CN113012217A - Image processing-based construction method of convolution neural network for welding seam positioning - Google Patents

Image processing-based construction method of convolution neural network for welding seam positioning Download PDF

Info

Publication number
CN113012217A
CN113012217A CN201911332505.7A CN201911332505A CN113012217A CN 113012217 A CN113012217 A CN 113012217A CN 201911332505 A CN201911332505 A CN 201911332505A CN 113012217 A CN113012217 A CN 113012217A
Authority
CN
China
Prior art keywords
layer
welding seam
image
neural network
convolution
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
Application number
CN201911332505.7A
Other languages
Chinese (zh)
Inventor
不公告发明人
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Individual
Original Assignee
Individual
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Individual filed Critical Individual
Priority to CN201911332505.7A priority Critical patent/CN113012217A/en
Publication of CN113012217A publication Critical patent/CN113012217A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0014—Image feed-back for automatic industrial control, e.g. robot with camera
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30152—Solder
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/06—Recognition of objects for industrial automation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Robotics (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a method for constructing a convolution neural network for welding seam positioning based on image processing, which comprises the following steps: firstly, acquiring a convolutional neural network training sample for welding seam positioning based on image processing; step two, establishing a convolutional neural network for positioning the welding seam based on image processing, which comprises an input layer, a convolutional layer C1, a pooling layer S2, a convolutional layer C3, a pooling layer S4, a convolutional layer C5, a pooling layer S6, a fully-connected layer F7, a fully-connected layer F8 and an output layer; and step three, training the welding seam positioning based on image processing by using a convolution neural network. The method has simple steps and convenient implementation, constructs the convolution neural network for welding seam positioning based on image processing, can realize the rapid detection of welding seams and non-welding seams, can realize the automatic detection and positioning of the welding seams, has strong practicability and is convenient for popularization and use.

