Vehicle body paint defect detection device and detection method thereof

Abstract

The present application relates to a kind of vehicle body paint defects detection device and its detection method, including servo controller, industrial computer, vehicle body conveying mechanism, several area array cameras, several projectors and several projection screens, the industrial computer is connected with the servo controller, the servo controller is connected with the vehicle body conveying mechanism, the vehicle body conveying mechanism can be driven under the action of the servo controller and industrial computer and be moved in length direction and height direction for vehicle to be detected, the area array camera is distributed around vehicle to be detected, the industrial computer is connected with the projector, the projection screen is set in the periphery of the projector and corresponds with the projector, the vehicle body paint defects detection device and its detection method, collection efficiency is high, imaging quality is high, system structure is compact, and cost is low.

Description

Technical field:

[0001] The present invention relates to the field of defect detection and machine vision technology, and in particular to a vehicle body paint surface defect detection device and a detection method thereof. Background technology:

[0002] Currently, paint defect detection for automotive bodies primarily relies on traditional manual visual inspection. However, due to low efficiency, lack of objective testing standards, and susceptibility to subjective factors such as human distraction and fatigue, this method is increasingly unable to meet the measurement and inspection requirements of the process. Consequently, there is a growing demand for automated defect detection devices that can not only strictly control product quality but also promptly trace defects to their source, providing data support for process quality improvement.

[0003] Car body paint defects vary widely, and different manufacturers define them differently. Based on their optical imaging, defects can be categorized into: chromatic aberration, dirt, texture, scratches, and bumps. Single 2D imaging and detection methods are inadequate for common defects. Simultaneous detection of all defects often requires a combination of 2D and 3D imaging. Among 3D imaging methods, laser triangulation and fringe projection reconstruct height. 3D reconstruction equipment based on fringe projection is primarily used for diffusely reflective objects. Laser triangulation can be used for height measurement of mirror-like objects, but it struggles with detecting micron-level defects. Among 3D imaging methods, photometric stereo and fringe reflectometry (phase deflectometry) reconstruct gradients. Photometric stereo, based on the Lambertian illumination model, achieves high gradient reconstruction accuracy for diffusely reflective surfaces, but is difficult to directly apply to mirror-like objects. Phase deflectometry reconstructs gradients for mirror-like objects with high accuracy, in principle reaching submicron levels. Summary of the invention:

[0004] In response to the above problems, the present invention proposes a vehicle body paint defect detection device and a detection method suitable for mirror defect detection on large and complex curved surfaces, which has high acquisition efficiency, high imaging quality, compact system structure and low cost, while ensuring micron-level detection accuracy and taking into account factors such as detection efficiency, system cost and compact structure.

[0005] The present invention is achieved through the following technical solution: a vehicle body paint defect detection device, including a servo controller, an industrial computer, a vehicle body conveying mechanism, several area array cameras, several projectors and several projection screens. The posture relationship between the coordinate system of the above components and the world coordinate system is calibrated in advance through a calibration plate.

[0006] The area array camera, projector, and projection screen together constitute an image acquisition system for collecting and analyzing surface data of the vehicle to be inspected. The industrial computer is connected to the projector, and the projection screen is arranged outside the projector and corresponds to the projector. The projector can project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto the corresponding projection screen. The area array cameras are distributed around the vehicle to be inspected, and each area array camera photographs the stripes on the projection screen through reflection from the vehicle body surface. Within a single field of view of the vehicle body, four sinusoidal phase-shifted fringe images, four horizontal and four vertical, can be captured, for a total of eight sinusoidal phase-shifted fringe images.

[0007] In order to improve the image acquisition efficiency and imaging quality and achieve the best image acquisition and data analysis effect, preferably, there are 8 projectors in total, of which 2 are set at the front position of the vehicle to be detected, 2 are set at the rear position of the vehicle to be detected, and 4 are set at the side position and roof position of the vehicle to be detected. The projection screen includes 2 parallel vertical screens parallel to the XOZ plane and parallel to the body of the vehicle to be detected, 4 vertical deflected vertical screens arranged at 40° to 50° with the XOZ plane and arranged around the body of the vehicle to be detected, and 2 deflected vertical screens arranged at the side of the vehicle to be detected. A top projection screen is measured above the vehicle body and at an angle of 100° to 110° to each other. The projection direction of the projector corresponding to the parallel vertical screen is an inclined projection facing the positive direction of the X-axis under the XOY viewing angle, and the angle between the projection direction and the positive direction of the X-axis is 40° to 50°. The projection direction of the projector corresponding to the deflected vertical screen is perpendicular to the deflected vertical screen. The projection direction of the projector corresponding to the top projection screen is an inclined projection facing the positive direction of the X-axis under the XOZ viewing angle, and the angle between the projection direction and the positive direction of the X-axis is 40° to 50°.

