Aerostat material appearance defect detection system and method based on machine vision

Abstract

The invention discloses an aerostat material appearance defect detection system and method based on machine vision, and relates to the technical field of aerostat material detection.The aerostat material appearance defect detection system comprises a conveying assembly, a machine vision detection assembly is arranged above the conveying assembly, and a feeding assembly is arranged on one side of the feeding end of the conveying assembly; the influence of external light on the material is reduced as much as possible by adding the working cover, and the lamplight assembly can perform concentrated irradiation, form a brightness difference and enhance the abnormal display possibility of the irradiated area of the material, so that the light surrounding environment and the conveying speed of the aerostat material are controlled, the illumination intensity and controllable human eye recognition movement are determined, and the visual effect of the aerostat is improved. And the abnormal phenomenon in the material is judged in a man-machine cooperation mode with assistance of machine vision contrast detection, on one hand, the real-time condition of the material in the light band is recorded for personnel, on the other hand, the abnormity is automatically judged, personnel omission is prevented, and the defect discrimination rate is increased.

Description

Technical Field

[0001] The present invention relates to the technical field of aerostat material detection, and in particular to a system and method for detecting appearance defects of aerostat materials based on machine vision. Background Art

[0002] The aerostat is composed of an airbag made of polymer composite materials and various equipment and payloads arranged on the airbag. The airbag is filled with a floating gas with a density lower than that of air. The density difference between the gas inside and outside the airbag provides the buoyancy of the aerostat, thereby completing the launch of the aerostat. In the process of airbag processing, the size of each part of the airbag needs to be calculated and cut first. Before the cutting step, the material needs to be inspected, and the material with abnormal problems needs to be cut and avoided to ensure the quality of the airbag processing process. However, some defects larger than 0.5mm, such as broken wires, streaks, yarn slips, damage, etc., can be detected by the human eye during the detection process. After the material is detected to have defects, the defective areas are clearly marked with a marker. However, human eye detection is greatly affected by external conditions. On the one hand, the eyes will become tired or misjudged under long-term observation, and personal experience will be inconsistent with the standard for judging material defects. On the other hand, external light will also affect the appearance of the material. Once the material moves on the conveying device, it will be missed. Therefore, the use of manual inspection alone will result in a decrease in defect recognition rate and unstable quality.

[0003] Therefore, the present invention aims to provide a machine vision-based aerostat material appearance defect detection system. During the material inspection process, by controlling the light surrounding environment and the transmission speed, the light intensity and the controllable human eye recognition movement are determined, and with the assistance of machine vision comparison detection, the abnormal phenomena in the material are determined in a human-machine collaborative manner to improve the defect identification rate, and the machine vision comparison detection can automatically save the abnormal situation record, which is convenient for quality traceability. Summary of the invention

[0004] The present invention provides an aerostat material appearance defect detection system based on machine vision, aiming to solve the problems of low defect recognition rate, missed problem detection, wrong problem detection and so on in the process of aerostat material inspection.

[0005] In order to solve the above technical problems, the present invention provides an aerostat material appearance defect detection system based on machine vision, comprising a conveying component, a machine vision detection component is arranged above the conveying component, and a loading component is arranged on one side of the feeding end of the conveying component;

[0006] The machine vision detection component includes a lighting component, a flat-panel industrial computer, a control terminal and a work cover. The work cover is an opaque bottomless shell. The work cover covers the top of the transmission component. The lighting component, the flat-panel industrial computer and the control terminal are all arranged in the work cover. A manual operation area for manual collaborative inspection is provided on one side of the bottom of the machine vision detection component.

[0007] In some embodiments, the conveying assembly includes a conveying platform, on which a conveying roller is rotatably mounted, and a conveying belt is frictionally driven on the surface of the conveying roller.

[0008] In some embodiments, the conveyor belt is composed of a cloth layer and a PVC layer, and a regular pattern is arranged on the surface.

[0009] In some embodiments, the lighting assembly includes a mounting base, which is fixedly mounted on the transmission assembly, a mounting bracket is hinged on the top of the mounting base, and angle locking is achieved between the mounting base and the mounting bracket by friction, a high-brightness light strip is hinged on the top of the mounting bracket, and angle locking is achieved between the mounting bracket and the high-brightness light strip by friction, and a camera device is provided on the surface of the mounting bracket, and the camera device is electrically connected to the tablet industrial computer via a cable.