Description

Image processing-based construction method of convolution neural network for welding seam positioning
Technical Field
The invention belongs to the technical field of automatic welding, and particularly relates to a construction method of a convolution neural network for welding seam positioning based on image processing.
Background
In the production flow of welding, the accuracy of weld positioning always restricts the development of automatic welding, and at present, the automatic identification of the positioned weld by using a visual method has extensive research at home and abroad: one type is a welding seam image obtained according to X-ray photography, and a support vector machine and a fuzzy neural network are utilized to identify welding seam defects; the other type is that a welding seam image is obtained according to a high-speed camera CCD, and a welding seam is positioned by utilizing single stripe laser, an image matching technology, a structured light visual three-point, a structured light and uniform light multi-feature technology; and the ultrasonic sensing technology is also used for three-dimensional positioning. In the case of performing weld seam positioning based on image processing, it is conceivable to classify a weld seam and a non-weld seam using a convolutional neural network, but a convolutional neural network model for weld seam positioning based on image processing is still lacking in the prior art.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a method for constructing a convolutional neural network for welding seam positioning based on image processing, aiming at the defects in the prior art, the method has simple steps and convenient implementation, constructs the convolutional neural network for welding seam positioning based on image processing, can realize the rapid detection of welding seams and non-welding seams, can realize the automatic detection and positioning of welding seams, and has strong practicability and convenient popularization and use.
In order to solve the technical problems, the invention adopts the technical scheme that: a method for constructing a convolution neural network for welding seam positioning based on image processing is characterized by comprising the following steps:
the method comprises the following steps of firstly, obtaining a convolutional neural network training sample for welding seam positioning based on image processing, and specifically comprises the following steps:
step 101, periodically shooting an image of a weld joint to be positioned by using a camera and transmitting the shot image of the weld joint to be positioned to an image processor;
102, carrying out blocking processing on a welding seam image to be positioned by an image processor;
103, performing dimensionality reduction on the to-be-positioned welding seam image processed in the step 102 by an image processor to obtain a plurality of to-be-positioned welding seam images subjected to dimensionality reduction;
step 104, taking the multiple weld joint images to be positioned after the dimension reduction processing as training samples;
step two, establishing a convolutional neural network for positioning the welding seam based on image processing, which comprises an input layer, a convolutional layer C1, a pooling layer S2, a convolutional layer C3, a pooling layer S4, a convolutional layer C5, a pooling layer S6, a fully-connected layer F7, a fully-connected layer F8 and an output layer;
an input layer for unifying the size of the input color image data into 224 × 3;
convolutional layer C1: performing convolution on input color image data, extracting features, performing convolution kernel extraction on the input color image data by 11 × 11, performing step length extraction on the input color image data by 4, and measuring the number of feature mapping maps by 12 to obtain feature maps of 55 × 12;
pooling layer S2: performing dimensionality reduction on the feature map after convolution of the convolution layer C1, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 2 in step size to obtain a 27 × 12 feature map;
convolutional layer C3: convolving the characteristic diagram data subjected to dimensionality reduction of the pooling layer S2, extracting characteristics, taking 5 × 5 as convolution kernel, taking 2 as step length, measuring 48 as characteristic mapping diagram number, and obtaining a characteristic diagram of 27 × 48;
pooling layer S4: performing dimensionality reduction on the feature map after convolution of the convolution layer C3, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 2 in step size to obtain a 13 × 48 feature map;
convolutional layer C5: convolving the characteristic diagram data subjected to dimensionality reduction of the pooling layer S4, extracting characteristics, taking 3 × 3 as convolution kernel, taking 1 as step length, measuring 96 as the number of characteristic mapping diagrams, and obtaining 13 × 96 characteristic diagrams;
pooling layer S6: performing dimensionality reduction on the feature map after convolution of the convolution layer C5, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 1 in step size to obtain a feature map of 6 × 96;
full connection layer F7: 1024 neurons are set, all the characteristics output by the pooling layer S6 are connected, and a ReLU function is set as an activation function;
full connection layer F8: 512 neurons are arranged and are fully connected with the neurons of a full connection layer F7, and a ReLU function is set as an activation function;
an output layer: 2 neurons are arranged and are in full connection with the neurons of the full connection layer F8, and output results are classified in a second mode; the result of the second classification is "weld" or "non-weld";
step three, training the welding seam positioning convolution neural network based on image processing: and (4) training the convolution neural network for positioning the welding seam based on the image processing constructed in the step two by adopting the training sample in the step one to obtain the trained convolution neural network for positioning the welding seam based on the image processing.
The method for constructing the convolution neural network for welding seam positioning based on image processing is characterized by comprising the following steps of: the image processor is a computer.
The method for constructing the convolution neural network for welding seam positioning based on image processing is characterized by comprising the following steps of: in step 102, the specific process of the image processor for blocking the weld joint image to be positioned is as follows: an image processor receives the weld joint image to be positioned and divides the weld joint image to be positioned into M multiplied by N weld joint subimages Y to be positioned according to the line interval width d and the column interval width H1、Y2、…、YM×NEach to-be-positioned welding seam subimage consists of M multiplied by N pixels, wherein M is the number of lines of the to-be-positioned welding seam subimage, N is the number of columns of the to-be-positioned welding seam subimage, d, H, M, N, M and N are natural numbers, and the units of d and H are pixels.
The method for constructing the convolution neural network for welding seam positioning based on image processing is characterized by comprising the following steps of: in step 103, the specific process of performing the dimensionality reduction processing on the to-be-positioned weld image processed in step 102 by the image processor is as follows: the image processor calls a dimensionality reduction matrix W which is trained by a principal component analysis method in advance and according to a formula Y'f＝WYfPerforming dimensionality reduction treatment on the MxN welding seam subimages to be positioned, and performing Y-dimensional reduction treatment on the MxN welding seam subimages to be positioned1、Y2、…、YM×NConverting the feature vectors into the feature vectors Y 'of the M multiplied by N welding seam sub-images to be positioned after dimensionality reduction processing'1、Y′2、…、Y′M×NWherein Y isfIs the f-th sub-image of the weld joint to be positioned
