WO2021179680A1 - 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法 - Google Patents

基于机器视觉的挡风玻璃划痕接续检测装置和检测方法 Download PDF

Info

Publication number
WO2021179680A1
WO2021179680A1 PCT/CN2020/131203 CN2020131203W WO2021179680A1 WO 2021179680 A1 WO2021179680 A1 WO 2021179680A1 CN 2020131203 W CN2020131203 W CN 2020131203W WO 2021179680 A1 WO2021179680 A1 WO 2021179680A1
Authority
WO
WIPO (PCT)
Prior art keywords
ball screw
glass
image
bottom plate
windshield
Prior art date
Application number
PCT/CN2020/131203
Other languages
English (en)
French (fr)
Inventor
石颉
白雅玲
池越
周亚同
胡凯
Original Assignee
苏州鑫睿益荣信息技术有限公司
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 苏州鑫睿益荣信息技术有限公司 filed Critical 苏州鑫睿益荣信息技术有限公司
Publication of WO2021179680A1 publication Critical patent/WO2021179680A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
    • G01N21/958—Inspecting transparent materials or objects, e.g. windscreens
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B08—CLEANING
    • B08B—CLEANING IN GENERAL; PREVENTION OF FOULING IN GENERAL
    • B08B1/00—Cleaning by methods involving the use of tools
    • B08B1/10—Cleaning by methods involving the use of tools characterised by the type of cleaning tool
    • B08B1/14—Wipes; Absorbent members, e.g. swabs or sponges
    • B08B1/143—Wipes
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B08—CLEANING
    • B08B—CLEANING IN GENERAL; PREVENTION OF FOULING IN GENERAL
    • B08B1/00—Cleaning by methods involving the use of tools
    • B08B1/20—Cleaning of moving articles, e.g. of moving webs or of objects on a conveyor
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B08—CLEANING
    • B08B—CLEANING IN GENERAL; PREVENTION OF FOULING IN GENERAL
    • B08B11/00—Cleaning flexible or delicate articles by methods or apparatus specially adapted thereto
    • B08B11/04—Cleaning flexible or delicate articles by methods or apparatus specially adapted thereto specially adapted for plate glass, e.g. prior to manufacture of windshields
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
    • G01N21/958—Inspecting transparent materials or objects, e.g. windscreens
    • G01N2021/9586—Windscreens