[0010] In some embodiments, the flat-panel industrial computer is fixedly mounted on the inner wall of the working cover, the flat-panel industrial computer and the manual operation area are respectively arranged on the left and right sides of the conveying component, a mounting plate is fixedly mounted on the bottom of the working cover, the control terminal is fixedly mounted on the mounting plate, the control terminal and the manual operation area are arranged on the same side of the conveying component, and the control terminal is electrically connected to the conveying component through a cable.

[0011] In some embodiments, the loading assembly includes a material support, which is arranged directly in front of the feed end of the conveying assembly. A material roller is rotatably mounted on the material support, and a float material belt is wound on the surface of the material roller.

[0012] The present invention also provides a method for using an aerostat material appearance defect detection system based on machine vision, which specifically comprises the following steps:

[0013] S1: Turn on the lighting components, tablet computer and control terminal;

[0014] S2: The personnel enter the working cover, adjust the conveyor belt to move forward, backward, and stop, and observe whether there is any problem with the imaging on the screen of the flat panel industrial computer. If there is any problem, check the camera equipment;

[0015] S3: When the aerostat material to be inspected is sent to the high-brightness light belt by the conveyor belt, the conveyor belt stops conveying, the tablet industrial computer is pressed to collect video, and the personnel observe whether there is any abnormal prompt on the tablet industrial computer;

[0016] S4: The conveyor belt is continuously controlled to move forward. When the personnel observe the screen of the tablet industrial computer and find any abnormality, they press the pause button of the control terminal and look down at the aerostat material for manual judgment;

[0017] S5: After the manual judgment is completed, mark the abnormal point with a marker, press Back, and when the abnormal point just judged is completely exited from the screen, press Stop again, and then press Forward to check whether the abnormal point just judged is displayed normally on the screen;

[0018] S6: When it is determined after review that the abnormal point is a material problem, press the screen record button to record the problem point as problem data and process it as machine data in the judgment database;

[0019] S7: Continue to press forward and repeat the actions of S3-S6 until all the required aerostat materials are inspected and the work is completed. Turn off the lighting components, the tablet industrial computer and the control terminal.

[0020] Compared with the related art, the aerostat material appearance defect detection system based on machine vision provided by the present invention has the following beneficial effects:

[0021] The present invention provides an aerostat material appearance defect detection system based on machine vision. A working cover is added to reduce the influence of external light on the material as much as possible, and a light component can irradiate in a concentrated manner to form a brightness difference, thereby enhancing the possibility of abnormal appearance of the irradiated area of ​​the material, thereby controlling the surrounding environment of the light and the speed of the aerostat material transmission, determining the light intensity and controllable human eye recognition movement, and assisting with machine vision comparison detection. The abnormal phenomenon in the material is determined in a human-machine collaborative manner. On the one hand, the real-time situation of the material in the light band is recorded for personnel, and on the other hand, the abnormality is automatically judged to prevent personnel from missing, thereby improving the defect identification rate. In addition, the machine vision comparison detection can automatically save the abnormal situation record, which is convenient for quality traceability.

[0022] The present invention provides a machine vision-based aerostat material appearance defect detection system, wherein the conveyor belt is composed of a cloth layer and a PVC layer, and a regular pattern is arranged on the surface. The aerostat material is also made of a polymer composite material, and only under strong light vertical irradiation, the material can be transparent and the background regular pattern is used as a reference to find out the abnormal problem. If not observed vertically, material defects such as broken wires, damage, dark patterns and the like are not easy to be observed.

[0023] The present invention provides an aerostat material appearance defect detection system based on machine vision. Manual judgment of material abnormalities, as abnormal problem data, can allow the machine database to self-learn and self-perfect, and finally form a complete data chain of the material, thereby providing a data judgment basis for the machine judgment. The database establishment process relies on the knowledge and experience of personnel, which is data knowledge that is currently unavailable on the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0025] Figure 2 It is a schematic diagram of the structure of the lighting assembly of the present invention;

[0026] Figure 3 It is the effect diagram of the use of the present invention.