Figure BDA0002330049100000041
Y′fFor the f-th position to be positioned after dimension reduction processingThe value of the characteristic vector f of the sub-image of the welding seam is a natural number of 1-MXN.
Compared with the prior art, the invention has the following advantages: the method has simple steps and convenient implementation, constructs the convolution neural network for welding seam positioning based on image processing, can realize the rapid detection of the welding seam and the non-welding seam, can realize the automatic detection and positioning of the welding seam, has strong practicability and is convenient for popularization and use.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
FIG. 1 is a block diagram of the process flow of the present invention.
Detailed Description
As shown in fig. 1, the method for constructing a convolutional neural network for positioning a welding seam based on image processing of the present invention comprises the following steps:
the method comprises the following steps of firstly, obtaining a convolutional neural network training sample for welding seam positioning based on image processing, and specifically comprises the following steps:
step 101, periodically shooting an image of a weld joint to be positioned by using a camera and transmitting the shot image of the weld joint to be positioned to an image processor;
102, carrying out blocking processing on a welding seam image to be positioned by an image processor;
103, performing dimensionality reduction on the to-be-positioned welding seam image processed in the step 102 by an image processor to obtain a plurality of to-be-positioned welding seam images subjected to dimensionality reduction;
step 104, taking the multiple weld joint images to be positioned after the dimension reduction processing as training samples;
step two, establishing a convolutional neural network for positioning the welding seam based on image processing, which comprises an input layer, a convolutional layer C1, a pooling layer S2, a convolutional layer C3, a pooling layer S4, a convolutional layer C5, a pooling layer S6, a fully-connected layer F7, a fully-connected layer F8 and an output layer;
an input layer for unifying the size of the input color image data into 224 × 3;
convolutional layer C1: performing convolution on input color image data, extracting features, performing convolution kernel extraction on the input color image data by 11 × 11, performing step length extraction on the input color image data by 4, and measuring the number of feature mapping maps by 12 to obtain feature maps of 55 × 12;
pooling layer S2: performing dimensionality reduction on the feature map after convolution of the convolution layer C1, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 2 in step size to obtain a 27 × 12 feature map;
convolutional layer C3: convolving the characteristic diagram data subjected to dimensionality reduction of the pooling layer S2, extracting characteristics, taking 5 × 5 as convolution kernel, taking 2 as step length, measuring 48 as characteristic mapping diagram number, and obtaining a characteristic diagram of 27 × 48;
pooling layer S4: performing dimensionality reduction on the feature map after convolution of the convolution layer C3, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 2 in step size to obtain a 13 × 48 feature map;
convolutional layer C5: convolving the characteristic diagram data subjected to dimensionality reduction of the pooling layer S4, extracting characteristics, taking 3 × 3 as convolution kernel, taking 1 as step length, measuring 96 as the number of characteristic mapping diagrams, and obtaining 13 × 96 characteristic diagrams;
pooling layer S6: performing dimensionality reduction on the feature map after convolution of the convolution layer C5, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 1 in step size to obtain a feature map of 6 × 96;
full connection layer F7: 1024 neurons are set, all the characteristics output by the pooling layer S6 are connected, and a ReLU function is set as an activation function;
full connection layer F8: 512 neurons are arranged and are fully connected with the neurons of a full connection layer F7, and a ReLU function is set as an activation function;
an output layer: 2 neurons are arranged and are in full connection with the neurons of the full connection layer F8, and output results are classified in a second mode; the result of the second classification is "weld" or "non-weld";
step three, training the welding seam positioning convolution neural network based on image processing: and (4) training the convolution neural network for positioning the welding seam based on the image processing constructed in the step two by adopting the training sample in the step one to obtain the trained convolution neural network for positioning the welding seam based on the image processing.
In this embodiment, the image processor is a computer.
In this embodiment, the specific process of the image processor performing blocking processing on the weld image to be positioned in step 102 is as follows: an image processor receives the weld joint image to be positioned and divides the weld joint image to be positioned into M multiplied by N weld joint subimages Y to be positioned according to the line interval width d and the column interval width H1、Y2、…、YM×NEach to-be-positioned welding seam subimage consists of M multiplied by N pixels, wherein M is the number of lines of the to-be-positioned welding seam subimage, N is the number of columns of the to-be-positioned welding seam subimage, d, H, M, N, M and N are natural numbers, and the units of d and H are pixels.
In this embodiment, the specific process of performing, by the image processor in step 103, the dimension reduction processing on the to-be-positioned weld image processed in step 102 is as follows: the image processor calls a dimensionality reduction matrix W which is trained by a principal component analysis method in advance and according to a formula Y'f＝WYfPerforming dimensionality reduction treatment on the MxN welding seam subimages to be positioned, and performing Y-dimensional reduction treatment on the MxN welding seam subimages to be positioned1、Y2、…、YM×NConverting the feature vectors into the feature vectors Y 'of the M multiplied by N welding seam sub-images to be positioned after dimensionality reduction processing'1、Y′2…、Y′M×NWherein Y isfIs the f-th sub-image of the weld joint to be positioned
Figure BDA0002330049100000061
Y′fAnd f is a natural number of 1-MXN for the characteristic vector of the f-th sub-image of the weld joint to be positioned after dimension reduction treatment.
In conclusion, the method provided by the invention has the advantages of simple steps and convenience in implementation, the convolutional neural network for welding seam positioning based on image processing is constructed, the rapid detection of the welding seam and the non-welding seam can be realized, the automatic detection and positioning of the welding seam can be realized, the practicability is high, and the popularization and the use are convenient.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing descriptions of specific exemplary embodiments of the present invention have been presented for purposes of illustration and description. It is not intended to limit the invention to the precise form disclosed, and obviously many modifications and variations are possible in light of the above teaching. The exemplary embodiments were chosen and described in order to explain certain principles of the invention and its practical application to enable one skilled in the art to make and use various exemplary embodiments of the invention and various alternatives and modifications as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the claims and their equivalents.

