CN115660966A - Machine vision intelligent auxiliary identification method and system - Google Patents
Machine vision intelligent auxiliary identification method and system Download PDFInfo
- Publication number
- CN115660966A CN115660966A CN202211088455.4A CN202211088455A CN115660966A CN 115660966 A CN115660966 A CN 115660966A CN 202211088455 A CN202211088455 A CN 202211088455A CN 115660966 A CN115660966 A CN 115660966A
- Authority
- CN
- China
- Prior art keywords
- iron
- scrap
- image
- residual
- clamp
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a machine vision intelligent auxiliary identification method and a system, comprising the following steps: marking the residual iron scrap picture data to form a target frame as a data set of a residual iron scrap detection model; training a residual scrap iron detection model by using the data set to obtain a trained residual scrap iron detection model; collecting an iron scrap image on a clamp and carrying out image preprocessing on the iron scrap image; inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and judging whether scrap iron residues exist. The detection efficiency and the accuracy of the residual scrap iron can be improved, and the automation of the detection of the residual scrap iron is increased.
Description
Technical Field
The invention belongs to the technical field of detection of clamp residual scrap iron, and particularly relates to a machine vision intelligent auxiliary identification method and system.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The bogie comprises a wheel, wherein the wheel is manufactured, shaft end machining is an important ring in the wheel manufacturing and machining process, scrap iron is easy to remain in the clamp after the shaft end machining is finished, the scrap iron can still remain after being washed, and the residual scrap iron in the clamp can influence the positioning of a workpiece added by a subsequent clamp in the using process, so that the machined part machined by the clamp has machining errors.
At present, the treatment of the residual scrap iron in the wheel set end face milling clamp is generally to detect the residual scrap iron through manual work, and the mode not only is influenced by subjective factors of workers, but also greatly reduces the working efficiency.
In recent years, a plurality of object detection algorithms have been proposed one after another, however, the conventional object detection algorithm is less applied to the recognition of the residual scrap iron because the similarity of the property of the scrap iron and the background of the jig is high.
Therefore, the existing recognition accuracy of scrap iron in the clamp is not high, and the derived error of the residual scrap iron of the clamp for milling the end face of the wheel set on the subsequent machining operation is caused.
Disclosure of Invention
In order to solve the problems, the invention provides a machine vision intelligent auxiliary identification method which can improve the detection efficiency and accuracy of the residual scrap iron and increase the automation of the detection of the residual scrap iron.
According to some embodiments, the invention adopts the following technical scheme:
in a first aspect, a machine vision intelligent auxiliary identification method is disclosed, which comprises:
marking the residual iron scrap picture data to form a target frame as a data set of a residual iron scrap detection model;
training a residual scrap iron detection model by using the data set to obtain a trained residual scrap iron detection model;
collecting iron scrap images on a clamp and carrying out image preprocessing on the iron scrap images;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and judging whether scrap iron residue exists.
According to the technical scheme, before the scrap iron image on the clamp is collected, a workpiece to be cut is placed on the clamp chassis, the workpiece is fixed at a designated position by the chuck, the workpiece is cut by the control tool, the workpiece is transported away after the cutting of the workpiece is completed, and the scrap iron on the clamp is subjected to image collection through the camera.
According to the further technical scheme, whether scrap iron residues exist or not is judged, and if the scrap iron residues do not exist or meet the requirements, an alarm is not generated;
if the scrap iron image is residual or not meeting the requirement, cleaning the scrap iron image, and acquiring the scrap iron image on the clamp again and carrying out image preprocessing on the scrap iron image;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and detecting again.
Further technical scheme carries out image preprocessing to iron fillings image, includes:
cutting and scaling the residual scrap iron image;
and carrying out filtering and denoising treatment on the residual scrap iron image by adopting an image filtering algorithm, and carrying out image enhancement on the image.
According to the further technical scheme, the residual scrap iron detection model automatically identifies scrap iron in the image through a YOLO algorithm and frames the position of the scrap iron with a target frame.