[0008] In order to reduce costs while ensuring the use effect, it is preferred that the projector is a DLP projector.

[0009] The resolution and arrangement quantity of the cameras are related to the minimum defect detection size requirement. The minimum defect detection size involved in the present invention is 1 mm, and the pixel resolution of the image is set to be about 0.2 mm / pixel. Therefore, preferably, the area array camera has 5 million pixels and a 2 / 3 target surface. The number of the area array cameras is 18, of which 4 are correspondingly arranged at the front position of the vehicle to be inspected for detecting the hood and the front face of the vehicle, 4 are correspondingly arranged at the rear position of the vehicle to be inspected for detecting the trunk lid and the rear of the vehicle, and 10 are correspondingly arranged at the side position and roof position of the vehicle to be inspected for detecting the hood, roof, and trunk lid.

[0010] The servo controller, industrial computer and vehicle body conveying mechanism are used to cooperate with the image acquisition system to drive the vehicle to be inspected to move to the image acquisition point. The industrial computer is connected to the servo controller, and the servo controller is connected to the vehicle body conveying mechanism. The vehicle body conveying mechanism can drive the vehicle to be inspected to move in the length direction and height direction under the action of the servo controller and the industrial computer.

[0011] The present invention also includes a method for detecting vehicle body paint defects, which uses the above-mentioned vehicle body paint defect detection device and includes the following steps:

[0012] 1) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the hood inspection section, and then the vehicle body conveying mechanism drives the vehicle to be inspected upward along the Z-axis until the hood of the vehicle to be inspected reaches the hood inspection height. The eight projectors sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens. The area array cameras located at the front and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringes on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0013] 2) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto corresponding projection screens. The area array cameras located at the front and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0014] 3) Repeat step 2) until the engine hood is completely mapped.

[0015] 4) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the roof inspection section, and then the vehicle body conveying mechanism drives the vehicle to be inspected to move downward along the Z-axis until the roof of the vehicle to be inspected reaches the roof inspection height and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens. The area array cameras located on the side and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringes on the projection screen through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0016] 5) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto corresponding projection screens. The area array cameras located on the side and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0017] 6) Repeat step 5) until the roof is mapped.

[0018] 7) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the front face inspection section. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras positioned at the front of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0019] 8) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto corresponding projection screens. The area array cameras positioned at the front of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0020] 9) Repeat step 8) until the front of the vehicle is captured.

[0021] 10) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the trunk lid inspection section, and then the vehicle body conveying mechanism drives the vehicle to be inspected to move upward along the Z-axis so that the trunk lid of the vehicle to be inspected reaches the trunk lid inspection height and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens. The area array cameras located at the side and rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringes on the projection screen through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0022] 11) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto corresponding projection screens. The area array cameras positioned at the side and rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0023] 12) Repeat step 11) until the trunk lid is completely captured.

[0024] 13) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the vehicle rear inspection section. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras located at the rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0025] 14) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto corresponding projection screens. The area array camera located at the rear of the vehicle to be inspected sequentially captures images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0026] 15) Repeat step 14) until the rear end of the vehicle is completely captured;

[0027] 16) Use the calibration information to stitch the collected sinusoidal phase-shifted fringe images, and use the principle of phase measurement deflectometry to process the sinusoidal phase-shifted fringe images to obtain curvature maps, diffuse reflection maps, specular reflection maps, and glossiness maps. Then, through traditional image processing algorithms and deep learning algorithms, detect dirt defects, texture defects, scratch and bump defects, and bump defects.

[0028] Among them, the calculation process of the curvature map is to calculate the first-order derivatives of the lateral wrapping phase and the vertical wrapping phase respectively to obtain the lateral curvature and vertical curvature of the surface, and then calculate the square sum of the two to obtain the curvature map; the calculation process of the glossiness map is to average the 8 stripe images to obtain the glossiness map; the diffuse reflection map is extracted by taking the smaller comparison with the glossiness map, and the mirror reflection map is extracted by taking the larger comparison with the glossiness map, and the bright imaging part of the stripe.

[0029] It should be noted that while capturing images of the roof, trunk lid, and other positions, the image capture of the side of the vehicle body has been completed at the same time, so there is no need to repeatedly capture images of the side of the vehicle body separately.