Claims (4)

1. A method for constructing a convolution neural network for welding seam positioning based on image processing is characterized by comprising the following steps:
the method comprises the following steps of firstly, obtaining a convolutional neural network training sample for welding seam positioning based on image processing, and specifically comprises the following steps:
step 101, periodically shooting an image of a weld joint to be positioned by using a camera and transmitting the shot image of the weld joint to be positioned to an image processor;
102, carrying out blocking processing on a welding seam image to be positioned by an image processor;
103, performing dimensionality reduction on the to-be-positioned welding seam image processed in the step 102 by an image processor to obtain a plurality of to-be-positioned welding seam images subjected to dimensionality reduction;
step 104, taking the multiple weld joint images to be positioned after the dimension reduction processing as training samples;
step two, establishing a convolutional neural network for positioning the welding seam based on image processing, which comprises an input layer, a convolutional layer C1, a pooling layer S2, a convolutional layer C3, a pooling layer S4, a convolutional layer C5, a pooling layer S6, a fully-connected layer F7, a fully-connected layer F8 and an output layer;
an input layer for unifying the size of the input color image data into 224 × 3;
convolutional layer C1: performing convolution on input color image data, extracting features, performing convolution kernel extraction on the input color image data by 11 × 11, performing step length extraction on the input color image data by 4, and measuring the number of feature mapping maps by 12 to obtain feature maps of 55 × 12;
pooling layer S2: performing dimensionality reduction on the feature map after convolution of the convolution layer C1, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 2 in step size to obtain a 27 × 12 feature map;
convolutional layer C3: convolving the characteristic diagram data subjected to dimensionality reduction of the pooling layer S2, extracting characteristics, taking 5 × 5 as convolution kernel, taking 2 as step length, measuring 48 as characteristic mapping diagram number, and obtaining a characteristic diagram of 27 × 48;
pooling layer S4: performing dimensionality reduction on the feature map after convolution of the convolution layer C3, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 2 in step size to obtain a 13 × 48 feature map;
convolutional layer C5: convolving the characteristic diagram data subjected to dimensionality reduction of the pooling layer S4, extracting characteristics, taking 3 × 3 as convolution kernel, taking 1 as step length, measuring 96 as the number of characteristic mapping diagrams, and obtaining 13 × 96 characteristic diagrams;
pooling layer S6: performing dimensionality reduction on the feature map after convolution of the convolution layer C5, performing feature extraction by adopting a self-adaptive pooling method, taking 3 × 3 in a rectangular area and 1 in step size to obtain a feature map of 6 × 96;
full connection layer F7: 1024 neurons are set, all the characteristics output by the pooling layer S6 are connected, and a ReLU function is set as an activation function;
full connection layer F8: 512 neurons are arranged and are fully connected with the neurons of a full connection layer F7, and a ReLU function is set as an activation function;
an output layer: 2 neurons are arranged and are in full connection with the neurons of the full connection layer F8, and output results are classified in a second mode; the result of the second classification is "weld" or "non-weld";
step three, training the welding seam positioning convolution neural network based on image processing: and (4) training the convolution neural network for positioning the welding seam based on the image processing constructed in the step two by adopting the training sample in the step one to obtain the trained convolution neural network for positioning the welding seam based on the image processing.
2. The method of constructing a convolutional neural network for image processing-based weld seam positioning according to claim 1, wherein: the image processor is a computer.
3. The method of constructing a convolutional neural network for image processing-based weld seam positioning according to claim 1, wherein: in step 102, the specific process of the image processor for blocking the weld joint image to be positioned is as follows: an image processor receives the weld joint image to be positioned and divides the weld joint image to be positioned into M multiplied by N weld joint subimages Y to be positioned according to the line interval width d and the column interval width H1、Y2、…、YM×NEach to-be-positioned welding seam subimage consists of M multiplied by N pixels, wherein M is the number of lines of the to-be-positioned welding seam subimage, N is the number of columns of the to-be-positioned welding seam subimage, d, H, M, N, M and N are natural numbers, and the units of d and H are pixels.
4. The method of constructing a convolutional neural network for image processing-based weld seam positioning according to claim 1, wherein: in step 103, the specific process of performing the dimensionality reduction processing on the to-be-positioned weld image processed in step 102 by the image processor is as follows: the image processor calls a dimensionality reduction matrix W which is trained by a principal component analysis method in advance and according to a formula Y'f＝WYfPerforming dimensionality reduction treatment on the MxN welding seam subimages to be positioned, and performing Y-dimensional reduction treatment on the MxN welding seam subimages to be positioned1、Y2、…、YM×NConverting the feature vectors into M multiplied by N subimage feature vectors Y of the weld joint to be positioned after dimension reduction treatment1′、Y2′、…、YM×N', wherein, YfIs the f-th sub-image of the weld joint to be positioned
Figure FDA0002330049090000031
Y′fAnd f is a natural number of 1-MXN for the characteristic vector of the f-th sub-image of the weld joint to be positioned after dimension reduction treatment.
CN201911332505.7A 2019-12-22 2019-12-22 Image processing-based construction method of convolution neural network for welding seam positioning Pending CN113012217A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911332505.7A CN113012217A (en) 2019-12-22 2019-12-22 Image processing-based construction method of convolution neural network for welding seam positioning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911332505.7A CN113012217A (en) 2019-12-22 2019-12-22 Image processing-based construction method of convolution neural network for welding seam positioning