The second aspect discloses supplementary identification system of iron fillings machine vision is remained to anchor clamps, includes:
a data set building module configured to: marking the residual iron scrap picture data to form a target frame as a data set of a residual iron scrap detection model;
a model training module configured to: training a residual iron scrap detection model by using the data set to obtain a trained residual iron scrap detection model;
a detection module configured to: collecting an iron scrap image on a clamp and carrying out image preprocessing on the iron scrap image;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and judging whether scrap iron residues exist.
The third aspect discloses supplementary recognition device of iron fillings machine vision is remained to anchor clamps, includes:
the image acquisition equipment comprises a plurality of visual sensors, the visual sensors are arranged right above the grippers on the clamp, and the image information shot by the visual sensors is transmitted to the processor for preprocessing;
and the processor stores the trained residual iron scrap detection model, identifies the preprocessed image by using the trained residual iron scrap detection model, and judges whether iron scrap residues exist or not.
According to the technical scheme, the vision sensor is provided with a lens protection cover, and the lens protection cover is provided with an automatic heating and purging device for avoiding cutting and cutting liquid splashing in the cutting process.
According to the technical scheme, the device comprises a shell, a telescopic rod is arranged on the shell, a sliding groove is formed in the shell, the cutter can translate an XY coordinate plane through the sliding groove, and the telescopic rod can move up and down an OZ coordinate axis.
According to the technical scheme, the vision sensor is fixed on the shell through a telescopic rod, the OZ coordinate axis moves up and down through the telescopic rod and is controlled by the processor, when algorithm detection in the processor is finished, the telescopic rod is controlled to rebound automatically, and the vision sensor is closed.
Compared with the prior art, the invention has the beneficial effects that:
the invention discloses a method for detecting residual scrap iron of a clamp, which comprises the following steps of firstly, marking a required label in a residual scrap iron picture sample to form a target frame as a data set of a residual scrap iron detection model; secondly, data enhancement is carried out on the data of the residual scrap iron data set; thirdly, placing the test set into a trained model for testing, so that the model precision reaches more than 95%; and finally, installing a visual sensor provided with a specific protective cover right above the clamp corresponding to the claw on the clamp, and putting the shot image into a trained model to judge whether residual scrap iron exists. The problem of the derived error that wheel set terminal surface milling process anchor clamps remain iron fillings and produce follow-up processing operation is solved, improved the detection efficiency and the degree of accuracy of remaining iron fillings, increase the automation that remains iron fillings detected.
Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a schematic view of the overall structure of the present invention;
FIG. 2 is a surface view of the clamp of the present invention;
FIG. 3 is a diagram of the scrap iron remaining on the jig of the present invention;
FIG. 4 is a diagram showing the effect of the present invention after the YOLO-V5s recognition;
FIG. 5 is a flow chart of the detection method of the present invention.
In the figure, 1, a cartridge; 2. a housing; 3. a vision sensor; 4. a cutter; 5. a hydraulic lever; 6. a clamp chassis; 7. a display; 8. and (4) clamping.
The specific implementation mode is as follows:
the invention is further described with reference to the following figures and examples.
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The first embodiment is as follows:
in this embodiment, a face milling fixture is used for illustration, but this does not mean that the method provided by the present invention can only be applied to the iron scrap detection of the face milling fixture. The method can also be applied to image detection in which other related objects to be detected are close to scenes where the objects are located and difficult to identify according to different identification objects.
Target detection algorithms based on the current mainstream can be divided into two categories: the first type mainly adopts a two-stage form, namely, distinguishing the front background and the back background, namely, positioning the target, and then specifically classifying the target, wherein the commonly used detection algorithms mainly comprise an SSD series, an R-CNN series and the like; the other type adopts a one-stage mode, namely, the purposes of positioning and specific classification can be achieved through one-time classification, and a common detection algorithm is a YOLO series algorithm. Based on the high similarity between the residual scrap iron and the background characters of the clamp, the method adopts a YOLO-V5s series algorithm to carry out target detection, and has the advantages of high detection speed and good small target identification effect.