[0030] The beneficial effects of the present invention are as follows: the vehicle body paint defect detection device and detection method, by designing a servo controller, an industrial computer and a vehicle body conveying mechanism for use in conjunction with an image acquisition system, can conveniently and efficiently acquire sinusoidal phase-shifted fringe images related to the vehicle body surface, thereby outputting mirror reflection images, diffuse reflection images, glossiness images, curvature images, etc., and then can detect dirt defects, texture defects, scratch and bump defects, and concave and convex defects on the vehicle body paint surface through traditional image processing algorithms and deep learning algorithms; the use of projection to replace LCD screens and LED light sources has low hardware costs and avoids the customization of large LCD screens and LED light sources; the detection accuracy is high, and it can be controlled by adding or removing array cameras to achieve micron-level defect detection accuracy; it has strong adaptability, and the vehicle body stop-and-go image acquisition and splicing method can adapt to different vehicle models; at the same time, it is also suitable for defect detection of mirror objects with large and complex curved surfaces in other industries. Description of the drawings:

[0031] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0032] Figure 2 This is a schematic diagram of the engine hood inspection section in the present invention;

[0033] Figure 3 This is a schematic diagram of the roof inspection section in the present invention;

[0034] Figure 4 This is a schematic diagram of the vehicle front face detection section in the present invention;

[0035] Figure 5 This is a schematic diagram of the trunk lid inspection section in the present invention;

[0036] Figure 6 This is a schematic diagram of the vehicle rear detection section in the present invention;

[0037] Figure 7 This is a schematic diagram of the four-step phase shift principle in the present invention;

[0038] Figure 8 Schematic diagram of the deflectometry imaging principle of the present invention;

[0039] Figure 9 Schematic diagram of the calculation process of the curvature map in the present invention;

[0040] Figure 10 Schematic diagram of the synthetic image in the present invention. Specific implementation method:

[0041] The preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more precise definition of the protection scope of the present invention.

[0042] like Figure 1 The device for detecting defects in vehicle paint surfaces shown in the figure includes a servo controller 1, an industrial computer 2, a vehicle body conveying mechanism 3, 18 area array cameras 4 with 5 million pixels and a 2 / 3 target surface, 8 projectors 5 and 8 projection screens 6. The projectors 5 are DLP projectors. The industrial computer 2 is connected to the servo controller 1, and the servo controller 1 is connected to the vehicle body conveying mechanism 3. The vehicle body conveying mechanism 3 can drive the vehicle to be inspected 9 to move in the length and height directions under the action of the servo controller 1 and the industrial computer 2. The industrial computer 2 is connected to the projectors 5. The projection screens 6 are arranged outside the projectors 5 and correspond to the projectors 5. The area array cameras 4 are distributed around the vehicle to be inspected 9.

[0043] There are a total of 8 projectors 5, of which 2 are set at the front position of the vehicle to be detected 9, 2 are set at the rear position of the vehicle to be detected 9, and 4 are set at the side position and roof position of the vehicle to be detected 9. The projection screen 6 includes 2 parallel vertical screens 61 parallel to the XOZ plane and parallel to the body of the vehicle to be detected 9, 4 vertical deflected vertical screens 62 arranged around the body of the vehicle to be detected 9 at an angle of 40° to 50° to the XOZ plane, and 2 set above the body of the vehicle to be detected 9 and at an angle of 100° to 11 The top projection screen 63 is 0°, the projection direction of the projector 5 corresponding to the parallel vertical screen 61 is an inclined projection facing the positive direction of the X-axis at an XOY viewing angle, and the angle between the projection direction and the positive direction of the X-axis is 40°~50°, the projection direction of the projector 5 corresponding to the deflected vertical screen 62 is perpendicular to the deflected vertical screen 62, and the projection direction of the projector 5 corresponding to the top projection screen 63 is an inclined projection facing the positive direction of the X-axis at an XOZ viewing angle, and the angle between the projection direction and the positive direction of the X-axis is 40°~50°.

