CN111947645A - Intelligent tool positioning system and positioning method for automobile assembly - Google Patents
Intelligent tool positioning system and positioning method for automobile assembly Download PDFInfo
- Publication number
- CN111947645A CN111947645A CN202010804853.6A CN202010804853A CN111947645A CN 111947645 A CN111947645 A CN 111947645A CN 202010804853 A CN202010804853 A CN 202010804853A CN 111947645 A CN111947645 A CN 111947645A
- Authority
- CN
- China
- Prior art keywords
- image
- positioning
- module
- positioning system
- intelligent tool
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Biology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Automation & Control Theory (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an intelligent tool positioning system and a positioning method for automobile assembly, wherein the positioning system comprises an image acquisition module, an operation storage device module, a software positioning module, a control device module and an output device module; the image acquisition module, the AI chip and the industrial control computer are integrated in the embedded internet of things system and used for completing image collection and storage locally; the cloud computing and storing comprises the steps of storing collected image samples to a cloud, and training and identifying images at the cloud; the positioning method adopts a deep learning model to perform tool positioning identification, and simultaneously collects the analysis data of the previous n frames of images to classify the analysis result of the current frame of image, thereby further improving the precision of the positioning result; the automobile part manual assembling device can effectively solve the problems that the operation position is easy to miss, the operation efficiency is low and the labor cost is high when the automobile parts are assembled manually in the prior art, and has a good market prospect.
Description
Technical Field
The invention relates to the technical field of automobile assembly, in particular to an intelligent positioning system and a positioning method for automobile assembly.
Background
In the existing automobile part assembling process, sometimes due to environmental requirements, the same series of products have slight differences in some characteristics such as appearance size and shape, the current mode for identifying the characteristics is mainly judged by human eyes, but due to fatigue of people, omission occurs, products mixed with other models in one model can be caused, and the identification speed is low. The assembly of automobile parts is high-precision assembly line type operation, manual assembly is adopted, the operation position is easy to miss, the operation efficiency is low, and the labor cost is high.
Disclosure of Invention
The purpose of the invention is as follows: the invention provides an intelligent tool positioning system and a positioning method for automobile assembly, which solve the problems of human misoperation, low tool operation efficiency and high labor cost in manual assembly in the background art.
The technical scheme is as follows: in order to achieve the purpose, the invention adopts the technical scheme that:
an intelligent tool positioning system for automobile assembly is characterized by comprising an image acquisition module, an operation storage device module, a software positioning module, a control device module and an output device module; the image acquisition module comprises a plurality of cameras and is used for acquiring image samples; the operation storage device module comprises an industrial control computer, and the industrial control computer identifies and processes the acquired image samples; the software positioning module carries out real-time positioning according to the image recognition result; the control equipment module sends out an instruction in real time, and adjusts the assembly position of the part until the assembly operation of the part is completed; and the output equipment module receives the positioning data sent by the software positioning module in real time and displays the positioning data to a user in real time through an interactive operation interface.
Furthermore, the system adopts a mode of combining an embedded internet of things system and cloud computing storage; specifically, the embedded internet of things system comprises an image acquisition module, an AI chip and an industrial control computer which are integrated locally and used for completing image collection and storage locally; the cloud computing and storing comprises the steps of storing collected image samples to a cloud, and training and identifying images at the cloud.
Furthermore, the system adopts a mode of combining an embedded internet of things system and cloud computing storage; specifically, the embedded internet of things system comprises an image acquisition module, an AI chip and an industrial control computer which are integrated locally and used for completing image collection and storage locally; the cloud computing and storing comprises the steps of storing collected image samples to a cloud, and training and identifying images at the cloud.
Further, the industrial control computer receives the image sample acquired by the image acquisition module through a USB interface and processes the image sample by adopting a bus structure.
Furthermore, the output equipment adopts an interactive operation interface to assist a user in completing the work of sample collection, training to identify the required parameter setting and debugging; and reminding the user of the proportion of the finished process to the total process in a progress bar mode in the guiding process.