Publications (1)

Publication Number Publication Date
CN113012217A true CN113012217A (en) 2021-06-22

Family

ID=76382808

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911332505.7A Pending CN113012217A (en) 2019-12-22 2019-12-22 Image processing-based construction method of convolution neural network for welding seam positioning

Country Status (1)

Country Link
CN (1) CN113012217A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102021117714A1 (en) 2021-07-08 2023-01-12 Endress+Hauser SE+Co. KG Automatic seam detection for a welding process

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102021117714A1 (en) 2021-07-08 2023-01-12 Endress+Hauser SE+Co. KG Automatic seam detection for a welding process

Similar Documents

Publication Publication Date Title
Liu et al. YOLOv5-Tassel: Detecting tassels in RGB UAV imagery with improved YOLOv5 based on transfer learning
CA3121440C (en) Assembly body change detection method, device and medium based on attention mechanism
CN112528878B (en) Method and device for detecting lane line, terminal equipment and readable storage medium
Choi et al. Thermal image enhancement using convolutional neural network
Fujita et al. Basic investigation on a robust and practical plant diagnostic system
CN110619638A (en) Multi-mode fusion significance detection method based on convolution block attention module
Wu et al. VH-HFCN based parking slot and lane markings segmentation on panoramic surround view
CN104299006A (en) Vehicle license plate recognition method based on deep neural network
CN115880558B (en) Farming behavior detection method and device, electronic equipment and storage medium
CN111709980A (en) Multi-scale image registration method and device based on deep learning
CN114581456B (en) Multi-image segmentation model construction method, image detection method and device
US20200202542A1 (en) Systems and methods for determining depth information in two-dimensional images
CN112819748B (en) Training method and device for strip steel surface defect recognition model
CN104850857B (en) Across the video camera pedestrian target matching process of view-based access control model spatial saliency constraint
CN109685774A (en) Varistor open defect detection method based on depth convolutional neural networks
CN111914596B (en) Lane line detection method, device, system and storage medium
CN113012217A (en) Image processing-based construction method of convolution neural network for welding seam positioning
CN112070181B (en) Image stream-based cooperative detection method and device and storage medium
CN113065575A (en) Image processing method and related device
CN106683044B (en) Image splicing method and device of multi-channel optical detection system
CN111881914A (en) License plate character segmentation method and system based on self-learning threshold
CN116228708A (en) Industrial defect detection method and system based on visual cognition calculation
CN113012092A (en) Automatic welding positioning method based on machine vision
CN111079585A (en) Image enhancement and pseudo-twin convolution neural network combined pedestrian re-identification method based on deep learning
CN115860139A (en) Deep learning-based multi-scale ship target detection method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20210622