[0044] The number of the area array cameras 4 is 18, of which 4 are arranged at the front of the vehicle 9 to be detected, 4 are arranged at the rear of the vehicle 9 to be detected, and 10 are arranged at the side and roof of the vehicle 9 to be detected. The specific layout of the area array cameras 4 is related to the projection screen 6 and the detection surface of the vehicle 9 to be detected, and is arranged according to the principle of mirror object phase measurement deflectometry, such as Figure 1In the figure, at the XOY perspective, the area array camera 4 used for detecting the front face of the vehicle 9 to be inspected shoots the front face of the vehicle at an angle along the Y direction; similarly, the area array camera 4 used for detecting the rear of the vehicle 9 to be inspected shoots the rear face at an angle along the Y direction; at the XOZ perspective, the area array camera 4 used for detecting the hood, roof, and trunk lid of the vehicle 9 to be inspected shoots at an angle along the X direction, and its background virtual image is the top projection screen 63; at the XOY perspective, the area array camera 4 used for detecting defects on both sides of the vehicle 9 to be inspected shoots at an angle along the X direction, and its background virtual image is the parallel vertical screen 61.

[0045] like Figure 1 As shown in the YOZ perspective, each part of the vehicle 9 can only be imaged within a certain area at a time. This requires the vehicle transport mechanism 3 to move the vehicle 9 along the X-axis for stop-and-go photography. Because the roof, hood, and trunk lid vary in height on most vehicle models, and the area array camera 4 used for roof photography is stationary, the vehicle transport mechanism 3 must move the vehicle 9 up and down along the Z-axis to accommodate these height differences during stop-and-go photography.

[0046] The specific steps for collecting vehicle body images using the above-mentioned vehicle body paint defect detection device are as follows:

[0047] 1) If Figure 2 As shown, the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis direction to the hood inspection section, and then the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move upward along the Z-axis direction so that the hood of the vehicle to be inspected 6 reaches the hood inspection height. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto the corresponding projection screen 6. The area array cameras 4 set at the front and roof positions of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 takes a picture of the stripes on the projection screen 6 through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0048] 2) The vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens 6. The area array cameras 4 located at the front and roof of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 photographs the fringes on the projection screen 6 through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0049] 3) Repeat step 2) until the engine hood is completely mapped.

[0050] 4) If Figure 3As shown, the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis direction to the roof inspection section, and then the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move downward along the Z-axis direction, so that the roof of the vehicle to be inspected 9 reaches the roof inspection height and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto the corresponding projection screen 6. The area array cameras 4 arranged at the side and roof positions of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 takes a picture of the stripes on the projection screen 6 through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0051] 5) The vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens 6. The area array cameras 4 located on the side and roof of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 photographs the fringes on the projection screen 6 through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0052] 6) Repeat step 5) until the roof is mapped.

[0053] 7) If Figure 4 As shown, the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis direction to the front face inspection section, and the eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto the corresponding projection screen 6. The area array cameras 4 arranged at the front of the vehicle to be inspected 9 sequentially collect images. Each area array camera 4 takes a picture of the stripes on the projection screen 6 through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0054] 8) The vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens 6. The area array cameras 4 located at the front of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 photographs the fringe on the projection screen 6 through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0055] 9) Repeat step 8) until the front of the vehicle is captured.

[0056] 10) If Figure 5As shown, the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis direction to the trunk lid inspection section, and then the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move upward along the Z-axis direction, so that the trunk lid of the vehicle to be inspected 9 reaches the trunk lid inspection height and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens 6. The area array cameras 4 arranged at the side and rear positions of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 takes a picture of the stripes on the projection screen 6 through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0057] 11) The vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens 6. The area array cameras 4 located on the side and rear of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 photographs the stripes on the projection screen 6 through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0058] 12) Repeat step 11) until the trunk lid is completely captured.

[0059] 13) If Figure 6 As shown, the vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis direction to the vehicle rear inspection section, and the eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto the corresponding projection screen 6. The area array cameras 4 arranged at the rear of the vehicle to be inspected 9 sequentially collect images. Each area array camera 4 takes a picture of the fringe on the projection screen 6 through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images.

[0060] 14) The vehicle body conveying mechanism 3 drives the vehicle to be inspected 9 to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors 5 sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens 6. The area array cameras 4 located at the rear of the vehicle to be inspected 9 sequentially capture images. Each area array camera 4 photographs the stripes on the projection screen 6 through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images.

[0061] 15) Repeat step 14) until the rear end of the vehicle is completely captured;

[0062] 16) Use the calibration information to stitch the collected sinusoidal phase-shifted fringe images, and use the principle of phase measurement deflectometry to process the sinusoidal phase-shifted fringe images to obtain curvature maps, diffuse reflection maps, specular reflection maps, and glossiness maps. Then, through traditional image processing algorithms and deep learning algorithms, detect dirt defects, texture defects, scratch and bump defects, and bump defects.