Further, the software running the interactive operation interface packages and packs the execution code and the dependent environment of the program into a portable mirror image through a packaging and packing mode; the software is encrypted by adopting a dongle encryption mode, and the dongle part is packaged by adopting DLL.
Further, the user opens the dongle device by inputting a password, acquires the unique serial number, reads the sector content, and performs authentication.
A positioning method adopting the intelligent tool positioning system for automobile assembly comprises the following steps:
step S1, collecting an image sample;
step S2, adopting a deep learning model, and using image samples collected by a user segment to carry out identification and training to obtain a pre-trained universal model; carrying out transfer learning by using the trained model parameters; according to the user requirements, adding iterative image samples at the edge equipment end, training and updating the model, and improving the model accuracy so as to adapt to the recognition tasks of various scenes;
step S3, extracting site features, color features and texture features from the image, and then performing fusion linear processing or nonlinear processing on the feature information to obtain a mapping template of the site image and the site position;
and step S4, in the process of real-time identification, matching the information of the processed input image with the trained templates to obtain a position result.
Further, a deep learning model framework, preferably an inclusion V3 network framework, is adopted in the step S2; specifically, firstly, a deep learning framework inclusion V3 performs model training on a large data set ImageNet, then, the trained model parameters are used for transfer learning, and parameter fine tuning is performed on a small sample obtained in the use of a product, so that the method is suitable for recognition tasks of various real scenes.
Further, the deep learning model framework adopted in the step S2 is preferably a VGG16 network framework.
Furthermore, the intelligent positioning system analyzes the result based on the current frame image model, integrates the analysis data of the previous n frames of images, and classifies the analysis result of the current frame image; and n is a preset threshold value and can be freely adjusted according to actual working conditions.
Has the advantages that:
(1) the intelligent tool positioning system and method provided by the invention adopt the most advanced deep learning model at present to recognize the operation position classification in real time. And in the development process, images acquired by the user segment are used for identification and training to obtain a pre-trained universally applicable model. And then, according to the requirements of the user, iterative pictures can be added at the edge equipment end, the updated model is trained, and the model accuracy is improved. Meanwhile, the previous n frames of images are subjected to fusion processing and classified, so that the accuracy of the positioning result is further improved, and the instability of a real-time result is relieved.
(2) The output equipment of the intelligent tool positioning system provided by the invention has a convenient and effective UI operation interface, and gradually assists a user to finish the operations of sample collection, training to recognize the required parameter setting and various debugging operations by using interactive and guided operation interfaces.
(3) The intelligent tool positioning system provided by the invention establishes corresponding application container mirror images aiming at different running platforms, and a user does not need to configure the running environment of a program, thereby simplifying the installation steps and reducing the use threshold. By only packaging necessary dependent packages, the occupied space of the software is reduced, and the installation speed of the software is improved. Meanwhile, the method prevents a user from directly contacting the program code, protects the code from being read, written and tampered, and is convenient for the user to manage the non-code file. The system software is encrypted by the aid of the dongle, and meanwhile confidentiality of the equipment is guaranteed.
(4) The invention adopts a mode of combining an embedded internet of things system and cloud computing storage, and integrates a camera, an AI chip and control, so that a user can locally complete a series of operations from image collection, training to identification, and the safety of data is ensured. According to the invention, the characteristics of cloud computing are reasonably utilized, and complex operation and huge storage are efficiently uploaded to the cloud to complete through a 5G technology; the method reduces the hardware cost of the user, reduces the size of front-end hardware of the user, and has the characteristics of small size and easy portability. Meanwhile, the use permission of each user software is activated at the cloud.
Drawings
FIG. 1 is a block diagram of an intelligent tool positioning system provided by the present invention;
Detailed Description
The present invention will be further described with reference to the accompanying drawings.