Definitions

  • the invention relates to the technical field of automobile detection, and in particular to a windshield scratch connection detection device and a detection method based on machine vision.
  • the front windshield of a car is the first barrier to protect the lives of drivers and passengers. If the front windshield is cracked and not repaired in time, it will have a great impact on the subsequent use of the car. Therefore, people are paying more attention to the detection of scratches on the front windshield.
  • the detection scheme of the Chinese patent with application number 201720843553.2 can adjust the height of the device and expand the scope of detection, but the light is irradiated on the glass through the transparent plate. If the transparent plate has dust or scratches, the detection accuracy will be reduced, and the glass is scratched. It is still obtained through manual observation.
  • the technical problem solved by the present invention is to provide a machine vision-based windshield scratch splicing detection device, which can adapt to any size of the front windshield, and can perform splicing detection without turning the front and back of the glass , At the same time, there is a dust removal device, which is used to clean the dust on the surface of the glass for scratch detection.
  • a windshield scratch splicing detection device based on machine vision including a support frame, a conveying device, a cleaning device, a lighting device, and an image acquisition and processing device.
  • the right side of the support frame The upper and lower parts are symmetrical darkrooms.
  • the conveying device is arranged at the front end of the support frame for fixing and sending different sizes of glass into the darkroom.
  • the cleaning device is installed in the center of the support frame on the left side of the darkroom for entering the glass. The dust on the glass surface is cleaned in the darkroom.
  • the top of the symmetrical darkroom is provided with support rods, and each of the support rods is equipped with a plurality of the lighting devices used to illuminate the glass, and the image acquisition and processing device It is fixed at the center of the middle support frame on the top of the upper and lower darkrooms, and is used to take images of the glass surface and perform processing and analysis to determine whether there are scratches on the glass surface.
  • the conveying device includes a large bottom plate, and two left and right first linear guide rails are fixed on the upper side of the large bottom plate.
  • the side sliding block, the inner side of the two left and right first linear guides and close to the first linear guide are respectively installed with the left ball screw and the right ball screw, and also includes two first ball screw nuts and The first nut connecting seat, the two first ball screw nuts are connected to the first nut connecting seat and are respectively connected to the left ball screw and the right ball screw, for correspondingly controlling the two first straight lines on the left and right
  • one end of the first ball screw is equipped with a first fixed bearing seat, and the other end is equipped with a first supporting bearing seat for fixing the first ball screw.
  • first small bottom plate and a second small bottom plate that are installed horizontally.
  • the nut connecting seat of the right ball screw is connected, the lower surface of the second small bottom plate is connected to the lower slide block of the left and right first linear guides respectively, and the left bottom surface of the second small bottom plate is also connected to the left ball screw
  • a second linear guide rail is installed on the first small bottom plate and the second small bottom plate respectively, two tightening plates are installed on the second linear guide rail, and both sides of the second linear guide rail are respectively installed
  • each second ball screw is equipped with a second nut connecting seat, which is connected to the second ball screw nut, and the second fixed bearing seat is installed on one side of the second ball screw.
  • a second supporting bearing seat is installed at one end for fixing the second ball screw, and it also includes a second motor.
  • the second motor is installed at one end of the second fixed bearing seat through a round nut.
  • the second nut connection base of the ball screw is connected, and also includes a plc controller and a touch screen.
  • the first motor and the second motor are respectively connected to the plc controller, and the plc controller is connected to the touch screen.
  • the touch screen controls the motor and then controls the slider And step up the movement of the board.
  • the cleaning device includes a third servo motor, a third ball screw, a third lifting platform, a roller, a roller bearing and a support rod, and the upper surface of the third lifting platform is installed on the top of the upper and lower darkrooms, And one side of the third lifting platform coincides with the center of the support rod on the top of the darkroom.
  • a third servo motor is installed on the inner side of the upper surface of the third lifting platform. The third servo motor is connected to the third ball screw.
  • the rotating shafts of the three lifting platforms are connected by cross brackets, the lower surface of the third lifting platform is connected with the support rod, the support rod is connected with the roller bearing, the roller is fixed on the outer surface of the roller bearing, and the surface of the roller is attached to wipe the glass surface
  • the dust glass cleaning cloth also includes a plc controller and a touch screen, the third servo motor is connected to the plc controller, and the plc controller is connected to the touch screen, and the third servo motor is controlled by the touch screen, and the third servo motor controls the ball screw The expansion and contraction, and then control the third lifting platform to move up and down.
  • the lighting device includes an LED tube, a fourth servo motor, a fourth ball screw, a fourth lifting platform, and an LED tube frame.
  • the LED tube frame is installed on the lower surface of the lifting platform.
  • the lamp tube is connected to the lamp tube frame.
  • a fourth servo motor is installed inside the upper surface of the lifting platform.
  • the fourth servo motor is connected to the fourth ball screw.
  • the fourth ball screw is connected to the shaft of the fourth lifting platform through a cross bracket.
  • the top of the darkroom is equipped with a plurality of the lighting devices independently controlled, and the multiple lighting devices are installed at the positions near the center on both sides of the support rod in the middle of the darkroom top, and are arranged linearly. It also includes a plc controller and a touch screen.
  • the four servo motors are connected to the plc controller, and the plc controller is connected to the touch screen.
  • the fourth servo motor is controlled by the touch screen.
  • the fourth servo motor controls the expansion and contraction of the fourth ball screw, and then controls the fourth lifting platform to move up and down.
  • the image acquisition and processing device includes an industrial camera, an industrial computer, an L-shaped camera bracket, and a universal adjustment base.
  • the industrial camera is fixed on the universal adjustment base, and the surface of the universal adjustment base is mounted on the L-shaped camera.
  • One side of the frame and the other side are fixed to the center of the middle support frame on the top of the upper and lower darkrooms.
  • the industrial camera includes a power socket connected to an external power source and a network cable socket connected to an industrial computer.
  • the right side of the support frame is hollowed out, and the upper and lower darkrooms are provided with the lighting device and the image acquisition device for continuously detecting the front and back of the glass.
  • the invention also discloses a windshield scratch connection detection method based on machine vision, the steps are:
  • the first step is to control the plc controller through the touch screen, and then control the sliders on the first linear guides on the left and right sides to drive the first small bottom plate and the second small bottom plate to move, so that the two small bottom plates move longitudinally and relative to each other at the same time. Stop when the glass width is level. After the four tightening plates on the first small bottom plate and the second small bottom plate tighten the four corners of the glass, the slider connected to the tightening plate is controlled by the plc controller to make them move toward the darkroom. Moving at a speed;
  • the third servo motor of the upper and lower cleaning device of the dark room is controlled by the touch screen to push the third ball screw to expand and contract, and the roller is lowered to stop when it just touches the glass surface. Rotate and use the cleaning cloth on the surface to remove dust on the surface to make subsequent detection more accurate.
  • the cleaning device rises and starts to detect scratches.
  • the width of the glass determines the number of LED lights that are turned on.
  • the scratch detection uses backlight illumination The detection method of the LED lamp.
  • the upper lighting device and the industrial camera in the lower dark room are turned off, the upper industrial camera is turned on, and the industrial camera adjusts its direction through the universal base to cover the glass surface.
  • the LED light below is driven by the motor to a position that can fully illuminate the glass surface.
  • the industrial camera starts to work, and transmits the windshield image to the industrial computer connected to the industrial camera, and uses the industrial computer to check the scratches. The image is automatically detected.
  • the industrial camera in the upper dark room turns off the lower dark room, and the lighting device in the lower dark room rises at the same time.
  • the LED light in the upper dark room is driven by the motor to be lowered to a position that can fully illuminate the glass surface.
  • the industrial camera below starts to work;
  • the third step is to transfer the image to the industrial computer connected to the industrial camera for image processing and analysis. After all the inspections are completed, the glass is driven by the motor to return to the outside of the dark room on the left to complete an inspection.
  • the image analysis step is:
  • Step 1 Image smoothing: The image will be disturbed by noise during the collection process. For the accuracy and efficiency of image post-processing, first smooth the collected glass image, and use median filtering to smooth the image.
  • the filter window adopts a square detection window, and the size template adopts a 7 ⁇ 7 template, thereby effectively suppressing the noise in the image;
  • Step two image segmentation: image segmentation is a key step in scratch detection.
  • the local dynamic threshold segmentation method is mainly used to extract all the defect information and convert it into a binary image;
  • Step three morphological processing: In order to obtain further information on the scratches, the method of expansion morphology is used to process the scratches, so that it can simplify the image data while maintaining its basic shape characteristics.
  • Step 4 Image extraction: In order to restore the defect shape, a skeleton extraction method is used;
  • Step 5 Compare the result with the standard feature to determine whether the defect is a scratch.
  • the present invention uses machine vision to detect glass scratches, and replaces manual operation with machine detection, which reduces labor intensity and improves detection accuracy.
  • the present invention can be adapted to the detection of glass of any size, and the applicable scope of the glass detection is improved.
  • the present invention adopts upper and lower double darkrooms, and does not need to be turned over to continuously detect the upper and lower sides of the glass, which improves the detection efficiency.
  • the present invention is equipped with a cleaning device, which can remove dust on the surface of the glass before detecting the scratches on the glass, so that the subsequent detection is more accurate.
  • the LED lamp of the present invention adopts a separate control method, and the corresponding number of LED lamps are turned on according to the width and length of the glass, so as to prevent the waste of unnecessary lighting lamps and save power resources.