[0063] The specific image processing methods are as follows:

[0064] During the image acquisition process, the projection screen 6 flashes four sinusoidal phase-shifted fringe images horizontally and vertically, and their sinusoidal phases differ by π / 2. The light intensity in the image acquired by the camera is:

[0065] I1(u,v)=I dc +I a cos[φ(u, v)]

[0066] I2(u,v)=I dc +I a cos[φ(u, v)+π / 2]

[0067] I3(u,v)=I dc +I a cos[φ(u,v)+π]

[0068] I4(u,v)=I dc +I a cos[φ(u,v)+3π / 2]

[0069] According to the four-step phase shift principle, the wrapping phase of the image can be calculated

[0070]

[0071] Wrapping phase calculated by phase shift algorithm is unique within a phase cycle, but since there are multiple fringes throughout the measurement space, Sawtooth distribution, such as Figure 7 shown.

[0072] The coordinate system relationship between each area array camera 4 and its corresponding projection screen 6 is calibrated in advance for image stitching. If a point on the car body paint surface has a concave-convex defect, it will inevitably cause a slight difference between the normal line of this point and other points in the neighborhood. The change in the normal line of this point relative to the neighborhood is decomposed into the X and Y directions of the area array camera 4, and the gradient change is expressed as tan(θ x ) and tan(θ y ).

[0073] like Figure 8As shown, for the imaging of a single area array camera 4, the gradient change causes the phase change of the fringe in the image. and The relationship between them is related to the distance from the point to the projection screen 6 and the fringe period of the projection screen 6, and the relationship is as follows:

[0074]

[0075] The calculation of Lx and Ly in phase measurement deflectometry is very computationally intensive. Considering that Lx and Ly are basically unchanged compared to the defect and its neighborhood, the defect detection method proposed in this invention does not solve Lx and Ly, but directly compares the neighborhood of the wrapped phase image to find the suspected defect area. Figure 9 As shown, the first-order derivatives of the horizontal and vertical wrapping phases are calculated to obtain the horizontal and vertical curvatures of the surface, and the square sum of the two is calculated to obtain the curvature map. The glossiness I0 is calculated by averaging the eight fringe images to obtain the glossiness map.

[0076] I0=0.125*(I x1 +I x2 +I x3 +I x4 +I y1 +I y2 +I y3 +I y4 )

[0077] Diffuse reflection image I diff It is to extract the dark field imaging part of the stripes. The calculation process is as follows:

[0078] dif x =min(min(I x1 , I0), min(I x3 ,I0))+min(min(I x2 , I0), min(I x4 , I0))

[0079] dif y =min(min(I y1 , I0), min(I y3 ,I0))+min(min(I y2 , I0), min(I y4 , I0))

[0080] I diff =0.5*(dif x +dif y )

[0081] Specular Reflection Figure I specIt is to extract the bright field imaging part of the stripes. The calculation process is as follows:

[0082] spe x =max(max(I x1 , I0), max(I x3 ,I0))+max(max(I x2 , I0), max(I x4 , I0))

[0083] spe y =max(max(I y1 , I0), min(I y2 ,I0))+min(min(I y2 , I0), min(I y4 , I0))

[0084] I spec =0.5*(spe x +spe y )

[0085] like Figure 10 As shown, a single area array camera 4 can output a composite image, including specular reflection images, diffuse reflection images, glossiness images, and curvature images. By stitching all area array cameras 4 and the field of view together, a composite image of the entire vehicle body to be inspected is generated. Defects such as dirt, texture defects, scratches, bruises, and bumps on the paint surface will appear as strong abnormal signals in one or more of the composite images. Defect detection can then be performed using traditional image processing methods such as blob analysis, edge extraction, template matching, and frequency domain analysis, or through deep learning algorithms.

[0086] In the description of the present invention, it should be noted that the terms "upper", "lower", "left", "right", "top", "bottom", "side", "end", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limiting the present invention.

[0087] In addition, in the description of the embodiments of the present invention, unless otherwise expressly specified or limited, the terms "installed," "connected," "connected," "set," and "provided with" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention in specific circumstances.