Fig. 1 shows an intelligent tool positioning system for automobile assembly, which includes an image acquisition module, an operation storage device module, a software positioning module, a control device module and an output device module; the image acquisition module adopts a plurality of cameras and is used for acquiring image samples. The camera unit may be replaced with a laser scanning device. The operation storage device module comprises an industrial control computer, and the industrial control computer adopts a bus structure to store, train and identify the acquired image samples; the software positioning module carries out real-time positioning according to the image recognition result; the control equipment module sends out an instruction in real time, and adjusts the assembly position of the part until the assembly operation of the part is completed; and the output equipment module receives the positioning data sent by the software positioning module in real time and displays the positioning data to a user in real time through an interactive operation interface.
The system adopts a mode of combining an embedded internet of things system and cloud computing storage; specifically, the embedded internet of things system comprises an image acquisition module, an AI chip and an industrial control computer which are integrated locally and used for completing image collection and storage locally; the cloud computing and storing comprises the steps of storing collected image samples to a cloud, and training and identifying images at the cloud. The cloud computing and storing comprises a training step of storing the collected image samples to the cloud and carrying out image recognition at the cloud. A GPU computing part required by a traditional storage and training process is transferred to a cloud end, and a front-end system only reserves a camera, a 5G wireless module and basic storage required by a storage model and basic computing power required by an operation model required by the acquisition of images. Complex operation and huge storage are efficiently uploaded to the cloud end through a 5G technology to complete the operation; the method reduces the hardware cost of the user, reduces the size of front-end hardware of the user, and has the characteristics of small size and easy portability. Meanwhile, the use permission of each user software is activated at the cloud.
The output equipment adopts interactive and guided operation interfaces to assist a user in completing the operations of sample collection, training to recognize the required parameter setting and various debugging operations; and reminding the user of the proportion of the finished process to the total process in a progress bar mode in the guiding process.
The output equipment adopts an interactive operation interface to assist a user in completing the work of sample collection, training to recognize the required parameter setting and debugging; and reminding the user of the proportion of the finished process to the total process in a progress bar mode in the guiding process.
And the software running the interactive operation interface packages and packs the execution codes and the dependent environment of the program into the portable mirror image through a packaging and packing mode. Corresponding application container images are established for different running platforms, and a user does not need to configure the running environment of a program. Meanwhile, only necessary dependent packages are packaged, so that the occupied space of the software is reduced.
The software is encrypted by adopting a dongle encryption mode. The program is written by high-level language such as python, and the encrypted dog part is written by low-level code such as C, and is packaged into DLL library. A high-level language is used to call a softdog DLL encapsulation library in the project. The dongle is used as an identity identification device, the authentication is an extremely important part, and the dongle device must be opened by a password. According to the development requirement and the use condition of the user, different verification modes such as primary verification, intermediate verification, high-level verification and the like can be adopted. The user opens the dongle device by inputting the password, obtains the unique serial number, reads the sector content, and performs authentication.
The following describes the positioning method of the intelligent tool positioning system for automobile assembly in detail with reference to the embodiments.
And step S1, acquiring a plurality of frame image samples through an image acquisition module in the system.
Step S2, a currently most advanced deep learning model is adopted, the embodiment is described by adopting an inclusion V3 network framework, and other deep learning models, such as a VGG16 framework, also belong to the category of the deep learning model used in the present invention. Identifying and training by using image samples acquired by a user segment to obtain a pre-trained universal model; carrying out transfer learning by using the trained model parameters; according to the user requirements, an iterative image sample is added at the edge equipment end, the model is trained and updated, and the model accuracy is improved so as to adapt to the recognition tasks of various scenes.
Specifically, the deep learning framework inclusion V3 performs model training on a large data set ImageNet, then performs transfer learning by using trained model parameters, and performs parameter fine tuning on a small sample to adapt to recognition tasks of various real scenes.
Likewise, the deep learning model framework can also be equivalently replaced by a VGG16 network framework and the like.