As shown in figures 1 and 2, the invention provides a jig residual scrap iron detection method based on a YOLO algorithm, which is used for detecting residual scrap iron left on a jig after a tool controlled by a PLC is cut, wherein figure 1 is a schematic overall structure diagram, the positions of a jig 6 and a vision sensor 3 in the equipment can be clearly seen from the schematic overall structure diagram, a mold box 1 is installed on one side of a shell 2, the vision sensor 3 is installed above the shell 2, the vision sensor 3 is installed at the upper end of the shell 2 through a hydraulic rod 5, a display 7 is connected to a PLC system, an image collected by the vision sensor 3 is transmitted to the PLC system for processing, the tool 4 is also arranged above the shell 2, and a jig 8 is arranged on a jig chassis 6.
Fig. 2 is a view of the fixture surface from which the fixture chassis 6 surface and the three-jaw structure can be clearly seen, and the vision sensor 3 captures an image of the fixture surface thereof.
As shown in fig. 5, to achieve the above object, the modeling in the present invention includes the following steps:
step S1: firstly, taking the residual iron scrap picture data with 800 fine labels as a data set, wherein fig. 3 is an example of a certain image in the data set, and then labeling the required label by using a labeling tool to form a target frame as the data set of the residual iron scrap detection model.
Step S2: preprocessing data of a residual scrap iron data set, establishing sample data, dividing the sample data into a training set and a test set according to a specific proportion (considering data balance), putting the training set and the test set data into a YOLO-V5s network, setting a hyper-parameter, and then starting to train the network;
the network structure is referred to GoogleLeNet. Consists of a convolutional layer and a full link layer.
(1) Inputting an image, firstly dividing the image into a SxS grid;
(2) For each mesh, B bounding boxes are predicted (including the confidence that each bounding box is a target and the probability that each bounding box region is in multiple categories).
(3) And predicting SxSxB target windows according to the previous step, removing the target windows with low possibility according to a threshold value, and finally removing the redundant windows by the NMS.
And step S3: verifying the performance of the model through the test set;
and step S4: adjusting and optimizing to enable the accuracy of the model to reach more than 95%, and then storing the model;
in this embodiment, the image preprocessing is performed on the residual iron scrap image, including:
and (4) clipping (640 x 640) and scaling transformation are carried out on the residual scrap iron image. Then, carrying out filtering and denoising treatment on the residual iron scrap image by adopting an image filtering algorithm, wherein the image filtering algorithm comprises the following steps: mean filtering, median filtering and Gaussian filtering, and performing Retines image enhancement technology processing on the image.
Specifically, the effect diagram of the original image, like fig. 3, after the image preprocessing is shown in fig. 4.
Wherein setting the hyper-parameters comprises selecting the hyper-parameters of the model, comprising:
batch_size＝32，epochs＝200，adam＝Adam()optimizer。
after the model is built, the modeling process can be seen in the display 7, and the field identification is performed by the following steps:
step S1: when the workpiece is placed on the clamp chassis 6, the three-jaw self-centering chuck on the clamp clamps the workpiece through the command of a PLC system, centers the workpiece at the center of the chuck, and then controls the cutter 4 to cut.
Step S2: when the work piece cutting is accomplished, the M instruction through the milling machine will be through the cooperation of tool changing process, lathe automatic operation, withdraws cutter 4 automatically, and the work piece is carried away, and anchor clamps 8 resets, and the chuck opens, and visual sensor 3 shoots this moment, carries out image acquisition to iron fillings on anchor clamps 8 through the camera to carry out image preprocessing to the iron fillings image, the image after the processing will show in display 7.
And step S3: and putting the preprocessed scrap iron image into a pre-trained model, and judging whether scrap iron residues exist.
And step S4: if the scrap iron is free of residues or meets the requirements, no alarm is generated, and the generated reminder can be circulated according to the original flow. If the system gives an alarm when the residue or the residue does not meet the requirement, the position shown as the frame in the figure 4 is remained, the process is started again after the position is cleaned, and the system can detect again.
Step S5: and after the judgment is finished, the surface of the clamp is free of scrap iron, a new workpiece is placed on the clamp base, and the steps are repeated.
