15 optics auxiliary accessories 16 lamp bodies
17 light emitting source 18 drive control modules
19 light source communication module 20 FPGA
21 DSP 22 data storage cells
23 SDRAM。
Embodiment
Below in conjunction with Figure of description and embodiment, a kind of on-line measuring device based on multimodal registration of the utility model is described in detail:
As shown in Figure 1, a kind of on-line measuring device based on multimodal registration, comprises mechanical carrying platform 1, intelligent lighting assembly 3, pattern registration unit 4, manual intervention unit 5, interaction data communication bus 6 and CPU (central processing unit) 7; Pattern registration unit 4 and intelligent lighting assembly 3 are installed on above mechanical carrying platform 1 from top to bottom successively; Manual intervention unit 5 is installed on the right side of pattern registration unit 4, pattern registration unit 4 is made up of multiple pattern registration subelement 8, and each pattern registration subelement 8 is all connected with CPU (central processing unit) 7 by interaction data communication bus 6 respectively with manual intervention unit 5.
As shown in Figure 2,3, described pattern registration subelement 8 is made up of camera lens 12, sensor 11, driving circuit 10, STM32 processor 9; Driving circuit 10 achieves driving to sensor 11 and control, and it is made up of signal acquisition module 13 and high-speed pulse module 14, produces signal acquisition module 13 and high-speed pulse module 14 to work required time sequential pulse by STM32 processor 9.High-speed pulse module 14 is responsible for the high speed time series pulse signals that generation sensor 11 works required, and the collection of signal acquisition module 13 primary responsibility sensor 11 signal, the signal collected finally is sent into STM32 processor 9 and processed.According to detecting the difference of target 2, the sensor that the sensor 11 in each pattern registration subelement 8 can select Visible-light CCD, infrared CCD etc. dissimilar, and the pattern registration subelement 8 that can configure repetition is as required to realize the object of duplicate detection.Camera lens 12 in each pattern registration subelement 8, driving circuit 10 all require to design according to sensor 11 difference in this pattern registration subelement 8 and concrete registration.STM32 processor 9 is responsible for the execution of pattern registration subelement 8 registration action and is carried out real-time, interactive with CPU (central processing unit) 7, and in the present embodiment, the STM32F103 Series MCU of ST company selected by STM32 processor 9.
Described manual intervention unit 5 is made up of the input equipment such as keyboard or touch-screen, in order to directly to input intervention rule, it is installed on the right side of last pattern registration subelement 8 of pattern registration unit 4, when all pattern registration subelements 8 all fail to export effective testing result or testing result undesirable, intervened according to actual conditions by manual intervention unit 5, artificial court verdict is provided.The registration rule that manual intervention unit 5 is formed can in the middle of real-time update to Expert Rules.
As shown in Figure 4, described intelligent lighting assembly 3 is made up of multiple lighting source and light source communication module 19, each lighting source is installed successively with the direction of motion (as shown by the arrows in Figure 1) detecting target 2, and each lighting source and each pattern registration subelement 8 one_to_one corresponding are arranged, each lighting source is by lamp body 16, multiple light emitting source 17, optics auxiliary accessories 15 and multiple drive control module 18 form, lamp body 16 is made up of the arc that two are oppositely arranged, light emitting source 17 is distributed in lamp body 16 inner periphery, optics auxiliary accessories 15 is installed on below lamp body 16, each light emitting source 17 connects one to one with each drive control module 18, each drive control module 18 is all connected with light source communication module 19, light source communication module 19 is connected with CPU (central processing unit) 7.Each lighting source in intelligent lighting assembly 3 is not all according to pattern registration subelement 8 each in pattern registration unit 4 lighting source that this subelement provides different, and each lighting source all can provide different illumination direction, irradiating angle, light source colour and different light source light spectrums as required.During print character as detection chip surface, light source can be controlled and throw light on low angle, thus outstanding character.In each lighting source, drive control module 18 is all connected to accept light source configuration order with light source communication module 19.Lamp body 16 can according to detect need be designed to vaulted or annular structure, light emitting source 17 can select LED, the light sources such as infrared Halogen lamp LED, drive control module 18 mainly realizes driving and the control of light source, reasonably select optics auxiliary accessories 15 effectively can improve the lighting quality of detection system in the utility model, generally the optical element such as polaroid or diffusing panel can be selected according to field working conditions and the actual requirement optics auxiliary accessories 15 detected, drive control module 18 is responsible for the drived control of light source, and CPU (central processing unit) 6 can realize the real-time control to certain lighting source independent by light source communication module 19.