And step S3, extracting site features, color features and texture features from the images, and then performing fusion linear processing or nonlinear processing on the feature information to obtain a mapping template of the site images and the site positions.
And step S4, in the process of real-time identification, matching the information of the processed input image with the trained templates to obtain a position result.
When the intelligent tool positioning method is used for analyzing results based on the current frame image model, the analysis data of the previous n frames of images are further fused, and the analysis results of the current frame of images are classified for improving the positioning result precision and slowing down the instability of real-time results, wherein n is a preset threshold value.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and these are intended to be within the scope of the invention.
Claims (10)
1. An intelligent tool positioning system for automobile assembly is characterized by comprising an image acquisition module, an operation storage device module, a software positioning module, a control device module and an output device module; the image acquisition module comprises a plurality of cameras and is used for acquiring image samples; the operation storage device module comprises an industrial control computer, and the industrial control computer identifies and processes the acquired image samples; the software positioning module carries out real-time positioning according to the image recognition result; the control equipment module sends out an instruction in real time, and adjusts the assembly position of the part until the assembly operation of the part is completed; and the output equipment module receives the positioning data sent by the software positioning module in real time and displays the positioning data to a user in real time through an interactive operation interface.
2. The intelligent tool positioning system for automobile assembly as claimed in claim 1, wherein the system adopts a combination of an embedded internet of things system and cloud computing storage; specifically, the embedded internet of things system comprises an image acquisition module, an AI chip and an industrial control computer which are integrated locally and used for completing image collection and storage locally; the cloud computing and storing comprises the steps of storing collected image samples to a cloud, and training and identifying images at the cloud.
3. The intelligent tool positioning system for automobile assembly as claimed in claim 1, wherein the industrial control computer receives the image samples collected by the image collection module through a USB interface and processes the image samples by adopting a bus structure.
4. The intelligent tool positioning system for automobile assembly as claimed in claim 1, wherein the output device employs an interactive operation interface to assist a user in completing the operations of sample collection, training to identify the required parameter setting and debugging; and reminding the user of the proportion of the finished process to the total process in a progress bar mode in the guiding process.
5. The intelligent tool positioning system for automobile assembly as claimed in claim 4, wherein the software running the interactive operation interface packages the execution code and the dependent environment of the program into a portable image through a package packaging mode; the software is encrypted by adopting a dongle encryption mode, and the dongle part is packaged by adopting DLL.
6. The system of claim 5, wherein the user opens the dongle device by entering a password, obtains the unique serial number, reads the contents of the sector, and performs authentication.
7. Positioning method using the intelligent tool positioning system for automobile assembly according to any of claims 1-6, characterized in that it comprises the following steps:
step S1, collecting an image sample;
step S2, adopting a deep learning model, and using image samples collected by a user segment to carry out identification and training to obtain a pre-trained universal model; carrying out transfer learning by using the trained model parameters; according to the user requirements, adding iterative image samples at the edge equipment end, training and updating the model, and improving the model accuracy so as to adapt to the recognition tasks of various scenes;
step S3, extracting site features, color features and texture features from the image, and then performing fusion linear processing or nonlinear processing on the feature information to obtain a mapping template of the site image and the site position;
and step S4, in the process of real-time identification, matching the information of the processed input image with the trained templates to obtain a position result.
8. The positioning method of the intelligent tool positioning system for automobile assembly according to claim 7, wherein a deep learning model framework, preferably an inclusion V3 network framework, is adopted in the step S2; specifically, firstly, a deep learning framework inclusion V3 performs model training on a large data set ImageNet, then, the trained model parameters are used for transfer learning, and parameter fine tuning is performed on a small sample obtained in the use of a product, so that the method is suitable for recognition tasks of various real scenes.
9. The positioning method for the intelligent tool positioning system for automobile assembly as recited in claim 7, wherein a deep learning model framework, preferably a VGG16 network framework, is adopted in the step S2.