[0027] Numbers in the figure: 1. Conveying component; 2. Machine vision inspection component; 3. Loading component; 11. Conveying platform; 12. Conveying roller; 13. Conveying belt; 21. Lighting component; 22. Flat-panel industrial computer; 23. Control terminal; 24. Working cover; 25. Manual operation area; 211. Mounting seat; 212. Mounting bracket; 213. Camera equipment; 214. High-brightness light belt; 31. Material bracket; 32. Material roller; 33. Airship material belt. DETAILED DESCRIPTION

[0028] Embodiment 1

[0029] This embodiment provides a system for detecting appearance defects of aerostat materials based on machine vision. Figure 1-3 As shown, the present invention includes a conveying component 1, a machine vision detection component 2 is arranged above the conveying component 1, and a loading component 3 is arranged on one side of the feeding end of the conveying component 1; the machine vision detection component 2 includes a lighting component 21, a flat-panel industrial computer 22, a control terminal 23 and a working cover 24, the working cover 24 is an opaque bottomless shell, the working cover 24 covers the top of the conveying component 1, the lighting component 21, the flat-panel industrial computer 22 and the control terminal 23 are all arranged in the working cover 24, and a manual operation area 25 for manual collaborative detection is provided on one side of the bottom of the machine vision detection component 2.

[0030] In this embodiment, the conveying component 1 is used for loading the aerostat material belt 33, and the machine vision detection component 2 is used for judging abnormal problems of the aerostat material belt 33. The judgment system is divided into two parts: manual judgment and machine vision comparison. The manual judgment relies on the light band of the lighting component 21 on the material to analyze and judge the light transmittance of the material; the machine vision comparison judgment relies on the camera device 213 to record the surface of the material under the light band, and then display it on the display of the flat-panel industrial computer 22. The system performs rapid comparison and processing based on the surface brightness data after illumination. Personnel can make judgments through the display of the flat-panel industrial computer 22, and record and save abnormal problems for reference for secondary judgment.

[0031] Among them, since broken threads, streaks and other defects are prone to visual ghosting and color aberration, natural light will affect personnel judgment, resulting in defects being ignored and missed. Therefore, it is necessary to add a work cover 24 to minimize the impact of external light on the material, and the light assembly 21 can focus on irradiation to form a brightness difference, thereby enhancing the possibility of abnormal appearance of the irradiated area of ​​the material.

[0032] Embodiment 2

[0033] Based on the first embodiment, Figure 1 As shown, the conveying assembly 1 of this embodiment includes a conveying platform 11, on which a conveying roller 12 is rotatably mounted, and a conveying belt 13 is frictionally driven on the surface of the conveying roller 12.

[0034] In this embodiment, the conveying platform 11 provides support for the detection process. Conveying rollers 12 are provided at both ends. The speed is controlled by a motor and a reducer. The conveying length of the aerostat material belt 33 does not exceed 5m per minute during the transmission process. Among them, the conveying belt 13 is composed of a cloth layer and a PVC layer. The bonding effect is achieved through a multi-layer processing technology. Its surface is provided with regular textures, which is convenient for reflection after light irradiation, so as to identify abnormal conditions. The conveying platform 11 can adjust the height, and the height is generally adjusted between 1.5m-1.8m. The width of the conveying component 1 is 1.2m. If the size needs to be increased, the size of the lighting component 21 and the working cover 24 needs to be modified at the same time.

[0035] Embodiment 3

[0036] On the basis of the second embodiment, the conveyor belt 13 of this embodiment is composed of a cloth layer and a PVC layer, and the surface is provided with regular patterns. The aerostat material is also made of a polymer composite material. Only when the strong light is vertically irradiated, the material can be transparent and the background regular pattern is used as a reference to find the abnormal problem. If it is not observed vertically, material defects such as broken wires, damage, dark patterns and other abnormal problems are not easy to be observed.

[0037] Embodiment 4

[0038] On the basis of embodiment one, the lighting assembly 21 of this embodiment includes a mounting base 211, which is fixedly mounted on the transmission assembly 1. A mounting bracket 212 is hinged on the top of the mounting base 211, and the mounting base 211 and the mounting bracket 212 are angle-locked by friction. A high-brightness light strip 214 is hinged on the top of the mounting bracket 212, and the mounting bracket 212 and the high-brightness light strip 214 are angle-locked by friction. A camera device 213 is provided on the surface of the mounting bracket 212, and the camera device 213 is electrically connected to the tablet industrial computer 22 via a cable.