As shown in Figure 5, described CPU (central processing unit) 7 is responsible for the coordination and control of whole detection system, its each pattern registration subelement 8 by interaction data communication bus 6 and pattern registration unit 4 and manual intervention unit 5 are connected, realize when registration runs and the real-time, interactive of each registration subelement data, it is by realizing to lighting source corresponding to each registration subelement configuration with the connection of intelligent lighting assembly 3.CPU (central processing unit) 6 forms primarily of on-site programmable gate array FPGA (Field Programmable Gate Array) 20, digital signal processor DSP (Digital Signal Processor) 21, synchronous DRAM SDRAM (Synchronous Dynamic Random Access Memory) 23 and data storage cell 22.Wherein FPGA 20 is as the coprocessor of DSP signal processor, mainly complete light source control, drive and pre-service, bus management and some low orders logic function, and DSP primary responsibility completes the high-order Processing Algorithm such as regular sample learning, result of decision process.Storage when SDRAM 23 and data storage cell 22 are respectively used to CPU (central processing unit) 6 high speed operation and reading, as the intelligent learning and real-time update etc. of Expert Rules.This platform architecture, improves system works efficiency, improves the overall performance of system.
Using method of the present utility model:
(I) initialization
According to the testing requirement detecting target 2, utilize artificial collection to detect sample, tentatively determine suitable registration pattern and registration rule (S1);
(II) optimizes registration rule
Comprehensive expert's registration rule is set, in conjunction with field working conditions, successively in deterministic model registration unit 4 each pattern registration subelement 8 registration rule and be optimized, in the data storage cell 22 of registration after optimization rule input CPU (central processing unit) 7 (S2);
(III) pattern registration unit configures
According to the rule in step (II) and the concrete testing requirement detecting target 2, each pattern registration subelement 8 in pattern registration unit 4 is configured, comprises the selection (S3) of sensor 11 and camera lens 12;
(IV) lighting source configures
The selection of respective sensor 11 and camera lens 12 in integrating step (III), select suitable illumination sources, controlled by each lighting source in CPU (central processing unit) 7 pairs of intelligent lighting assemblies 3, the pattern registration subelement 8 corresponding according to each lighting source, configures suitable light level scheme (S4);
(V) starts registration procedure
The registration rule of each pattern registration subelement 8 is issued to this pattern registration subelement 8 by interaction data communication bus 6 by CPU (central processing unit) 7, when detection target 2 is successively through i-th pattern registration subelement, the i-th i pattern registration subelement 2 ... during the subelements such as the n-th pattern registration subelement, associative mode registration subelement 8 brings into operation S5;
(VI) registration result judges
In pattern registration unit 4, registration result is delivered to CPU (central processing unit) 7 by each pattern registration subelement 8 respectively, and CPU (central processing unit) 7 carries out comprehensive calculation process (S6) to each registration result;
(VII) revises Expert Rules
If with the expection arranged, to detect target inconsistent, directly revises comprehensive expert's registration rule (S7), restart step (I) and perform detection, until testing result meets expection by manual intervention unit 5;
(VIII) terminates
If detect target with the expection arranged to be consistent, then this time detects and terminates (S8).
Major advantage of the present utility model has:
(i) in a detection system, achieve multiple pattern registration unit, multiple characteristic informations that can simultaneously realize detecting target carry out registration analysis.