Specifically, for step S1, the lens of the vision sensor 3 used is mounted with a protective cover with an automatic heating and purging device, and the vision sensor 3 is mounted right above the gripper corresponding to the gripper on the jig 8, i.e. three vision sensors 3 are taken one by one, for a total of three, and the taken image information is pre-processed in the processor and displayed on the display 7.
The automatic heating and blowing device blows and blows at the outer wall of the protective cover and heats the inner wall; the work piece is worked when the cutting, avoids cutting fluid to splash and influences the shooting.
Specifically, in step S2, the three images captured by the vision sensor are put into the previously built model, and the model for recognizing the object in the image automatically recognizes the iron filings in the image by using the YOLO algorithm and frames the position with the target frame.
And inputting the three pictures as models, and recognizing iron filings in the pictures to frame by using the models for recognizing the objects in the pictures.
In addition, the cutter 4 is fixed on the shell 2 through a telescopic rod, a sliding chute is arranged in the shell, the cutter 4 can translate on an XY coordinate plane through the sliding chute, and the telescopic rod can move up and down on an OZ coordinate axis and is controlled by a PLC system.
And the vision sensor 3 is fixed on the shell 2 through a telescopic rod, the OZ coordinate axis is moved up and down through the telescopic rod and is controlled by the PLC system, when the detection of the YOLO algorithm of the model in the PLC system is finished, the PLC system controls the telescopic rod to automatically rebound, and the vision sensor is closed.
The lathe fixture is controlled by a PLC system, is fixed on a chassis and is a three-jaw self-centering chuck, and the three-jaw self-centering chuck is linked and has the characteristics of simplicity in clamping, large clamping range and automatic centering. The image acquisition device is controlled by a system based on a YOLO algorithm, a lens of a visual sensor of the image acquisition device is provided with a protective cover, an automatic heating and purging device is arranged, and the iron scrap condition on the clamp can be displayed in a display through the system.
The method for detecting the iron scrap residual on the basis of the YOLO algorithm and the image acquisition equipment can solve the problem that the iron scrap is easy to remain in the existing clamp, and the detection method has the following advantages in the process of detecting the iron scrap residual on the clamp:
the YOLO-V5s algorithm is used as a small target detection algorithm, the detection capability is strong, the accuracy of the model is high, the process of iron scrap detection is greatly improved by processing images shot by a visual sensor, the workload of workers is reduced, the production and manufacturing efficiency is improved, the period of the working process is accelerated, and the production and manufacturing are favorably promoted to be unmanned;
the image acquisition equipment is added in the working process, and a main part of the image acquisition equipment is provided with a lens protective cover (provided with an automatic heating and purging device), so that the cutting and cutting liquid splashing conditions in the cutting process are effectively prevented.
Through the cooperation between the visual sensor and the information acquisition module, the detection result image can be displayed in real time in the display.
Three images shot by the vision sensor are added to the original database, and the accumulated images are labeled and modeled again, so that the data volume is increased.
According to the method, the accuracy is introduced to be used as the evaluation of the counting result, YOLO-V5s is used as the residual scrap iron detection algorithm, and the experimental result shows that the result predicted by the method is high in precision, and the small target detection capability is obviously enhanced. The method can accurately display the scrap iron condition in one picture and give a warning in time.
In the invention, the accuracy is used as an index for evaluating the quality of a model, more than 95 are divided into 100, and the accuracy reaches more than 95%. Accuracy = identify correct number/total number.
Example two:
based on the method of the first embodiment, the first embodiment discloses a machine vision intelligent auxiliary recognition system for clamp residual scrap iron, which comprises:
a data set building module configured to: marking the residual iron scrap picture data to form a target frame as a data set of a residual iron scrap detection model;
a model training module configured to: training a residual scrap iron detection model by using the data set to obtain a trained residual scrap iron detection model;
a detection module configured to: collecting an iron scrap image on a clamp and carrying out image preprocessing on the iron scrap image;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and judging whether scrap iron residues exist.
In this embodiment, the specific implementation method of the module may refer to the description of the specific method in embodiment one.