[0039] In this embodiment, the light assembly 21 is inside the work cover 24 and opposite to the direction in which the personnel stand. The camera resolution of the camera device 213 is 2048*1536, the lens focal length is 8mm, the natural light filter is used, and the maximum laser lighting time is 70us. This is a shelf product and will not be described here. The mounting bracket 212 and the high-brightness light strip 214 are foldable. When in use, the high-brightness light strip 214 is kept horizontal and irradiated on the aerostat material strip 33. The peak illumination of the high-brightness light strip 214 at 1 meter is not greater than 1000lx. The personnel rotate for 2 hours. The personnel observe the display to make a judgment. After judging the abnormality, they directly check the material to confirm the existence of the abnormality. The light strip uses a high-concentration monochrome chip with a response time of less than 1 microsecond. The light assembly 21 has the advantages of miniaturization, high concentration and energy saving, providing a lighting basis for manual judgment and machine vision comparison.

[0040] Embodiment 5

[0041] On the basis of the first embodiment, the flat-panel industrial computer 22 of this embodiment is fixedly mounted on the inner wall of the working cover 24, the flat-panel industrial computer 22 and the manual operation area 25 are respectively arranged on the left and right sides of the transmission component 1, a mounting plate is fixedly mounted on the bottom of the working cover 24, the control terminal 23 is fixedly mounted on the mounting plate, the control terminal 23 and the manual operation area 25 are arranged on the same side of the transmission component 1, and the control terminal 23 is electrically connected to the transmission component 1 through a cable.

[0042] In this embodiment, the flat industrial computer 22 is installed on the inner wall of the working cover 24 by screws, and the signal transmission is connected to the camera device 213 by cables. The flat industrial computer 22 is composed of a host and a touch display. The internal software system is a visual sensor with a frame that meets the requirements and a shooting line frequency that meets the requirements to realize the collection of image data. By cooperating with the energy-saving, miniaturized, and high-illuminance lighting component 21, the high-definition collection of the measured cross-sectional image is realized. The data after the camera is analyzed and processed by the host, and the judgment situation is displayed on the display for manual reference. When the manual verification is abnormal, the image can be directly marked and saved through the touch display for reference for subsequent cutting work. The analysis database includes problems such as different densities, broken wires, stripes, yarn walking, and damage. The database can be sampled and expanded, and can be supplemented after manual screening. Among them, the control terminal 23 controls the running speed of the conveyor belt 13, and the control method is manual adjustment. The buttons are divided into forward, backward, and emergency stop, and the maximum speed does not exceed 5m per minute.

[0043] Embodiment 6

[0044] Based on the first embodiment, Figure 3 As shown, the loading assembly 3 of this embodiment includes a material support 31, which is arranged just in front of the feeding end of the conveying assembly 1. A material roller 32 is rotatably mounted on the material support 31, and a float material belt 33 is wound around the surface of the material roller 32.

[0045] In this embodiment, the material support 31 is placed in front of the conveying component 1 , and after the aerostat material belt 33 is pulled out from the support, it enters the interior of the machine vision inspection component 2 through the conveying belt 13 .

[0046] Working principle: By controlling the surrounding environment of the light and the speed of the material transmission of the aerostat, the light intensity and the controllable human eye recognition movement are determined, and the machine vision comparison detection is assisted. The abnormal phenomena in the material are judged in a collaborative way. On the one hand, the real-time situation of the material in the light band is recorded for the personnel, and on the other hand, the abnormality is automatically judged to prevent personnel from missing, which improves the defect identification rate. In addition, the machine vision comparison detection can automatically save the abnormal situation record to facilitate the traceability of quality. Manual judgment of material abnormalities, as abnormal problem data, can allow the machine database to self-learn and self-improve, and finally form a complete data chain of the material, and then make data judgment basis for the machine judgment. The process of establishing this database relies on the knowledge and experience of personnel, which is data knowledge that is currently unavailable on the market.

[0047] Implementation steps:

[0048] S1: Turn on the lighting assembly 21, the tablet computer 22 and the control terminal 23;

[0049] S2: The personnel enter the working cover, adjust the conveyor belt to move forward, backward, and stop, and observe whether there is any problem with the imaging on the screen of the flat panel industrial computer 22. If there is any problem, check the camera device 213;

[0050] S3: When the aerostat material to be inspected is sent to the high-brightness light belt 214 by the conveyor belt 13, the conveyor belt 13 stops conveying, the tablet industrial computer 22 is pressed to collect video, and the personnel observe whether there is any abnormal prompt on the tablet industrial computer 22;

[0051] S4: The conveyor belt 13 is continuously controlled to move forward. After observing the screen of the tablet computer 22 and finding any abnormality, the personnel press the pause button of the control terminal 23 and look down to observe the aerostat material for manual judgment;