(ii) the final Output rusults of detection system draws after all registration unit results carry out comprehensive analysis processing, the simultaneously real-time participation of manual intervention unit in testing process, have modified the detection inefficacy or mistake that bring because field working conditions is complicated, the intellectuality being conducive to Expert Rules is perfect.
(iii) based on the on-line checkingi of multimodal registration achieve multiple characteristic information classification, collect and comprehensively, contribute to the sample classification statistics realizing detecting sample, realize efficiently, on-line checkingi identification accurately, improve the automated production level of whole production line.
(iv) combine the result of multiple characteristic information registration, effectively achieve the compression of multicharacteristic information, decrease redundancy feature information, alleviate the pressure of hardware system processing feature information, improve systems axiol-ogy precision.
(v) the result combining multimodal registration information is that the analysis and decision that production line automation controls provides effective support.
Claims (4)
1. based on an on-line measuring device for multimodal registration, it is characterized in that: described pick-up unit comprises mechanical carrying platform (1), intelligent lighting assembly (3), pattern registration unit (4), manual intervention unit (5), interaction data communication bus (6) and CPU (central processing unit) (7); Pattern registration unit (4) and intelligent lighting assembly (3) are installed on mechanical carrying platform (1) top from top to bottom successively; Manual intervention unit (5) is installed on the right side of pattern registration unit (4), pattern registration unit (4) is made up of multiple pattern registration subelement (8), and each pattern registration subelement (8) is all connected with CPU (central processing unit) (7) by interaction data communication bus (6) respectively with manual intervention unit (5).
2. a kind of on-line measuring device based on multimodal registration according to claim 1, is characterized in that: described pattern registration subelement (8) is made up of camera lens (12), sensor (11), driving circuit (10), STM32 processor (9); Driving circuit (10) is made up of signal acquisition module (13) and high-speed pulse module (14), STM32 processor (9) produces signal acquisition module (13) and the time sequential pulse needed for high-speed pulse module (14) work, high-speed pulse module (14) produces the high speed time series pulse signals needed for sensor (11) work, the signal of signal acquisition module (13) pick-up transducers (11), and signal is conveyed into STM32 processor (9) process.
3. a kind of on-line measuring device based on multimodal registration according to claim 1, it is characterized in that: described intelligent lighting assembly (3) is made up of multiple lighting source and light source communication module (19), each lighting source is installed successively with the direction of motion detecting target (2), and each lighting source and each pattern registration subelement (8) one_to_one corresponding are arranged, each lighting source is by lamp body (16), multiple light emitting source (17), optics auxiliary accessories (15) and multiple drive control module (18) composition, lamp body (16) is made up of the arc that two are oppositely arranged, light emitting source (17) is distributed in lamp body (16) inner periphery, optics auxiliary accessories (15) is installed on lamp body (16) below, each light emitting source (17) and each drive control module (18) connect one to one, each drive control module (18) is all connected with light source communication module (19), light source communication module (19) is connected with CPU (central processing unit) (7).