  • Figure 1 is a schematic diagram of the overall structure of the present invention plus a dark room
  • Figure 2 is a schematic diagram of the overall structure of the present invention to remove the darkroom
  • FIG. 3 is a schematic diagram of the three-dimensional structure of the conveying device of the present invention.
  • FIG. 4 is a schematic diagram of the three-dimensional structure of the cleaning device of the present invention.
  • FIG. 5 is a schematic diagram of the three-dimensional structure of the lighting device of the present invention.
  • FIG. 6 is a schematic diagram of the three-dimensional structure of the industrial camera of the present invention.
  • Figure 7 is a flow chart of the windshield scratch image detection flow chart of the present invention
  • the windshield scratch splicing detection device based on machine vision as shown in Fig. 1 and Fig. 2 includes a supporting frame 5, a conveying device 1, a cleaning device 2, a lighting device 3 and an image acquisition and processing device 4, the supporting frame 5
  • the upper and lower parts on the right are symmetrical darkrooms.
  • the conveying device 1 is arranged at the front end of the support frame 5 for fixing and sending different sizes of glass into the darkroom.
  • the cleaning device 2 is installed in the center of the support frame 5 on the left side of the darkroom.
  • the top of the symmetrical dark room is provided with support rods 24, each of which is equipped with a plurality of the illuminators used to illuminate the glass Device 3, the image acquisition and processing device 4 is fixed at the center of the middle support frame 5 on the top of the upper and lower darkrooms, and is used to take images of the glass surface and perform processing and analysis to determine whether there are scratches on the glass surface.
  • the right side of the support frame 5 It is hollow, and the upper and lower darkrooms are equipped with the lighting device 3 and the image acquisition device, which can be used to detect the front and back of the glass.
  • the conveying device 1 conveys the windshield to be inspected forward.
  • the cleaning device 2 cleans it first, and then sends it into the darkroom.
  • the upper darkroom lighting device 3 is turned on, and the image acquisition processing device 4 performs image acquisition processing to detect whether there are scratches on the upper surface of the windshield.
  • the lighting device 3 in the lower dark room is turned on, and the image acquisition processing device 4 performs image acquisition processing to detect whether there are scratches on the lower surface of the windshield, so as to realize the detection of both surfaces.
  • This device uses machine vision to detect glass scratches. Machine detection is used instead of manual operation, which reduces labor intensity and improves detection accuracy.
  • the present invention adopts upper and lower double darkrooms, and does not need to be turned over to continuously detect the upper and lower sides of the glass, which improves the detection efficiency.
  • the present invention uses a cleaning device
  • the setting of 2 can remove dust on the surface of the glass before detecting scratches on the glass, making subsequent inspections more accurate.
  • the conveying device 1 may be a screw conveying module, a cylinder conveying mechanism, etc.
  • the conveying device 1 in this embodiment includes a large bottom plate 110, and the top of the large bottom plate 110 is fixed
  • the left and right first linear guides 12 are arranged longitudinally.
  • the upper and lower sliders are installed on the left and right first linear guides 12 respectively.
  • the inner sides of the left and right first linear guides 12 are close to the first
  • the position of the linear guide rail 12 is correspondingly installed with the left ball screw and the right ball screw, and also includes two first ball screw nuts and a first nut connecting seat, the two first ball screw nuts Connected with the first nut connecting seat and respectively connected to the left ball screw and the right ball screw, used to correspondingly control the movement of the sliders on the two left and right first linear guides 12.
  • One end of the first ball screw is installed with the first ball screw.
  • a fixed bearing seat the other end of which is equipped with a first supporting bearing seat for fixing the ball screw, and also includes a first motor 11 which is mounted on one end of the first fixed bearing seat through a round nut;
  • first small bottom plate 111 and a second small bottom plate 113 that are installed horizontally.
  • the lower side surface is also connected to the nut connecting seat of the right ball screw
  • the lower surface of the second small bottom plate 113 is connected to the lower sliders of the left and right first linear guides 12, and the left bottom surface of the second small bottom plate 113 It is also connected to the nut connecting seat of the left ball screw.
  • the first small bottom plate 111 and the second small bottom plate 113 are respectively mounted with a second linear guide 13 and two tightening plates are mounted on the second linear guide 13 112.
  • a second ball screw is installed on both sides of the second linear guide 13 respectively, and a second nut connecting seat is installed on each second ball screw, which is connected to the second ball screw nut.
  • each tightening plate 112 is respectively connected to the second nut connecting seat of the corresponding second ball screw, and further includes a plc controller 7 and a touch screen 6, the first motor 11 and the second motor 14 respectively It is connected to the plc controller 7, which is connected to the touch screen 6, and the touch screen 6 controls the motor to control the movement of the slider and the tightening plate 112.
  • the cleaning device 2 can be a dust suction device, etc.
  • the cleaning device 2 of this embodiment includes a third servo motor 27, a third ball screw 28, a third lifting platform 23, and a roller 25 ,
  • the roller bearing 26 and the support rod 24, the upper surface of the third lifting platform 23 is installed on the top of the upper and lower darkrooms, and one side of the third lifting platform 23 coincides with the center of the support rod 24 on the top of the darkroom, on the third lifting platform 23
  • a third servo motor 27 is installed inside the surface.
  • the third servo motor 27 is connected to the third ball screw 28.
  • the third ball screw 28 is connected to the shaft 21 of the third lifting platform 23 through a cross bracket 22.
  • the third lifting platform The lower surface of 23 is connected to the support rod 24, the support rod 24 is connected to the roller bearing 26, the roller 25 is fixed on the outer surface of the roller bearing 26, and the surface of the roller 25 is attached with a dust glass cleaning cloth used to wipe the glass surface, and It includes a plc controller 7 and a touch screen 6.
  • the third servo motor 27 is connected to the plc controller 7, and the plc controller 7 is connected to the touch screen 6.
  • the touch screen 6 controls the third servo motor 27, and the third servo motor 27 controls the ball
  • the expansion and contraction of the screw rod controls the third lifting platform 23 to move up and down.
  • the lighting device 3 includes an LED tube 31, a fourth servo motor, a fourth ball screw, a fourth lifting platform, an LED tube frame 32, and the LED tube frame 32 is installed On the lower surface of the lifting platform, the LED tube 31 is connected to the tube frame, and a fourth servo motor is installed inside the upper surface of the lifting platform.
  • the fourth servo motor is connected to the fourth ball screw, and the fourth ball screw is connected to the fourth lifting
  • the rotating shaft 21 of the table is connected by a cross bracket 22, and the top of the upper and lower darkrooms are each equipped with a plurality of independently controlled illuminating devices 3, and the multiple illuminating devices 3 are installed at the positions near the center on both sides of the middle support rod 24 on the top of the darkroom, and
  • the linear arrangement also includes a plc controller 7 and a touch screen 6.
  • the fourth servo motor is connected to the plc controller 7, and the plc controller 7 is connected to the touch screen 6, and the fourth servo motor is controlled through the touch screen 6, and the fourth servo motor is controlled
  • the expansion and contraction of the fourth ball screw then controls the up and down movement of the fourth lifting platform.
  • the image acquisition and processing device 4 includes an industrial camera 45, an industrial computer, an L-shaped camera bracket 41, and a universal adjustment base 42.
  • the industrial camera 45 is fixed on the universal adjustment base 42
  • the universal adjustment base 42 is mounted on one side of the L-shaped camera frame on the surface, and the other side is fixed to the center of the middle support frame 5 on the top of the upper and lower darkrooms.
  • the industrial camera 45 includes a power socket 43 connected to an external power source and an industrial computer Network cable socket 44.
  • the plc controller 7 is controlled through the touch screen 6, and then the sliders on the first linear guides 12 on the left and right sides are controlled to drive the first small bottom plate 111 and the second small bottom plate 113 to move, so that the two small bottom plates are longitudinally opposite at the same time. Movement, the movement stops when it is flush with the width of the glass.
  • the PLC controller 7 controls the connection with the tightening plate 112 Slide block to move at a common speed in the direction of the darkroom;
  • the third servo motor 27 of the upper and lower cleaning device 2 of the dark room is controlled by the touch screen 6 to push the third ball screw 28 to expand and contract, and the roller 25 is lowered until it just touches the glass surface. Stop, through the rotation of the roller 25, use the cleaning cloth on the surface to remove the surface dust, so that the subsequent detection is more accurate.
  • the cleaning device 2 rises and starts to detect scratches, and the number of LED lights to be turned on is determined by the width of the glass. , Scratch detection adopts a backlight illuminating LED lamp detection method.
  • the upper lighting device 3 and the lower industrial camera 45 in the dark room are turned off, the upper industrial camera 45 is turned on, and the industrial camera 45 passes Adjust its direction to the base to cover the entire detection area of the glass surface.
  • the LED light below is driven by the motor to be lowered to a position that can fully illuminate the glass surface.
  • the industrial camera 45 starts to work and transmits the windshield image to the industry.
  • the industrial computer connected to the camera 45 is used to automatically detect the scratch image.
  • the industrial camera 45 in the upper dark room turns off the lower dark room, and at the same time the lighting device 3 in the lower dark room rises, and the LED in the upper dark room
  • the lamp is driven by the motor and lowered to a position that can completely illuminate the glass surface, and the industrial camera 45 below starts to work;
  • the image is transmitted to the industrial computer connected to the industrial camera 45 for image processing and analysis.
  • the glass is driven by the motor to return to the outside of the dark room on the left to complete one inspection.
  • Step 1 Image smoothing: The image will be disturbed by noise during the collection process. For the accuracy and efficiency of image post-processing, first smooth the collected glass image, and use median filtering to smooth the image.
  • the filter window adopts a square detection window, and the size template adopts a 7 ⁇ 7 template, thereby effectively suppressing the noise in the image;
  • Step two image segmentation: image segmentation is a key step in scratch detection.
  • the local dynamic threshold segmentation method is mainly used to extract all the defect information and convert it into a binary image;
  • Step three morphological processing: In order to obtain further information on the scratches, the method of expansion morphology is used to process the scratches, so that it can simplify the image data while maintaining its basic shape characteristics.
  • Step 4 Image extraction: In order to restore the defect shape, a skeleton extraction method is used;
  • Step 5 Compare the result with the standard feature to determine whether the defect is a scratch.