Example three:
based on the method of the first embodiment, the first embodiment discloses a machine vision intelligent auxiliary recognition device for clamp residual scrap iron, which comprises:
the image acquisition equipment comprises a plurality of visual sensors, the visual sensors are arranged right above the grippers on the clamp, and the image information shot by the visual sensors is transmitted to the processor for preprocessing;
and the PC end stores a trained residual iron scrap detection model, and identifies the preprocessed image by using the trained residual iron scrap detection model to judge whether iron scrap residues exist.
The vision sensor is provided with a lens protection cover, and the lens protection cover is provided with an automatic heating and purging device for avoiding cutting and cutting liquid splashing in the cutting process.
The cutting tool is fixed on the shell through a telescopic rod, a sliding groove is formed in the shell, the cutting tool can translate an XY coordinate plane through the sliding groove, and the OZ coordinate axis moves up and down through the telescopic rod.
The vision sensor is fixed on the shell through the telescopic rod, the OZ coordinate axis moves up and down through the telescopic rod and is controlled by the processor, when algorithm detection in the processor is finished, the telescopic rod is controlled to automatically rebound, and the vision sensor is closed.
Iron fillings detecting system is remained to anchor clamps in this embodiment, solves the problem that easily remains iron fillings on the anchor clamps in current lathe processing, and through the image of vision sensor transmission, the system can be accurate discerns the iron fillings condition of remaining on the anchor clamps, in time makes corresponding action, has not only improved the detection efficiency and the degree of accuracy of remaining iron fillings, also increases the automation that remains iron fillings and detect.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive changes in the technical solutions of the present invention.
Claims (10)
1. A machine vision intelligent auxiliary identification method is characterized by comprising the following steps:
marking the residual iron scrap picture data to form a target frame as a data set of a residual iron scrap detection model;
training a residual scrap iron detection model by using the data set to obtain a trained residual scrap iron detection model;
collecting an iron scrap image on a clamp and carrying out image preprocessing on the iron scrap image;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and judging whether scrap iron residues exist.
2. The machine vision intelligent auxiliary identification method as claimed in claim 1, wherein before the scrap iron image on the clamp is collected, a workpiece to be cut is placed on a clamp chassis, the workpiece is fixed at a designated position by a chuck, a cutter is controlled to cut the workpiece, after the workpiece is cut, the workpiece is transported away, and the scrap iron on the clamp is collected through a camera.
3. The machine vision intelligent auxiliary identification method as claimed in claim 1 or 2, wherein judging whether scrap iron residues exist, if scrap iron does not have residues or meets the requirements, no alarm is generated;
if the scrap iron image is residual or not meeting the requirement, cleaning the scrap iron image, and acquiring the scrap iron image on the clamp again and carrying out image preprocessing on the scrap iron image;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and detecting again.
4. The machine vision intelligent auxiliary identification method as claimed in claim 1 or 2, wherein the image preprocessing is performed on the scrap iron image, and comprises the following steps:
cutting and scaling the residual scrap iron image;
and carrying out filtering and denoising treatment on the residual scrap iron image by adopting an image filtering algorithm, and carrying out image enhancement on the image.
5. The machine-vision intelligent auxiliary identification method as claimed in any one of claims 1 to 4, wherein the residual iron scrap detection model automatically identifies the position of the iron scrap in the image by using a YOLO algorithm and frames the position by using a target frame.
6. The utility model provides an supplementary identification system of iron fillings machine vision intelligence is remained to anchor clamps, characterized by includes:
a data set building module configured to: marking the residual iron scrap picture data to form a target frame as a data set of a residual iron scrap detection model;
a model training module configured to: training a residual iron scrap detection model by using the data set to obtain a trained residual iron scrap detection model;
a detection module configured to: collecting an iron scrap image on a clamp and carrying out image preprocessing on the iron scrap image;
inputting the preprocessed scrap iron image into the trained residual scrap iron detection model, and judging whether scrap iron residues exist.
7. The utility model provides an supplementary recognition device of iron fillings machine vision intelligence is remained to anchor clamps, characterized by includes:
the image acquisition equipment comprises a plurality of visual sensors, the visual sensors are arranged right above the grippers on the clamp, and the image information shot by the visual sensors is transmitted to the processor for preprocessing;
and the processor stores the trained residual iron scrap detection model, identifies the preprocessed image by using the trained residual iron scrap detection model, and judges whether iron scrap residues exist or not.
8. The machine vision intelligent auxiliary recognition device for clamp residual iron filings according to claim 7, wherein the vision sensor is installed with a lens protection cover, and an automatic heating and purging device is installed on the lens protection cover for avoiding cutting and cutting liquid splashing during cutting.
9. The machine vision intelligent auxiliary recognition device for clamp residual iron scraps according to claim 7 or 8, which is characterized by further comprising a cutter, wherein the cutter is fixed on a shell through a telescopic rod, a sliding groove is formed in the shell, the cutter can perform translation of an OXY coordinate plane through the sliding groove, and the telescopic rod can perform up-and-down movement of an OZ coordinate axis.
10. The machine vision intelligent auxiliary recognition device for clamp residual iron scraps according to claim 9, wherein the vision sensor is fixed on the shell through a telescopic rod, the telescopic rod is used for moving an OZ coordinate axis up and down, the vision sensor is controlled by the processor, when algorithm detection in the processor is finished, the telescopic rod is controlled to automatically rebound, and the vision sensor is closed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211088455.4A CN115660966A (en) | 2022-09-07 | 2022-09-07 | Machine vision intelligent auxiliary identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211088455.4A CN115660966A (en) | 2022-09-07 | 2022-09-07 | Machine vision intelligent auxiliary identification method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115660966A true CN115660966A (en) | 2023-01-31 |
Family
ID=84983040
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211088455.4A Pending CN115660966A (en) | 2022-09-07 | 2022-09-07 | Machine vision intelligent auxiliary identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115660966A (en) |
-
2022
- 2022-09-07 CN CN202211088455.4A patent/CN115660966A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109822398B (en) | Numerical control machine tool broken cutter detection system and method based on deep learning | |
| CN111069976B (en) | Intelligent mobile monitoring system and method for damage of cutter for workshop or production line | |
| CN111300144B (en) | Automatic detection method for tool wear state based on image processing | |
| CN101456159B (en) | Spark identification tool-setting method and abrasive machining automatic system | |
| CN112016409A (en) | Deep learning-based process step specification visual identification determination method and system | |
| CN108760747A (en) | A kind of 3D printing model surface defect visible detection method | |
| CN112614105B (en) | Depth network-based 3D point cloud welding spot defect detection method | |
| CN113427731B (en) | Injection molding machine mold detection system and method based on machine vision | |
| CN113727022B (en) | Method and device for collecting inspection image, electronic equipment and storage medium | |
| CN109500657B (en) | Vision-based broken cutter detection method and system | |
| CN110977767A (en) | Casting defect distribution detection method and casting polishing method | |
| CN107234488A (en) | A kind of tool failure degree detection method based on Digit Control Machine Tool vision-based detection | |
| CN113487570A (en) | High-temperature continuous casting billet surface defect detection method based on improved yolov5x network model | |
| CN113744247A (en) | PCB welding spot defect identification method and system | |
| CN115456999A (en) | Automatic detection system and method for saw chain surface defects based on machine vision | |
| CN116258682A (en) | PCB solder paste defect detection method based on PSPNet and improved YOLOv7 | |
| CN115660966A (en) | Machine vision intelligent auxiliary identification method and system | |
| CN117409005A (en) | Defective product detection system and method for plate receiving machine based on image | |
| CN112801959A (en) | Auxiliary assembly system based on visual feature recognition | |
| CN112561885A (en) | YOLOv 4-tiny-based gate valve opening detection method | |
| CN112541367A (en) | Multiple two-dimensional code identification method based on deep learning and image processing | |
| CN107121063A (en) | The method for detecting workpiece | |
| CN108705689B (en) | Automatic tool setting process method for repairing surface microdefects of large-caliber precise optical crystal | |
| CN114663763A (en) | Aircraft skin assembly quality detection method based on digital twinning | |
| CN111054640A (en) | Automatic change mechanical detection device |
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 |