4. a kind of on-line measuring device based on multimodal registration according to claim 1, it is characterized in that: described CPU (central processing unit) (7) is by FPGA(20), DSP(21), SDRAM(23) and data storage cell (22) form, FPGA(20) interaction data communication bus (6) is connected, FPGA(20) connect and DSP(21), DSP(21) be connected with data storage cell (22), DSP(21) with SDRAM(23) be connected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201420807794.8U CN204330614U (en) | 2014-12-19 | 2014-12-19 | A kind of on-line measuring device based on multimodal registration |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201420807794.8U CN204330614U (en) | 2014-12-19 | 2014-12-19 | A kind of on-line measuring device based on multimodal registration |
Publications (1)
Publication Number | Publication Date |
---|---|
CN204330614U true CN204330614U (en) | 2015-05-13 |
Family
ID=53166963
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201420807794.8U Expired - Fee Related CN204330614U (en) | 2014-12-19 | 2014-12-19 | A kind of on-line measuring device based on multimodal registration |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN204330614U (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104483325A (en) * | 2014-12-19 | 2015-04-01 | 核工业理化工程研究院华核新技术开发公司 | Multi-mode registration-based online detecting device and multi-mode registration-based online detecting method |
CN106323600A (en) * | 2016-08-31 | 2017-01-11 | 武汉精测电子技术股份有限公司 | Cascaded distributed AOI defect detection system and detection method thereof |
CN104483324B (en) * | 2014-12-19 | 2017-10-20 | 核工业理化工程研究院华核新技术开发公司 | On-line measuring device based on multimodal registration |
-
2014
- 2014-12-19 CN CN201420807794.8U patent/CN204330614U/en not_active Expired - Fee Related
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104483325A (en) * | 2014-12-19 | 2015-04-01 | 核工业理化工程研究院华核新技术开发公司 | Multi-mode registration-based online detecting device and multi-mode registration-based online detecting method |
CN104483324B (en) * | 2014-12-19 | 2017-10-20 | 核工业理化工程研究院华核新技术开发公司 | On-line measuring device based on multimodal registration |
CN104483325B (en) * | 2014-12-19 | 2017-12-15 | 华核(天津)新技术开发有限公司 | On-line measuring device and detection method based on multimodal registration |
CN106323600A (en) * | 2016-08-31 | 2017-01-11 | 武汉精测电子技术股份有限公司 | Cascaded distributed AOI defect detection system and detection method thereof |
CN106323600B (en) * | 2016-08-31 | 2020-04-24 | 武汉精测电子集团股份有限公司 | Cascade distributed AOI defect detection system and detection method thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102218406B (en) | Intelligent detection device of defects of mobile phone outer shell based on machine vision | |
CN111008640B (en) | Image recognition model training and image recognition method, device, terminal and medium | |
CN102621156B (en) | Image-processing-based automatic micro part sorting system | |
CN204359710U (en) | A kind of glass surface defects pick-up unit | |
CN204330614U (en) | A kind of on-line measuring device based on multimodal registration | |
CN109461149A (en) | The intelligent checking system and method for lacquered surface defect | |
CN113870318B (en) | Moving target detection system and method based on multi-frame point cloud | |
CN109840900A (en) | A kind of line detection system for failure and detection method applied to intelligence manufacture workshop | |
CN102854194A (en) | Object surface defect detection method and apparatus based on linear array CCD | |
TWI709188B (en) | Fusion-based classifier, classification method, and classification system | |
Chen et al. | Defect detection method of aluminum profile surface using deep self-attention mechanism under hybrid noise conditions | |
CN104316033A (en) | Visual inspection system for automobile parts | |
CN113837994B (en) | Photovoltaic panel defect diagnosis method based on edge detection convolutional neural network | |
CN109693140A (en) | A kind of intelligent flexible production line and its working method | |
Yin et al. | PV-YOLO: lightweight YOLO for photovoltaic panel fault detection | |
CN104483325A (en) | Multi-mode registration-based online detecting device and multi-mode registration-based online detecting method | |
CN117085969B (en) | Artificial intelligence industrial vision detection method, device, equipment and storage medium | |
Cai et al. | An improved plant disease detection method based on YOLOv5 | |
CN104483324A (en) | Multi-mode registration-based online detecting device | |
CN117541568A (en) | Deep learning-based automobile brake disc surface defect detection method | |
CN109374637A (en) | A kind of automobile panel Surface Quality Inspection System and method | |
Zhang et al. | Unsound wheat kernel recognition based on deep convolutional neural network transfer learning and feature fusion | |
Ahmed et al. | Fresh or Stale: Leveraging Deep Learning to Detect Freshness of Fruits and Vegetables | |
CN117969553B (en) | On-line visual detection system for appearance of TPV knitted composite pipe | |
CN104156989A (en) | Method for detecting position of gear keyway by using Hough linear transformation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150513 Termination date: 20201219 |