Landscapes

  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

基于机器视觉的挡风玻璃划痕接续检测装置和检测方法，该装置包括传送装置(1)，清洁装置(2)，照明装置(3)以及图像采集处理装置(4)，传送装置(1)整体安装在支撑架(5)上，支撑架(5)右侧的上下两部分均为暗室，且支撑架(5)右侧表面镂空。暗室内部设有清洁装置(2)，照明装置(3)以及图像采集处理装置(4)。通过传送装置(1)固定挡风玻璃并向暗室传送，清洁装置(2)对进入暗室的玻璃进行清洁，照明装置(3)使用个数由玻璃的尺寸决定。上下方的工业相机(45)接续检测玻璃的正反面，并将拍到的图像送入工控机对图像进行处理分析，最终判断玻璃表面是否有划痕。本装置基于机器视觉分析图像，能够适应不同尺寸挡风玻璃，且照明装置(3)可调节个数，节约电力。

Description

基于机器视觉的挡风玻璃划痕接续检测装置和检测方法 技术领域
本发明涉及汽车检测技术领域，特别涉及基于机器视觉的挡风玻璃划痕接续检测装置和检测方法。
背景技术
汽车前挡风玻璃作为汽车车身的重要组成部分，是保护驾驶人和乘客生命安全的第一道屏障。如果前挡风玻璃出现裂纹不及时修复，对汽车后续的使用会产生很大的影响。所以对前挡风玻璃划痕的检测，也是人们比较关注的。
目前，对于汽车前挡风玻璃的检测，一般是靠人工观察检测，这样不仅会耗费大量的人力，而且人工能观察到的划痕是有限的，一些特别微小的划痕可能无法观察出来，容易造成检测结果不精确。申请号为201720879709.2的中国专利提供了一种划痕检测装置，该装置通过电机带动承载板以及承载板上的玻璃转动，提高了检测效率，降低了人员的劳动强度，但是由于玻璃是双曲面结构，这种方式只能检测一面，并且本质上是人工检测，存在不足。申请号为201720843553.2的中国专利的检测方案能够调节装置的高度，扩大检测适用范围，但是光线是通过透明板照射到玻璃上，如果透明板上有灰尘或者划痕会降低检测精度，同时玻璃划痕依然是通过人工观察得到。
为了解决上述问题，设计一种基于机器视觉的自适应任何尺寸挡风玻璃，并且能够无需翻面高效接续检测玻璃正反两面划痕的装置很有必要。
发明内容
本发明解决的技术问题是提供提供一种基于机器视觉的挡风玻璃划痕接续检测装置，该装置能够自适应任何尺寸的前挡风玻璃，并且能够无需翻面对玻璃的正反面进行接续检测，同时有除灰装置，用于清扫玻璃表面的灰尘进而进行划痕检测。
本发明解决其技术问题所采用的技术方案是：基于机器视觉的挡风玻璃划痕接续检测装置，包括支撑架、传送装置，清洁装置，照明装置和图像采集处理装置，所述支撑架右侧上下部分为对称的暗室，所述传送装置设置在支撑架的前端，用于将不同尺寸的玻璃固定并送入暗室，所述清洁装置安装在暗室左侧支撑架的中央，用于在玻璃进入暗室时清除玻璃表面的灰尘，所述上下对称的暗室的顶部均设置有支撑杆，所述支撑杆上各安装有多个用来给玻璃提供照明的所述照明装置，所述图像采集处理装置固定在上下暗室顶部中间支撑架的中央，用于拍摄玻璃表面的图像并进行处理分析，从而判定玻璃表面是否有划痕。
进一步的是：所述传送装置包括大底板，所述大底板的上方固定有左右两个纵向设置的第一直线导轨，左右两个第一直线导轨上分别安装有上侧滑块和下侧滑块，左右两个第一直线导轨的内侧并靠近第一直线导轨的位置分别对应安装有左侧滚珠丝杆和右侧滚珠丝杆，还包括两个第一滚珠丝杆螺母与第一螺母连接座，所述两个第一滚珠丝杆螺母与第一螺母连接座相连并分别连接在左侧滚珠丝杆和右侧滚珠丝杆，用于对应控制左右两个第一直线导轨上滑块的运动，第一滚珠丝杆一端安装第一固定轴承座，另一端安装第一支撑轴承座，用于固定第一滚珠丝杆，还包括第一电机，所述第一电机通过圆螺母安装在第一固定轴承座一端；
还包括横向安装的第一小底板和第二小底板，所述第一小底板的下表面分别与左右第一直线导轨的上侧滑块连接，第一小底板的右侧下表面还与右侧滚珠丝杆的螺母连接座连接，第二小底板的下表面分别与左右第一直线导轨的下侧滑块连接，第二小底板的左侧下表面还与左侧滚珠丝杆的螺母连接座连接，所述第一小底板和第二小底板上面分别安装有第二直线导轨，所述第二直线导轨上安装有两个加紧板，所述第二直线导轨的两侧分别安装有第二滚珠丝杆，每个第二滚珠丝杆上都安装有一个第二螺母连接座，与第二滚珠丝杆螺母相连，第二滚珠丝杆的一侧安装第二固定轴承座，另一端安装第二支撑轴承座，用于固定第二滚珠丝杆，还包括第二电机，所述第二电机通过圆螺母安装在第二固定轴承座一端，每个加紧板分别与对应的第二滚珠丝杆的第二螺母连接座连接，还包括plc控制器和触摸屏，所述第一电机、第二电机分别与plc控制器相连，plc控制器与触摸屏连接，通过触摸屏控制电机进而控制滑块和加紧板的运动。
进一步的是：所述的清洁装置包括第三伺服电机，第三滚珠丝杆，第三升降台，滚轮，滚轮轴承和支撑杆，所述第三升降台的上表面安装在上下暗室的顶部，并且第三升降台的一侧与暗室顶部支撑杆中央重合，第三升降台上表面内侧安装有第三伺服电机，第三伺服电机与第三滚珠丝杆相连，第三滚珠丝杆杠与第三升降台的转轴通过十字支架连接，第三升降台的下表面与支撑杆相连接，支撑杆与滚轮轴承相连，滚轮固定在滚轮轴承的外表面，且滚轮的表面附着有用来擦除玻璃表面的灰尘玻璃清洁布，还包括plc控制器和触摸屏，所述第三伺服电机与plc控制器相连，plc控制器再与触摸屏连接，通过触摸屏控制第三伺服电机，第三伺服电机控制滚珠丝杆的伸缩，进而控制第三升降台上下移动。
进一步的是：所述的照明装置包括LED灯管，第四伺服电机，第四滚珠丝杆，第四升降台，LED灯管架，所述LED灯管架安装在升降台的下表面，LED灯管与灯管架连接，升降台上表面内侧安装有第四伺服电机，第四伺服电机与第四滚珠丝杆相连，第四滚珠丝杆与第四升降台的转轴通过十字支架连接，上下暗室的顶部各安装有多个各自独立控制的所述照明装置，多个照明装置安装在暗室顶部中间支撑杆两侧靠近中央的位置，并线性排列，还包括 plc控制器和触摸屏，所述第四伺服电机与plc控制器相连，plc控制器再与触摸屏连接，通过触摸屏控制第四伺服电机，第四伺服电机控制第四滚珠丝杆的伸缩，进而控制第四升降台上下移动。
进一步的是：所述的图像采集处理装置包括工业相机、工控机、L型相机支架和万向调节底座，所述工业相机固定在万向调节底座上，万向调节底座表面安装在L型相机架的一侧，另一侧与上下暗室顶部中间支撑架的中央固定，工业相机包括与外部电源连接的电源插口和与工控机相连的网线插口。
进一步的是：所述支撑架右侧为镂空，且上下暗室均设有所述照明装置和图像采集装置，用于可以接续检测玻璃的正反面。
本发明还公开了基于机器视觉的挡风玻璃划痕接续检测方法，步骤为：
第一步，通过触摸屏控制plc控制器，进而控制左右两侧第一直线导轨上的滑块带动第一小底板和第二小底板运动，使两个小底板同时纵向相对运动，运动到与玻璃宽度平齐时停止，第一小底板和第二小底板上的四个加紧板加紧玻璃的四个角之后，通过plc控制器控制与加紧板相连的滑块，使其向暗室方向以共同的速度移动；
第二步，当玻璃运动到即将进入暗室时，通过触摸屏控制暗室上下清洁装置的第三伺服电机使其推动第三滚珠丝杆伸缩，将滚轮下放到与玻璃表面刚好接触时停止，通过滚轮的旋转，使用其表面的清洁布清除表面灰尘，使后续检测更加精确，玻璃完全进入暗室之后，清洁装置回升，开始检测划痕，通过玻璃的宽度确定开启LED灯的数量，划痕检测采用背光照射LED灯的检测方式，当检测上表面玻璃划痕时，上方的照明装置以及下方暗室的工业相机处于关闭状态，上方的工业相机开启，工业相机通过万向底座调节其方向，以覆盖玻璃表面的整个检测区域，下方的LED灯通过电机带动下放至能将玻璃表面完全照射的位置，此时工业相机开始工作，将挡风玻璃图像传给与工业相机相连的工控机，并用工控机对划痕图像进行自动检测，玻璃上表面检测完成后，上方暗室的工业相机关闭下方暗室的开启，同时下方暗室的照明装置回升，上方暗室的LED灯通过电机带动下放至能将玻璃表面完全照射的位置，下方的工业相机开始工作；
第三步，将图像传给与工业相机相连的工控机进行图像处理分析，全部检测完成后，玻璃通过电机的带动退回左侧暗室外部，完成一次检测。
进一步的是：所述图像分析步骤为：
步骤一，图像平滑处理：图片在采集过程中会受到噪声的干扰，为了图像后期处理的准确性和高效性，首先将采集到的玻璃图像进行平滑处理，主要采用中值滤波的方式平滑图像，滤波窗口采用方形检测窗口，尺寸模板采用7×7模板，从而有效的抑制图像中的噪声；
步骤二，图像分割：图像分割是划痕检测的关键步骤，主要采用局部动态阈值分割方法，将缺陷信息全部提取出来，并转化为二值化图像；
步骤三，形态学处理：为了进一步获取划痕信息，再用膨胀形态学的方法处理划痕，使之在简化图像数据的同时保持它的基本形状特征，
步骤四，图像提取：为了还原缺陷形状，采用骨架提取的方法；
步骤五，将结果与标准特征进行对比判断该缺陷是否为划痕。
本发明的有益效果是：
1、本发明采用机器视觉的方式检测玻璃划痕，用机器检测代替人工操作，降低了劳动强度的同时提高了检测的精度。
2、本发明能够适应任何尺寸的玻璃检测，提高了玻璃检测的适用范围。
3、本发明采用上下双暗室，无需翻面接续检测玻璃的上下两面，提高了检测效率。
4、本发明带有清洁装置，能够在检测玻璃划痕之前清除玻璃表面的灰尘，使后续检测更加精确。