[0052] S5: After the manual judgment is completed, mark the abnormal point with a marker, press Back, and when the abnormal point just judged is completely exited from the screen, press Stop again, and then press Forward to check whether the abnormal point just judged is displayed normally on the screen;

[0053] S6: When it is determined after review that the abnormal point is a material problem, press the screen record button to record the problem point as problem data and process it as machine data in the judgment database;

[0054] S7: Continue to press the forward button and repeat the actions of S3-S6 until all the aerostat materials that require inspection are inspected. The work is completed and the lighting assembly 21, the tablet industrial computer 22 and the control terminal 23 are turned off.

Claims

1. A machine vision-based aerostat material appearance defect detection system, including a transmission component, characterized in that: A machine vision detection component is arranged above the conveying component, and a loading component is arranged on one side of the feeding end of the conveying component; The machine vision detection component includes a lighting component, a flat-panel industrial computer, a control terminal and a work cover. The work cover is an opaque bottomless shell. The work cover covers the top of the transmission component. The lighting component, the flat-panel industrial computer and the control terminal are all arranged in the work cover. A manual operation area for manual collaborative inspection is provided on one side of the bottom of the machine vision detection component.

2. The machine vision-based aerostat material appearance defect detection system according to claim 1, characterized in that: The conveying assembly comprises a conveying platform, on which a conveying roller is rotatably mounted, and a conveying belt is frictionally driven on the surface of the conveying roller.

3. The machine vision-based aerostat material appearance defect detection system according to claim 1, characterized in that: The conveyor belt is composed of a cloth layer and a PVC layer, and the surface is provided with regular patterns.

4. The machine vision-based aerostat material appearance defect detection system according to claim 1, characterized in that: The lighting assembly includes a mounting base, which is fixedly mounted on the transmission assembly. A mounting bracket is hinged on the top of the mounting base, and angle locking is achieved between the mounting base and the mounting bracket through friction. A high-brightness light strip is hinged on the top of the mounting bracket, and angle locking is achieved between the mounting bracket and the high-brightness light strip through friction. A camera device is arranged on the surface of the mounting bracket, and the camera device is electrically connected to the tablet industrial computer through a cable.

5. The machine vision-based aerostat material appearance defect detection system according to claim 1, characterized in that: The flat-panel industrial computer is fixedly mounted on the inner wall of the working cover, the flat-panel industrial computer and the manual operation area are respectively arranged on the left and right sides of the conveying component, a mounting plate is fixedly mounted on the bottom of the working cover, the control terminal is fixedly mounted on the mounting plate, the control terminal and the manual operation area are arranged on the same side of the conveying component, and the control terminal is electrically connected to the conveying component through a cable.

6. The machine vision-based aerostat material appearance defect detection system according to claim 1, characterized in that: The feeding assembly comprises a material support, which is arranged in front of the feeding end of the conveying assembly. A material roller is rotatably mounted on the material support, and a float material belt is wound on the surface of the material roller.

7. A method for using the aerostat material appearance defect detection system based on machine vision as claimed in any one of claims 1 to 6, characterized in that: The specific steps include: S1: Turn on the lighting components, tablet computer and control terminal; S2: The personnel enter the working cover, adjust the conveyor belt to move forward, backward, and stop, and observe whether there is any problem with the imaging on the screen of the flat panel industrial computer. If there is any problem, check the camera equipment; S3: When the aerostat material to be inspected is sent to the high-brightness light belt by the conveyor belt, the conveyor belt stops conveying, the tablet industrial computer is pressed to collect video, and the personnel observe whether there is any abnormal prompt on the tablet industrial computer; S4: The conveyor belt is continuously controlled to move forward. When the personnel observe the screen of the tablet industrial computer and find any abnormality, they press the pause button of the control terminal and look down at the aerostat material for manual judgment; S5: After the manual judgment is completed, mark the abnormal point with a marker, press Back, and when the abnormal point just judged is completely exited from the screen, press Stop again, and then press Forward to check whether the abnormal point just judged is displayed normally on the screen; S6: When it is determined after review that the abnormal point is a material problem, press the screen record button to record the problem point as problem data and process it as machine data in the judgment database; S7: Continue to press forward and repeat the actions of S3-S6 until all the required aerostat materials are inspected and the work is completed. Turn off the lighting components, the tablet industrial computer and the control terminal.

Images