[0088] Finally, it should be noted that the above embodiments are only specific implementation methods of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A vehicle body paint defect detection device, characterized by: The system comprises a servo controller, an industrial computer, a vehicle body conveying mechanism, a plurality of area array cameras, a plurality of projectors and a plurality of projection screens. The industrial computer is connected to the servo controller, which is connected to the vehicle body conveying mechanism. The vehicle body conveying mechanism can drive the vehicle to be inspected to move in the length and height directions under the action of the servo controller and the industrial computer. The industrial computer is connected to the projector. The projection screen is arranged outside the projector and corresponds to the projector. The area array cameras are distributed around the vehicle to be inspected. There are a total of 8 projectors, of which 2 are set at the front position of the vehicle to be detected, 2 are set at the rear position of the vehicle to be detected, and 4 are set at the side position and roof position of the vehicle to be detected. The projection screen includes 2 parallel vertical screens parallel to the XOZ plane and parallel to the body of the vehicle to be detected, 4 vertical deflected vertical screens arranged at 40°~50° with the XOZ plane and arranged around the body of the vehicle to be detected, and two top projection screens arranged above the vehicle body to be inspected and with an angle of 100° to 110° to each other. The projection direction of the projector corresponding to the parallel vertical screen is an inclined projection facing the positive direction of the X-axis under the XOY viewing angle, and the angle between the projection direction and the positive direction of the X-axis is 40° to 50°. The projection direction of the projector corresponding to the deflected vertical screen is perpendicular to the deflected vertical screen. The projection direction of the projector corresponding to the top projection screen is an inclined projection facing the positive direction of the X-axis under the XOZ viewing angle, and the angle between the projection direction and the positive direction of the X-axis is 40° to 50°. The number of the area array cameras is 18, of which 4 are correspondingly arranged at the front position of the vehicle to be detected, 4 are correspondingly arranged at the rear position of the vehicle to be detected, and 10 are correspondingly arranged at the side position and roof position of the vehicle to be detected.

2. The vehicle body paint defect detection device according to claim 1, characterized in that: The area array camera has 5 million pixels and a 2 / 3 target surface.

3. The vehicle body paint defect detection device according to claim 1, characterized in that: The projector is a DLP projector.

4. A method for detecting vehicle body paint defects, using the vehicle body paint defect detection device according to any one of claims 1 to 3, characterized in that: The steps include: 1) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the hood inspection section, and then the vehicle body conveying mechanism drives the vehicle to be inspected upward along the Z-axis until the hood of the vehicle to be inspected reaches the hood inspection height. The eight projectors sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens. The area array cameras located at the front and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringes on the projection screen through reflection from the vehicle body surface, generating eight sinusoidal phase-shifted fringe images. 2) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras located at the front and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images. 3) Repeat step 2) until the hood is completely mapped. 4) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the roof inspection section, and then the vehicle body conveying mechanism drives the vehicle to be inspected downward along the Z-axis until the roof of the vehicle to be inspected reaches the roof inspection height and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens. The area array cameras located on the side and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringes on the projection screen through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images. 5) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras located on the side and roof of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images. 6) Repeat step 5) until the roof is mapped. 7) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the front face inspection section. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras positioned at the front of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images. 8) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras positioned at the front of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images. 9) Repeat step 8) until the front of the vehicle is captured. 10) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the trunk lid inspection section, and then the vehicle body conveying mechanism drives the vehicle to be inspected to move upward along the Z-axis until the trunk lid of the vehicle to be inspected reaches the trunk lid inspection height and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images horizontally and four vertically onto corresponding projection screens. The area array cameras located at the side and rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringes on the projection screen through reflection from the vehicle body surface, thereby obtaining eight sinusoidal phase-shifted fringe images. 11) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, one horizontal and one vertical, onto corresponding projection screens. The area array cameras located on the side and rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, generating eight sinusoidal phase-shifted fringe images. 12) Repeat step 11) until the trunk lid is completely captured. 13) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the vehicle rear inspection section. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras located at the rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images. 14) The vehicle body conveying mechanism drives the vehicle to be inspected to move along the X-axis to the next jigsaw puzzle photographing position and remains stationary. The eight projectors sequentially project four sinusoidal phase-shifted fringe images, four in the horizontal and four in the vertical, onto corresponding projection screens. The area array cameras located at the rear of the vehicle to be inspected sequentially capture images. Each area array camera photographs the fringe on the projection screen through reflection from the vehicle body surface, obtaining eight sinusoidal phase-shifted fringe images. 15) Repeat step 14) until the rear of the vehicle is captured. 16) Use the calibration information to stitch the collected sinusoidal phase-shifted fringe images, and apply the principle of phase measurement deflectometry to process the sinusoidal phase-shifted fringe images to obtain curvature maps, diffuse reflection maps, specular reflection maps, and glossiness maps. Then, through traditional image processing algorithms and deep learning algorithms, detect dirt defects, texture defects, scratch and bump defects, and bump defects.

Images