5、本发明LED灯采用单独控制的方式，根据玻璃的宽度和长度开启对应个数的LED灯，以防止不必要照明灯的浪费，节约了电力资源。
附图说明
图1为本发明的整体结构加暗室示意图；
图2为本发明的整体结构去暗室示意图；
图3为本发明的传送装置的立体结构示意图；
图4为本发明的清洁装置的立体结构示意图；
图5为本发明的照明装置的立体结构示意图；
图6为本发明的工业相机的立体结构示意图；
图7为本发明的挡风玻璃划痕图像检测流程图
附图标记：
1、传送装置；2、清洁装置；3、照明装置；4、图像采集处理装置；5、支撑架；6、触摸屏；7、plc控制器；11、第一电机；12、第一直线导轨；13第二直线导轨；14第二电机；110、大底板；111、第一小底板；112、加紧板；113第二小底板；21、转轴；22、十字支架；23、第三升降台；24、支撑杆；25、滚轮；26、滚轮轴承；27第三伺服电机，28第三滚珠丝杆；31、LED灯管；32、LED灯管架；41、L型相机支架；42、万向调节底座；43、电源插口；44、网线插口；45、工业相机。
具体实施方式
下面结合附图和具体实施方式对本发明进一步说明。
如图1和图2所示的基于机器视觉的挡风玻璃划痕接续检测装置，包括支撑架5、传送装置1，清洁装置2，照明装置3和图像采集处理装置4，所述支撑架5右侧上下部分为对称的暗室，所述传送装置1设置在支撑架5的前端，用于将不同尺寸的玻璃固定并送入暗室，所述清洁装置2安装在暗室左侧支撑架5的中央，用于在玻璃进入暗室时清除玻璃表面的灰尘，所述上下对称的暗室的顶部均设置有支撑杆24，所述支撑杆24上各安装有多个用来给玻璃提供照明的所述照明装置3，所述图像采集处理装置4固定在上下暗室顶部中间支撑架5的中央，用于拍摄玻璃表面的图像并进行处理分析，从而判定玻璃表面是否有划痕，所述支撑架5右侧为镂空，且上下暗室均设有所述照明装置3和图像采集装置，用于可以接续检测玻璃的正反面，本装置在工作时，传送装置1将待检测挡风玻璃向前输送，在输送过程中，清洁装置2先对其进行清洁操作，再将其送入暗室，上部暗室照明装置3开启，图像采集处理装置4进行图像采集处理，检测挡风玻璃上表面是否有划痕，上表面检测完成后，下部暗室的照明装置3开启，图像采集处理装置4进行图像采集处理，检测挡风玻璃下表面是否有划痕，从而实现两表面的检测，本装置采用机器视觉的方式检测玻璃划痕，用机器检测代替人工操作，降低了劳动强度的同时提高了检测的精度，同时本发明采用上下双暗室，无需翻面接续检测玻璃的上下两面，提高了检测效率，同时本发明通过清洁装置2的设置，能够在检测玻璃划痕之前清除玻璃表面的灰尘，使后续检测更加精确。
具体的，如图3所示，所述传送装置1可选为丝杆传送模组、气缸传送机构等，本实施例中的传送装置1包括大底板110，所述大底板110的上方固定有左右两个纵向设置的第一直线导轨12，左右两个第一直线导轨12上分别安装有上侧滑块和下侧滑块，左右两个第一直线导轨12的内侧并靠近第一直线导轨12的位置分别对应安装有左侧滚珠丝杆和右侧滚珠丝杆，还包括两个第一滚珠丝杆螺母与第一螺母连接座，所述两个第一滚珠丝杆螺母与第一螺母连接座相连并分别连接在左侧滚珠丝杆和右侧滚珠丝杆，用于对应控制左右两个第一直线导轨12上滑块的运动，第一滚珠丝杆一端安装第一固定轴承座，另一端安装第一支撑轴承座，用于固定滚珠丝杆，还包括第一电机11，所述第一电机11通过圆螺母安装在第一固定轴承座一端；
还包括横向安装的第一小底板111和第二小底板113，所述第一小底板111的下表面分别与左右第一直线导轨12的上侧滑块连接，第一小底板111的右侧下表面还与右侧滚珠丝杆的螺母连接座连接，第二小底板113的下表面分别与左右第一直线导轨12的下侧滑块连接，第二小底板113的左侧下表面还与左侧滚珠丝杆的螺母连接座连接，所述第一小底板111和第 二小底板113上面分别安装有第二直线导轨13，所述第二直线导轨13上安装有两个加紧板112，所述第二直线导轨13的两侧分别安装有第二滚珠丝杆，每个第二滚珠丝杆上都安装有一个第二螺母连接座，与第二滚珠丝杆螺母相连，第二滚珠丝杆的一侧安装第二固定轴承座，另一端安装第二支撑轴承座，用于固定第二滚珠丝杆，还包括第二电机14，所述第二电机14通过圆螺母安装在第二固定轴承座一端，每个加紧板112分别与对应的第二滚珠丝杆的第二螺母连接座连接，还包括plc控制器7和触摸屏6，所述第一电机11、第二电机14分别与plc控制器7相连，plc控制器7与触摸屏6连接，通过触摸屏6控制电机进而控制滑块和加紧板112的运动。
具体的，如图4所示，所述清洁装置2可选用吸尘装置等，本实施例的清洁装置2包括第三伺服电机27，第三滚珠丝杆28，第三升降台23，滚轮25，滚轮轴承26和支撑杆24，所述第三升降台23的上表面安装在上下暗室的顶部，并且第三升降台23的一侧与暗室顶部支撑杆24中央重合，第三升降台23上表面内侧安装有第三伺服电机27，第三伺服电机27与第三滚珠丝杆28相连，第三滚珠丝杆28杠与第三升降台23的转轴21通过十字支架22连接，第三升降台23的下表面与支撑杆24相连接，支撑杆24与滚轮轴承26相连，滚轮25固定在滚轮轴承26的外表面，且滚轮25的表面附着有用来擦除玻璃表面的灰尘玻璃清洁布，还包括plc控制器7和触摸屏6，所述第三伺服电机27与plc控制器7相连，plc控制器7再与触摸屏6连接，通过触摸屏6控制第三伺服电机27，第三伺服电机27控制滚珠丝杆的伸缩，进而控制第三升降台23上下移动。
具体的，如图5所示，所述照明装置3包括LED灯管31，第四伺服电机，第四滚珠丝杆，第四升降台，LED灯管架32，所述LED灯管架32安装在升降台的下表面，LED灯管31与灯管架连接，升降台上表面内侧安装有第四伺服电机，第四伺服电机与第四滚珠丝杆相连，第四滚珠丝杆与第四升降台的转轴21通过十字支架22连接，上下暗室的顶部各安装有多个各自独立控制的所述照明装置3，多个照明装置3安装在暗室顶部中间支撑杆24两侧靠近中央的位置，并线性排列，还包括plc控制器7和触摸屏6，所述第四伺服电机与plc控制器7相连，plc控制器7再与触摸屏6连接，通过触摸屏6控制第四伺服电机，第四伺服电机控制第四滚珠丝杆的伸缩，进而控制第四升降台上下移动。
具体的，如图6所示，所述的图像采集处理装置4包括工业相机45、工控机、L型相机支架41和万向调节底座42，所述工业相机45固定在万向调节底座42上，万向调节底座42表面安装在L型相机架的一侧，另一侧与上下暗室顶部中间支撑架5的中央固定，工业相机45包括与外部电源连接的电源插口43和与工控机相连的网线插口44。
具体工作流程为：
第一步，通过触摸屏6控制plc控制器7，进而控制左右两侧第一直线导轨12上的滑块带动第一小底板111和第二小底板113运动，使两个小底板同时纵向相对运动，运动到与玻璃宽度平齐时停止，第一小底板111和第二小底板113上的四个加紧板112加紧玻璃的四个角之后，通过plc控制器7控制与加紧板112相连的滑块，使其向暗室方向以共同的速度移动；
第二步，当玻璃运动到即将进入暗室时，通过触摸屏6控制暗室上下清洁装置2的第三伺服电机27使其推动第三滚珠丝杆28伸缩，将滚轮25下放到与玻璃表面刚好接触时停止，通过滚轮25的旋转，使用其表面的清洁布清除表面灰尘，使后续检测更加精确，玻璃完全进入暗室之后，清洁装置2回升，开始检测划痕，通过玻璃的宽度确定开启LED灯的数量，划痕检测采用背光照射LED灯的检测方式，当检测上表面玻璃划痕时，上方的照明装置3以及下方暗室的工业相机45处于关闭状态，上方的工业相机45开启，工业相机45通过万向底座调节其方向，以覆盖玻璃表面的整个检测区域，下方的LED灯通过电机带动下放至能将玻璃表面完全照射的位置，此时工业相机45开始工作，将挡风玻璃图像传给与工业相机45相连的工控机，并用工控机对划痕图像进行自动检测，玻璃上表面检测完成后，上方暗室的工业相机45关闭下方暗室的开启，同时下方暗室的照明装置3回升，上方暗室的LED灯通过电机带动下放至能将玻璃表面完全照射的位置，下方的工业相机45开始工作；
第三步，将图像传给与工业相机45相连的工控机进行图像处理分析，全部检测完成后，玻璃通过电机的带动退回左侧暗室外部，完成一次检测。
上述具体图像处理分析步骤为：
步骤一，图像平滑处理：图片在采集过程中会受到噪声的干扰，为了图像后期处理的准确性和高效性，首先将采集到的玻璃图像进行平滑处理，主要采用中值滤波的方式平滑图像，滤波窗口采用方形检测窗口，尺寸模板采用7×7模板，从而有效的抑制图像中的噪声；
步骤二，图像分割：图像分割是划痕检测的关键步骤，主要采用局部动态阈值分割方法，将缺陷信息全部提取出来，并转化为二值化图像；
步骤三，形态学处理：为了进一步获取划痕信息，再用膨胀形态学的方法处理划痕，使之在简化图像数据的同时保持它的基本形状特征，
步骤四，图像提取：为了还原缺陷形状，采用骨架提取的方法；
步骤五，将结果与标准特征进行对比判断该缺陷是否为划痕。
以上所述的具体实施例，对本发明的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本发明具体实施例而已，并不用于限制本发明，凡在本发明的精神和原则之内，所做的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (8)

  1. 基于机器视觉的挡风玻璃划痕接续检测装置，包括支撑架(5)、传送装置(1)，清洁装置(2)，照明装置(3)和图像采集处理装置(4)，其特征在于：所述支撑架(5)右侧上下部分为对称的暗室，所述传送装置(1)设置在支撑架(5)的前端，用于将不同尺寸的玻璃固定并送入暗室，所述清洁装置(2)安装在暗室左侧支撑架(5)的中央，用于在玻璃进入暗室时清除玻璃表面的灰尘，所述上下对称的暗室的顶部均设置有支撑杆(24)，所述支撑杆(24)上各安装有多个用来给玻璃提供照明的所述照明装置(3)，所述图像采集处理装置(4)固定在上下暗室顶部中间支撑架(5)的中央，用于拍摄玻璃表面的图像并进行处理分析，从而判定玻璃表面是否有划痕。
  2. 根据权利要求1所述的基于机器视觉的挡风玻璃划痕接续检测装置，其特征在于：所述传送装置(1)包括大底板(110)，所述大底板(110)的上方固定有左右两个纵向设置的第一直线导轨(12)，左右两个第一直线导轨(12)上分别安装有上侧滑块和下侧滑块，左右两个第一直线导轨(12)的内侧并靠近第一直线导轨(12)的位置分别对应安装有左侧滚珠丝杆和右侧滚珠丝杆，还包括两个第一滚珠丝杆螺母与第一螺母连接座，所述两个第一滚珠丝杆螺母与第一螺母连接座相连并分别连接在左侧滚珠丝杆和右侧滚珠丝杆，用于对应控制左右两个第一直线导轨(12)上滑块的运动，第一滚珠丝杆一端安装第一固定轴承座，另一端安装第一支撑轴承座，用于固定滚珠丝杆，还包括第一电机(11)，所述第一电机(11)通过圆螺母安装在第一固定轴承座一端；
    还包括横向安装的第一小底板(111)和第二小底板(113)，所述第一小底板(111)的下表面分别与左右第一直线导轨(12)的上侧滑块连接，第一小底板(111)的右侧下表面还与右侧滚珠丝杆的螺母连接座连接，第二小底板(113)的下表面分别与左右第一直线导轨(12)的下侧滑块连接，第二小底板(113)的左侧下表面还与左侧滚珠丝杆的螺母连接座连接，所述第一小底板(111)和第二小底板(113)上面分别安装有第二直线导轨(13)，所述第二直线导轨(13)上安装有两个加紧板(112)，所述第二直线导轨(13)的两侧分别安装有第二滚珠丝杆，每个第二滚珠丝杆上都安装有一个第二螺母连接座，与第二滚珠丝杆螺母相连，第二滚珠丝杆的一侧安装第二固定轴承座，另一端安装第二支撑轴承座，用于固定第二滚珠丝杆，还包括第二电机(14)，所述第二电机(14)通过圆螺母安装在第二固定轴承座一端，每个加紧板(112)分别与对应的第二滚珠丝杆的第二螺母连接座连接，还包括plc控制器(7)和触摸屏(6)，所述第一电机(11)、第二电机(14)分别与plc控制器(7)相连，plc控制器(7)与触摸屏(6)连接，通过触摸屏(6)控制电机进而控制滑块和加紧板(112)的运 动。
  3. 根据权利要求1所述的基于机器视觉的挡风玻璃划痕接续检测装置，其特征在于：所述清洁装置(2)包括第三伺服电机(27)，第三滚珠丝杆(28)，第三升降台(23)，滚轮(25)，滚轮轴承(26)和支撑杆(24)，所述第三升降台(23)的上表面安装在上下暗室的顶部，并且第三升降台(23)的一侧与暗室顶部支撑杆(24)中央重合，第三升降台(23)上表面内侧安装有第三伺服电机(27)，第三伺服电机(27)与第三滚珠丝杆(28)相连，第三滚珠丝杆(28)杠与第三升降台(23)的转轴(21)通过十字支架(22)连接，第三升降台(23)的下表面与支撑杆(24)相连接，支撑杆(24)与滚轮轴承(26)相连，滚轮(25)固定在滚轮轴承(26)的外表面，且滚轮(25)的表面附着有用来擦除玻璃表面的灰尘玻璃清洁布，还包括plc控制器(7)和触摸屏(6)，所述第三伺服电机(27)与plc控制器(7)相连，plc控制器(7)再与触摸屏(6)连接，通过触摸屏(6)控制第三伺服电机(27)，第三伺服电机(27)控制滚珠丝杆的伸缩，进而控制第三升降台(23)上下移动。
  4. 根据权利要求1所述的基于机器视觉的挡风玻璃划痕接续检测装置，其特征在于：所述的照明装置(3)包括LED灯管(31)，第四伺服电机，第四滚珠丝杆，第四升降台，LED灯管架(32)，所述LED灯管架(32)安装在升降台的下表面，LED灯管(31)与灯管架连接，升降台上表面内侧安装有第四伺服电机，第四伺服电机与第四滚珠丝杆相连，第四滚珠丝杆与第四升降台的转轴(21)通过十字支架(22)连接，上下暗室的顶部各安装有多个各自独立控制的所述照明装置(3)，多个照明装置(3)安装在暗室顶部中间支撑杆(24)两侧靠近中央的位置，并线性排列，还包括plc控制器(7)和触摸屏(6)，所述第四伺服电机与plc控制器(7)相连，plc控制器(7)再与触摸屏(6)连接，通过触摸屏(6)控制第四伺服电机，第四伺服电机控制第四滚珠丝杆的伸缩，进而控制第四升降台上下移动。
  5. 根据权利要求1所述的基于机器视觉的挡风玻璃划痕接续检测装置，其特征在于：所述的图像采集处理装置(4)包括工业相机(45)、工控机、L型相机支架(41)和万向调节底座(42)，所述工业相机(45)固定在万向调节底座(42)上，万向调节底座(42)表面安装在L型相机架的一侧，另一侧与上下暗室顶部中间支撑架(5)的中央固定，工业相机(45)包括与外部电源连接的电源插口(43)和与工控机相连的网线插口(44)。
  6. 根据权利要求1所述基于机器视觉的挡风玻璃划痕接续检测装置，其特征在于：所述支撑架(5)右侧为镂空，且上下暗室均设有所述照明装置(3)和图像采集装置，用于可以接续检测玻璃的正反面。
  7. 基于机器视觉的挡风玻璃划痕接续检测方法，其特征在于：步骤为：
    第一步，通过触摸屏(6)控制plc控制器(7)，进而控制左右两侧第一直线导轨(12) 上的滑块带动第一小底板(111)和第二小底板(113)运动，使两个小底板同时纵向相对运动，运动到与玻璃宽度平齐时停止，第一小底板(111)和第二小底板(113)上的四个加紧板(112)加紧玻璃的四个角之后，通过plc控制器(7)控制与加紧板(112)相连的滑块，使其向暗室方向以共同的速度移动；
    第二步，当玻璃运动到即将进入暗室时，通过触摸屏(6)控制暗室上下清洁装置(2)的第三伺服电机(27)使其推动第三滚珠丝杆(28)伸缩，将滚轮(25)下放到与玻璃表面刚好接触时停止，通过滚轮(25)的旋转，使用其表面的清洁布清除表面灰尘，玻璃完全进入暗室之后，清洁装置(2)回升，开始检测划痕，通过玻璃的宽度确定开启LED灯的数量，划痕检测采用背光照射LED灯的检测方式，当检测上表面玻璃划痕时，上方的照明装置(3)以及下方暗室的工业相机(45)处于关闭状态，上方的工业相机(45)开启，工业相机(45)通过万向底座调节其方向，以覆盖玻璃表面的整个检测区域，下方的LED灯通过电机带动下放至能将玻璃表面完全照射的位置，此时工业相机(45)开始工作，将挡风玻璃图像传给与工业相机(45)相连的工控机，并用工控机对划痕图像进行自动检测，玻璃上表面检测完成后，上方暗室的工业相机(45)关闭下方暗室的开启，同时下方暗室的照明装置(3)回升，上方暗室的LED灯通过电机带动下放至能将玻璃表面完全照射的位置，下方的工业相机(45)开始工作；
    第三步，将图像传给与工业相机(45)相连的工控机进行图像处理分析，全部检测完成后，玻璃通过电机的带动退回左侧暗室外部，完成一次检测。
  8. 根据权利要求7所述的基于机器视觉的挡风玻璃划痕接续检测方法，其特征在于，所述图像分析步骤为：
    步骤一，图像平滑处理：图片在采集过程中会受到噪声的干扰，为了图像后期处理的准确性和高效性，首先将采集到的玻璃图像进行平滑处理；
    步骤二，图像分割：图像分割是划痕检测的关键步骤，主要采用局部动态阈值分割方法，将缺陷信息全部提取出来，并转化为二值化图像；
    步骤三，形态学处理：为了进一步获取划痕信息，再用膨胀形态学的方法处理划痕，使之在简化图像数据的同时保持它的基本形状特征，
    步骤四，图像提取：为了还原缺陷形状，采用骨架提取的方法；
    步骤五，将结果与标准特征进行对比判断该缺陷是否为划痕。
PCT/CN2020/131203 2020-03-13 2020-11-24 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法 WO2021179680A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010173120.7 2020-03-13
CN202010173120.7A CN111239163A (zh) 2020-03-13 2020-03-13 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法

Publications (1)

Publication Number Publication Date
WO2021179680A1 true WO2021179680A1 (zh) 2021-09-16

Family

ID=70870263

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/131203 WO2021179680A1 (zh) 2020-03-13 2020-11-24 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法

Country Status (2)

Country Link
CN (1) CN111239163A (zh)
WO (1) WO2021179680A1 (zh)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113639802A (zh) * 2021-10-13 2021-11-12 江苏立晶工业科技有限公司 一种汽车前挡风玻璃的检测装置及方法
CN113899782A (zh) * 2021-09-30 2022-01-07 苏州市鼎诚钢化玻璃有限公司 一种真空玻璃隔热性能测试装置及测试方法
CN114018945A (zh) * 2022-01-04 2022-02-08 中机智能装备创新研究院(宁波)有限公司 一种锻造斗齿的检测装置以及检测方法
CN114235845A (zh) * 2021-11-19 2022-03-25 东莞理工学院 用于注塑件产品缺陷检测的机构、系统及方法
CN114414593A (zh) * 2022-03-31 2022-04-29 欧普康视科技股份有限公司 一种光学镜片投影检测装置
CN114858813A (zh) * 2022-06-07 2022-08-05 浙江理工大学 一种毛纺纱线生产用在线质量检测智能装置
CN115712210A (zh) * 2022-11-10 2023-02-24 江西夺胜电子科技有限公司 Lcd自动贴片生产线及其生产方法
CN117470766A (zh) * 2023-11-01 2024-01-30 上海中浦电磁科技有限公司 一种自动定位数据采集装置
CN117900155A (zh) * 2024-01-02 2024-04-19 劲佳科技有限公司 一种包装盒的防伪印刷检测装置
CN118253557A (zh) * 2024-05-30 2024-06-28 昆山鸿仕达智能科技股份有限公司 一种手机摄像头玻璃盖板侧壁镀层全自动浮动刮削机构

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111239163A (zh) * 2020-03-13 2020-06-05 苏州鑫睿益荣信息技术有限公司 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法
CN111912848A (zh) * 2020-08-13 2020-11-10 蚌埠中光电科技有限公司 一种高世代大尺寸玻璃基板的人工复检装置
CN112730252A (zh) * 2020-12-30 2021-04-30 湖南三安半导体有限责任公司 晶圆检测装置
CN112816475B (zh) * 2021-01-13 2023-06-16 太原市玉磊预拌混凝土有限公司 一种基于图像扫描的天然砂含泥量检测装置
CN117346843B (zh) * 2023-12-04 2024-02-20 江苏南晶红外光学仪器有限公司 一种镀膜玻璃检测设备及其检测方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2016114481A (ja) * 2014-12-16 2016-06-23 旭硝子株式会社 評価装置及び評価方法
CN105842256A (zh) * 2015-01-30 2016-08-10 哈里斯股份有限公司 透明板检查装置及透明板清扫检查系统
CN205593939U (zh) * 2016-03-25 2016-09-21 丹阳市精通眼镜技术创新服务中心有限公司 一种镜片面状疵病自动检测装置
CN207036721U (zh) * 2017-08-16 2018-02-23 福耀集团（上海）汽车玻璃有限公司 玻璃边缘检测装置及玻璃加工线
CN208547589U (zh) * 2018-07-30 2019-02-26 中山市光维智能科技有限公司 应用于手机光学镜头的分层自动检测装置
CN110567989A (zh) * 2019-09-04 2019-12-13 北京信息科技大学 屏蔽玻璃缺陷检测方法
CN111239163A (zh) * 2020-03-13 2020-06-05 苏州鑫睿益荣信息技术有限公司 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法
CN211856405U (zh) * 2020-03-13 2020-11-03 苏州鑫睿益荣信息技术有限公司 基于机器视觉的挡风玻璃划痕接续检测装置

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7551274B1 (en) * 2007-02-28 2009-06-23 Lite Sentry Corporation Defect detection lighting system and methods for large glass sheets
CN101718714B (zh) * 2009-11-25 2012-07-11 东旭集团有限公司 一种检测平板玻璃表面缺陷的系统及方法
CN203534984U (zh) * 2013-09-30 2014-04-09 福耀集团长春有限公司 一种轿车前挡风玻璃在线检光学畸变装置
CN103837552A (zh) * 2014-03-14 2014-06-04 苏州精创光学仪器有限公司 触摸屏保护玻璃外观缺陷检测系统
CN107064173A (zh) * 2017-01-03 2017-08-18 中国科学院上海光学精密机械研究所 大型平面光学元件表面疵病的检测装置和检测方法
FR3075374A1 (fr) * 2017-12-19 2019-06-21 Saint-Gobain Glass France Systeme d'inspection de vitrage ameliore
CN210119453U (zh) * 2019-03-27 2020-02-28 爱丁堡（南京）光电设备有限公司 一种光滑表面缺陷的检测装置
CN110290257B (zh) * 2019-07-09 2020-07-07 北京邮电大学 手机后盖视觉检测生产线
CN110487821A (zh) * 2019-08-30 2019-11-22 太原科技大学 一种基于机器视觉的透明板材缺陷检测装置及其检测方法
CN110756488A (zh) * 2019-10-30 2020-02-07 安徽省六二八光电科技有限公司 一种玻璃检测清洁装置

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2016114481A (ja) * 2014-12-16 2016-06-23 旭硝子株式会社 評価装置及び評価方法
CN105842256A (zh) * 2015-01-30 2016-08-10 哈里斯股份有限公司 透明板检查装置及透明板清扫检查系统
CN205593939U (zh) * 2016-03-25 2016-09-21 丹阳市精通眼镜技术创新服务中心有限公司 一种镜片面状疵病自动检测装置
CN207036721U (zh) * 2017-08-16 2018-02-23 福耀集团（上海）汽车玻璃有限公司 玻璃边缘检测装置及玻璃加工线
CN208547589U (zh) * 2018-07-30 2019-02-26 中山市光维智能科技有限公司 应用于手机光学镜头的分层自动检测装置
CN110567989A (zh) * 2019-09-04 2019-12-13 北京信息科技大学 屏蔽玻璃缺陷检测方法
CN111239163A (zh) * 2020-03-13 2020-06-05 苏州鑫睿益荣信息技术有限公司 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法
CN211856405U (zh) * 2020-03-13 2020-11-03 苏州鑫睿益荣信息技术有限公司 基于机器视觉的挡风玻璃划痕接续检测装置

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113899782B (zh) * 2021-09-30 2024-04-30 苏州市鼎诚钢化玻璃有限公司 一种真空玻璃隔热性能测试装置及测试方法
CN113899782A (zh) * 2021-09-30 2022-01-07 苏州市鼎诚钢化玻璃有限公司 一种真空玻璃隔热性能测试装置及测试方法
CN113639802B (zh) * 2021-10-13 2022-01-18 江苏立晶工业科技有限公司 一种汽车前挡风玻璃的检测装置及方法
CN113639802A (zh) * 2021-10-13 2021-11-12 江苏立晶工业科技有限公司 一种汽车前挡风玻璃的检测装置及方法
CN114235845A (zh) * 2021-11-19 2022-03-25 东莞理工学院 用于注塑件产品缺陷检测的机构、系统及方法
CN114018945A (zh) * 2022-01-04 2022-02-08 中机智能装备创新研究院(宁波)有限公司 一种锻造斗齿的检测装置以及检测方法
CN114018945B (zh) * 2022-01-04 2022-04-19 中机智能装备创新研究院(宁波)有限公司 一种锻造斗齿的检测装置以及检测方法
CN114414593A (zh) * 2022-03-31 2022-04-29 欧普康视科技股份有限公司 一种光学镜片投影检测装置
CN114858813A (zh) * 2022-06-07 2022-08-05 浙江理工大学 一种毛纺纱线生产用在线质量检测智能装置
CN115712210A (zh) * 2022-11-10 2023-02-24 江西夺胜电子科技有限公司 Lcd自动贴片生产线及其生产方法
CN115712210B (zh) * 2022-11-10 2024-05-28 江西夺胜电子科技有限公司 Lcd自动贴片生产线及其生产方法
CN117470766A (zh) * 2023-11-01 2024-01-30 上海中浦电磁科技有限公司 一种自动定位数据采集装置
CN117470766B (zh) * 2023-11-01 2024-05-28 上海中浦电磁科技有限公司 一种自动定位数据采集装置
CN117900155A (zh) * 2024-01-02 2024-04-19 劲佳科技有限公司 一种包装盒的防伪印刷检测装置
CN118253557A (zh) * 2024-05-30 2024-06-28 昆山鸿仕达智能科技股份有限公司 一种手机摄像头玻璃盖板侧壁镀层全自动浮动刮削机构

Also Published As

Publication number Publication date
CN111239163A (zh) 2020-06-05

Similar Documents

Publication Publication Date Title
WO2021179680A1 (zh) 基于机器视觉的挡风玻璃划痕接续检测装置和检测方法
CN211978714U (zh) 一种汽车漆面瑕疵自动检测装置
CN209736110U (zh) 一种ccd视觉检测设备
CN211426316U (zh) Aoi检测设备
CN109596633B (zh) 导光板缺陷检测设备及检测方法
CN103837552A (zh) 触摸屏保护玻璃外观缺陷检测系统
CN211856405U (zh) 基于机器视觉的挡风玻璃划痕接续检测装置
CN203745392U (zh) 触摸屏保护玻璃外观缺陷检测系统
CN112345553A (zh) 一种硬盘零件检测装置及检测方法
CN115452848A (zh) 一种人造板表面瑕疵自动检测设备
CN204789365U (zh) 用于pcb菲林片检测的aoi装置的图像采集装置
CN212301350U (zh) 一种用于钢管内壁涂层缺陷检测的光学成像显示系统
CN219326945U (zh) 一种无纺布瑕疵检测收卷一体装置
CN107884417A (zh) 一种基板边缘检查机及边缘检查方法
CN204790307U (zh) 用于pcb菲林片检测的aoi装置
CN117007601A (zh) 一种导光板缺陷检测装置及方法
CN108918538B (zh) 一种反光面板ccd检测设备
CN216525494U (zh) 一种关于显示屏双面及四周外观缺陷aoi检测的装置
CN116359228A (zh) 一种单晶硅棒表面瑕疵检测装置及检测系统
CN212031329U (zh) 一种平板玻璃瑕疵的抽检装置
CN110849887B (zh) 一种钢管内壁涂层缺陷检测装置及其检测方法
CN211318242U (zh) 电脑液晶屏导光板留白视觉检测装置
CN110947635B (zh) 一种轴承外圈表面缺陷自动检测装置
CN219179250U (zh) 一种pcb线路板检测装置
CN218885767U (zh) 一种模组产品外观缺陷的光学检测装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20923967

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20923967

Country of ref document: EP

Kind code of ref document: A1