10. The positioning method of the intelligent tool positioning system for automobile assembly according to claim 7, wherein the intelligent positioning system performs analysis results based on a current frame image model, fuses analysis data of previous n frames of images, and classifies the analysis results of the current frame of images; and n is a preset threshold value and can be freely adjusted according to actual working conditions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010804853.6A CN111947645A (en) | 2020-08-12 | 2020-08-12 | Intelligent tool positioning system and positioning method for automobile assembly |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010804853.6A CN111947645A (en) | 2020-08-12 | 2020-08-12 | Intelligent tool positioning system and positioning method for automobile assembly |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111947645A true CN111947645A (en) | 2020-11-17 |
Family
ID=73332258
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010804853.6A Pending CN111947645A (en) | 2020-08-12 | 2020-08-12 | Intelligent tool positioning system and positioning method for automobile assembly |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111947645A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112712501A (en) * | 2020-12-28 | 2021-04-27 | 江苏合泰飞梵科技有限公司 | Rearview mirror assembly production method based on artificial intelligence |
| CN114548947A (en) * | 2022-03-08 | 2022-05-27 | 南昌哈恩网络科技有限公司 | Online office safety processing method and server applied to digitization |
-
2020
- 2020-08-12 CN CN202010804853.6A patent/CN111947645A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112712501A (en) * | 2020-12-28 | 2021-04-27 | 江苏合泰飞梵科技有限公司 | Rearview mirror assembly production method based on artificial intelligence |
| CN114548947A (en) * | 2022-03-08 | 2022-05-27 | 南昌哈恩网络科技有限公司 | Online office safety processing method and server applied to digitization |
| CN114548947B (en) * | 2022-03-08 | 2024-01-12 | 杨建鑫 | Online office security processing method and server applied to digitization |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210174072A1 (en) | Microexpression-based image recognition method and apparatus, and related device | |
| CN111947645A (en) | Intelligent tool positioning system and positioning method for automobile assembly | |
| CN108197318A (en) | Face identification method, device, robot and storage medium | |
| CN106845416B (en) | Obstacle identification method and device, computer equipment and readable medium | |
| WO2022174605A1 (en) | Gesture recognition method, gesture recognition apparatus, and smart device | |
| CN108319916A (en) | Face identification method, device, robot and storage medium | |
| CN111062328A (en) | Image processing method and device and intelligent robot | |
| CN113378764B (en) | Video face acquisition method, device, equipment and medium based on clustering algorithm | |
| US20210174104A1 (en) | Finger vein comparison method, computer equipment, and storage medium | |
| CN114511589A (en) | Human body tracking method and system | |
| CN108197608A (en) | Face identification method, device, robot and storage medium | |
| CN115205586A (en) | Knowledge distillation-based multi-self-supervision task fusion method and device and storage medium | |
| CN115830399A (en) | Classification model training method, apparatus, device, storage medium, and program product | |
| CN115761834A (en) | Multi-task mixed model for face recognition and face recognition method | |
| CN112613425B (en) | Target identification system for small sample underwater image | |
| CN105345822A (en) | Intelligent robot control method and device | |
| CN111144374B (en) | Facial expression recognition method and device, storage medium and electronic equipment | |
| Bian et al. | Real-time object detection for smart connected worker in 3d printing | |
| CN108805121B (en) | License plate detection and positioning method, device, equipment and computer readable medium | |
| CN115097974A (en) | Intelligent auxiliary consultation system and method for BIM (building information modeling) | |
| CN115480963A (en) | Function test method and system for automatic driving hardware platform | |
| CN107340962A (en) | Input method, device and virtual reality device based on virtual reality device | |
| Madhusudhana et al. | Design and implementation of access monitoring system for banks using OpenCV and Django | |
| CN111660294B (en) | Augmented reality control system of hydraulic heavy-duty mechanical arm | |
| CN112686156A (en) | Emotion monitoring method and device, computer equipment and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |