Inspection method for paint defects of carrier and inspection tunnel

Abstract

The invention relates to a method for inspecting paint defects on a vehicle, which method enables the 3D coordinates of paint defects on a vehicle (7) to be determined with the aid of an inspection tunnel. The inspection tunnel (1) comprises a number of cameras (2), a conveyor (8), a position encoder (9) for measuring the movement of the vehicle (7), a total station (11) for measuring fixed points of the inspection tunnel structure (12), a calibration chessboard (3) and a calibration pattern (4), and a processing and storage device (10) for storing images acquired by the cameras, computer-aided design files of the vehicle and a defect recognition algorithm.

Description

[0001] Purpose of the Invention

[0002] The object of the present invention is a method for inspecting vehicles for paint defects and an inspection tunnel equipped with the necessary devices for carrying out the method for inspecting vehicles for paint defects.

[0003] The method of the invention seeks to automate the inspection process of the final finish of the paint in the carriers performed in the carrier production chain. Optionally, defects found are repaired. Technical Field

[0004] The present invention belongs to the technical field of inspection devices intended to detect and analyze paint defects on vehicles, as well as detection and analysis methods performed by the aforementioned devices. Background Art

[0005] Today, systems for scanning reflective surfaces using fixed cameras with artificial vision are used to determine the location of paint defects in 2D. Besides classifying them into several types, this information is used by production plants to: indicate the location of the defect to maintenance operators; generate statistics for the quality department; control the response of the painting process; and detect recurring errors as a single unit.

[0006] Since the painted carrier is a 3D object, obtaining the paint defects of the carrier in 2D still requires an operator to convert the detected 2D points into a 3D object. Therefore, the 2D-based system is not yet fully automatic.

[0007] It would therefore be desirable to find a fully automated solution to the problem of paint defects in carriers. Summary of the Invention

[0008] The invention discloses a method for inspecting paint defects of a carrier and an inspection tunnel for the paint defects of a carrier.

[0009] A first aspect of the present invention discloses a vehicle inspection tunnel for paint defects. The inspection tunnel comprises: a camera for acquiring images of the vehicle; a conveyor for linearly moving the vehicle longitudinally through the inspection tunnel; a position encoder for measuring the movement of the vehicle; a total station for accurately measuring fixed points of the inspection tunnel structure in 3D, with a calibration board on which a calibration pattern is located; and a processing and storage device for storing at least the images acquired by the camera, a computer-aided design (CAD) file of the vehicle, and a defect recognition algorithm. Furthermore, the processing and storage device is connected to the camera, the conveyor, and the position encoder. Finally, the processing and storage device is adapted to perform the steps of the vehicle inspection method for paint defects defined below in the second aspect of the present invention. Optionally, the inspection tunnel may include a repair tool.

[0010] Note that a position encoder or encoder is a pulse generator that is connected to the shaft of the motor that moves the conveyor and generates pulses as the motor rotates. In the present invention, the encoder is used to measure the progress of the carrier along the conveyor as the pulses are converted into length measurements.

[0011] In an embodiment for an inspection tunnel, the calibration pattern is formed by a grid of squares arranged in a staggered pattern. Furthermore, the calibration pattern includes a data matrix code and fiducial markers. The data matrix code and fiducial markers contain information related to the calibration pattern, such as the number of rows and columns, the location of the center square, the size of the squares, and the color of the squares. Preferably, the squares are black and white, as they provide better contrast by requiring identification of their connection points.

[0012] In another embodiment of the inspection tunnel, the inspection tunnel further comprises an inverted U-shaped support structure and a front support structure to support a camera in the inspection tunnel. These supports for the camera have the advantage of placing the camera in such a way that a full scan of the vehicles in the inspection tunnel can be performed.

[0013] The second aspect of the present invention discloses a method for inspecting paint defects on a vehicle associated with the inspection tunnel of the first aspect of the present invention. The method for inspecting paint defects on a vehicle comprises the following steps:

[0014] Calibration involves inspecting the camera in the tunnel and calculating the camera's intrinsic and extrinsic parameters;

[0015] Establishing a common reference system -SCE- for inspecting the tunnel and linking the cameras to the common reference system -SCE-;

[0016] Calculate the 3D coordinates of at least four reference points of the vehicle using stereo vision based on a common reference system (SCE), obtaining the X, Y, and Z coordinates of each reference point;

[0017] Calculating the X, Y, Z coordinates of each reference point based on the vehicle reference system - SCV - from the computer-aided design - CAD - file and 3D measurements of the vehicle;

[0018] acquiring at least two synchronized 2D images of the vehicle by means of a camera, wherein an identifier -ID- of each synchronized 2D image is associated with a spatial position of the vehicle relative to the inspection tunnel, and applying a defect recognition algorithm storing the X and Y coordinates of each defect and the identifier -ID- based on a common reference system -SCE;

[0019] combining the synchronized 2D images into a 3D image, wherein defects in the 3D image have X, Y, Z coordinates linked to a common reference system - SCE;

[0020] The X, Y, Z coordinates of the defect in the 3D image linked to the vehicle reference system -SCV- are calculated with the help of the following relations:

[0021] SCV = inverse (MR) x SCE

[0022] Where SCV is a matrix defining the X, Y, Z coordinates linked to the vehicle reference frame - SCV -; SCE is a matrix defining the X, Y, Z coordinates linked to the common reference frame - SCE -, and MR is the relation matrix between the two reference frames, defining the necessary translations, rotations, and scales from one to the other.

[0023] Through the previous steps, the 3D coordinates of each defect found in the paint of the carrier are obtained. This information can be used to repair the defect. To this end, the method of inspecting paint defects on a carrier of the present invention further includes the following steps:

[0024] Calculation of the camera-vehicle vector between the position of the camera and the center point coinciding with the X, Y, Z coordinates of the defect linked to the vehicle reference system -SCV-;

[0025] generating four additional vectors parallel to the camera-vehicle vector without rotation but with equidistant and radial translations of predetermined distances;

[0026] Calculate the 3D coordinates of the intersection of each additional vector on the vehicle to obtain four 3D points (P1x, P1y, P1z) (P2x, P2y, P2z) (P3x, P3y, P3z) (P4x, P4y, P4z);

[0027] Calculate the distance from each 3D point obtained from the previous step to the center point;

[0028] Discard vectors on the edges or surfaces of the modeling lines with carriers (as these will distort the final result of the angle);

[0029] Calculating triangles on the surface of the vehicle, wherein each triangle is calculated by connecting the 3D coordinates of the center point with two intersection points of the non-discarded vectors;

[0030] Calculate the normal vector to the calculated surface;

[0031] Calculates the average of the previous normal vectors to get a single normal vector.

[0032] The normal vector represents an approach vector of a repair tool used to repair the paint defect of the vehicle.

[0033] The camera calibration process described in the camera calibration step requires the calculation of intrinsic and extrinsic parameters. To calculate the intrinsic parameters, the following sub-steps are performed:

[0034] acquiring at least two images of a calibration checkerboard including at least a data matrix code and fiducial marks;

[0035] Decode the data matrix code to obtain the square size, number of rows and columns of the calibration chessboard;

[0036] Determining the center square of the calibration chessboard based on the data matrix code;

[0037] Calculate all connections of squares starting from the center square;

[0038] The optical center, the focal length, at least six radial distortion parameters (K1-K6) and at least two tangential distortion parameters (P1, P2) are calculated based on the connections of the squares, the size of the optics included in the camera and the size of the CCD unit of the camera.

[0039] Furthermore, in order to calculate the extrinsic parameters, the method for inspecting paint defects of a vehicle includes the following sub-steps:

[0040] placing a calibration chessboard within the inspection tunnel at a location where it is visible to at least one camera;

[0041] The measurements of the calibration chessboard are obtained using a total station by:

[0042] Four fixed points on the tunnel structure were measured and checked with the help of a total station;

[0043] Iteratively stationing the total station to obtain a common reference system -SCE- relative to the vehicle conveyor in the inspection tunnel;

[0044] measuring at least twelve auxiliary points located on a calibration chessboard in a common reference system - SCE - with the aid of a total station;

[0045] The relationship between the common reference frame - SCE - and the calibration chessboard is calculated by using rigid transformations and estimates;

[0046] Each camera saves at least one image of the calibration chessboard;

[0047] A local coordinate system is calculated for each camera and a transformation of the local coordinate system to the common reference system - SCE - is calculated.

[0048] Once the cameras are calibrated and a common reference frame - SCE - is established, the cameras are linked to the common reference frame - SCE -. The next step of the application consists in calculating the 3D coordinates of the four reference points of the vehicle based on the common reference frame - SCE - by stereo vision, obtaining the X, Y, and Z coordinates of each reference point. To calculate the 3D coordinates of the four reference points of the vehicle by stereo vision, the following steps are performed:

[0049] Select two cameras on each side of the vehicle that have visual access to the four reference points to be measured;

[0050] The reference points to be calculated on the carrier are selected taking into account its synchronous movement on the conveyor relative to the inspection tunnel.

[0051] Alternatively and in addition to the two previous steps, a recognition pattern can be generated by means of a comparison vector search algorithm for identifying subsequent identical vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 An inspection tunnel for inspecting the paint of a vehicle according to the present invention is shown, and the body of the vehicle within the inspection tunnel is shown.

[0053] Figure 2 An inspection tunnel of the present invention is shown with a calibration chessboard already placed inside.

[0054] Figure 3 The calibration pattern carried by the calibration chessboard is shown.

[0055] Figure 4 A total station is shown which acquires the 3D coordinates of four fixed points of the inspection tunnel.

[0056] Figure 5 The calculation of four reference points of the vehicle with the aid of a camera is shown.

[0057] Figure 6 A vehicle extracted from a CAD file is shown, based on which the 3D coordinates of any point of the vehicle can be derived.

[0058] Figure 7 The common reference system SCE, the vehicle reference system SCV and the coordinate transformation from one system to the other are shown.

[0059] Figure 8 A 2D image of a paint defect of a carrier and the 2D position by means of the X, Y coordinates of the defect are shown.

[0060] Figure 9 A 3D image of a paint defect of a carrier is shown along with the 3D position by means of the X, Y, Z coordinates of the defect.

[0061] Figure 10 A 3D image of a paint defect of a carrier and the approach of a repair tool for repairing the defect are shown. DETAILED DESCRIPTION

[0062] Label list

[0063] 1. Inspection tunnel

[0064] 2. Camera

[0065] 3. Calibrate the chessboard

[0066] 4. Calibration pattern: 4a - Data Matrix Code; 4b - Fiducial Mark; 4c - Grid; 4d - Center Grid

[0067] 5. Inverted U-shaped support structure for camera

[0068] 6.Front support structure

[0069] 7. Carrier

[0070] 8. Transporter of the carrier

[0071] 9. Position encoder or encoder

[0072] 10. Processing and storage devices

[0073] 11. Total Station

[0074] 12. Check the fixing points of the tunnel structure

[0075] 13. Intrinsic parameters

[0076] 14. External parameters

[0077] 15. Vehicle reference point

[0078] 16. Common Reference System -SCE-

[0079] 17. Vehicle Reference System -SCV-

[0080] 18. XY 2D coordinates of the defect

[0081] 19. XYZ 3D coordinates of the defect

[0082] 20.2D Synchronous Image

[0083] 21.3D Images

[0084] 22. Camera-Carrier Vector

[0085] 23. Vector parallel to the camera-vehicle vector

[0086] 24. Repair tools

[0087] 25. 3D Intersection of Parallel Vectors

[0088] 26. Defects.

[0089] Preferred embodiments of the present invention are described below based on the illustrated drawings.

[0090] Figure 1An inspection tunnel 1 according to the invention for carrying out an inspection of vehicles for paint defects is shown. Figure 1 An inspection tunnel 1 is shown, with the chassis of a carrier 7 on a conveyor 8 therein. Figure 1 , only the chassis of the carrier is shown, but the inspection tunnel 1 of the present invention can also perform inspections for paint defects on fully assembled carriers. The conveyor 8 has the purpose of moving the carrier 7 within the inspection tunnel 1. The inspection tunnel 1 is mainly composed of a camera 2, which is supported by an inverted U-shaped support structure 5 and a front support structure 6. In addition, the inspection tunnel 1 has a carrier conveyor 8, a position encoder 9 and a processing and storage device 10. The processing and storage device 10 is a processor and a memory, which are configured to perform the steps of the method described in the present invention and to be interconnected with the other elements that constitute the inspection tunnel. In addition, the encoder 9 is used to enable the inspection tunnel 1 to always know the position of the carrier. As will be further described, this allows synchronized images to be obtained between the position of the carrier and the image obtained by the camera at a given moment.

[0091] Before starting the inspection process of the vehicle, it is necessary to calibrate the camera 2. The calibration of the camera 2 includes calculating the intrinsic and extrinsic parameters of the camera.

[0092] To calculate the intrinsic parameters, a calibration chessboard 3 is placed inside the inspection tunnel 1, as shown in Figure 2 Its shape is shown in Figure 3 The calibration pattern 4 shown in FIG is placed on a calibration chessboard 3. The calibration pattern 4 consists of alternating black and white squares 4c in a staggered pattern, similar to a chessboard. The calibration pattern 4 includes a data matrix code 4a and fiducial markers 4b arranged in different white squares. The processing and storage device 10 performs the following steps to calculate intrinsic parameters: it acquires at least two images of the calibration chessboard 3 with the calibration pattern 4 using a camera 2; it decodes the data matrix code 4a to obtain the dimensions of the squares 4c, the center square 4d, and the number of rows and columns of the calibration chessboard 3. Using this information, the processing and storage device 10 calculates all connections of the squares starting from the center square. Using the connections of the squares, the dimensions of the optics included in the camera, and the dimensions of the camera's CCD cell, it calculates intrinsic parameters 13: the optical center, the focal length, at least six radial distortion parameters (K1-K6), and at least two tangential distortion parameters (P1, P2).

[0093] Regarding the extrinsic parameters, in addition to using the calibration chessboard 3, a total station 11 is used, such as Figure 4As shown in . First, a calibration board 3 is placed within the inspection tunnel 1 at a location visible to at least one camera 2. Next, measurements of the calibration board 3 are performed using a total station 11. This requires the creation of a common reference system (SCE) 16, for which iterative measurements are performed using the total station 11 at four points on the inspection tunnel 1 structure and twelve points on the calibration board 3. In other words, the same four points of the inspection tunnel 1 and twelve points on the calibration board 3 are measured from different positions of the total station 11 relative to the inspection tunnel 1. These different positions are preferably locations that the vehicle will cover on the conveyor 8. In other words, the relationship between the measurements taken at the four fixed points 12 of the inspection tunnel 1 and the twelve points on the calibration board 3 allows the creation of the common reference system (SCE) 16. Once the common reference system (SCE) 16 is defined, each camera 2 must be linked relative to it so that the 3D coordinates of the vehicle's paint defects can be determined. To this end, the relationship between the common reference system (SCE) and the calibration board is first calculated using rigid transformations and estimates. Subsequently, at least one image of the calibration chessboard 3 is stored for each camera 2 , the local coordinate system of each camera 2 is calculated and finally the transformation of the local coordinate system into the common reference system —SCE— is calculated.

[0094] like Figure 5 As shown in , once the camera 2 is calibrated and a common reference system - SCE- 16 is established, the camera is linked to the common reference system - SCE- 16. The next step of the application consists in calculating the 3D coordinates of the four reference points 15 of the vehicle based on the common reference system - SCE- by stereo vision, obtaining the X, Y, Z coordinates of each reference point. Figure 5 As shown in , on each side of the carrier 7, two cameras 2 acquire an image of a reference point 15 and obtain the 3D coordinates of the reference point 15 relative to a common reference system - SCE - 16. Thus, the 3D coordinates of two reference points 15 are obtained, one on each side of the carrier. Simultaneously or sequentially the 3D coordinates of two further reference points 15, also on both sides of the carrier, are calculated. This can be performed simultaneously if there are cameras 2 that can acquire images of the other two reference points 15, or sequentially by moving the carrier 7 with the aid of the conveyor 8 until both cameras 2 have access to the other two reference points 15. For the sequential approach, the 3D coordinates of the reference points 15 have a correction factor known by the position encoder or encoder 9 located in the conveyor 8 for eliminating the movement. In order not to have to repeat the calculation for the same carrier that will subsequently be inspected by the inspection tunnel 1, the 3D coordinates of the reference point 15 are calculated. Figure 5 The steps described, the processing and storage means 10 can generate recognition patterns with the aid of a contrast vector search algorithm.

[0095] Once the 3D coordinates of the four reference points 15 of the vehicle 7 are calculated relative to the common reference system - SCE-16, it is possible to establish a correspondence between the 3D coordinates of the four reference points 15 of the vehicle relative to the common reference system - SCE-16 and the 3D coordinates of those same four reference points 15 of the vehicle relative to the vehicle reference system - SCV-17 ( Figure 7 ), locate four reference points in the Computer Aided Design -CAD- file, which contains the 3D measurements / coordinates of the vehicle, such as Figure 6 In other words, at each reference point 15 ( Figure 5 ) and the same reference point 15 extracted from the CAD file ( Figure 6 ) to establish a corresponding relationship between ( Figure 7 ).

[0096] Next, the surface of the carrier 7 is analyzed to check for paint defects of the carrier. To this end, several 2D synchronized images 20 ( Figure 8 ). The 2D images are called "synchronous" because for each "synchronous" image there is a direct relationship between the identifier of the image -ID-, the spatial position of the camera and the spatial position of the vehicle, since the vehicle is on the conveyor 8 and its spatial relationship to the inspection tunnel 1 is known by means of the position encoder 9. Figure 8 As shown in , a defect recognition algorithm is applied to each 2D synchronized image 20 , with which the processing and storage means can calculate the X, Y coordinates of each defect 18 based on a common reference system - SCE- 16 , associate it with an identifier - ID-, and store it for later processing thereof.

[0097] Once several 2D simultaneous images 20 (at least two) of the defect 18 are acquired, they are combined into a 3D image 21 to obtain a 3D image, wherein the defect in the 3D image has X, Y, Z coordinates 19 linked to a common reference system - SCE. Since the aim is to find paint defects on the vehicle itself, the conversion of the X, Y, Z coordinates of the defect in the 3D image of the common reference system - SCE - to the vehicle reference system - SCV - is calculated with the aid of the following equation relationship:

[0098] SCV = inverse (MR) x SCE

[0099] Where SCV is a matrix defining the X, Y, Z coordinates linked to the vehicle reference system -SCV-; SCE is a matrix defining the X, Y, Z coordinates linked to the common reference system -SCE-, and MR is a relational matrix defining the necessary translation, rotation, and scale from the -SCV- to the -SCE- reference system. Using this equation, the 3D coordinates (X, Y, Z) of the paint defect 19 of the vehicle 7 are obtained in the 3D image linked to the vehicle reference system -SCV-, as Figure 9 As shown in .

[0100] A possible application of knowing the 3D coordinates (X, Y, Z) of the paint defect 19 of the carrier in the 3D image linked to the carrier reference system (SCV) is the repair of the defect 26 with the aid of a repair tool. The repair tool needs to know the angle at which it should reach the defect 26, i.e. the 3D coordinates 19 of the paint defect of the carrier. This is why the method of the invention can further include the following steps to calculate the angle at which the repair tool 24 must reach the defect on the carrier. Figure 10 The steps represented in are: · calculation of the camera-vehicle vector 22 between the position of the camera 2 and the center point coinciding with the X, Y, Z coordinates of the defect 19 linked to the vehicle reference system -SCV-;

[0101] Four additional vectors 23 are generated parallel to the camera-vehicle vector 22 without rotation but with equidistant and radial translations at predetermined distances of approximately 1 mm;

[0102] Calculating the 3D coordinates of the intersection point 25 of each additional vector on the carrier to obtain four 3D points (P1x, P1y, P1z)(P2x, P2y, P2z)(P3x, P3y, P3z)(P4x, P4y, P4z), and applying the aforementioned technique to calculate the 3D position of the paint defect;

[0103] Calculate the distance from each 3D point obtained from the previous step to the center point;

[0104] Discard vectors on the edges or surfaces of modeling lines with carriers, as these will distort the final result of the angle;

[0105] Calculating triangles on the surface of the vehicle, where each triangle is calculated by connecting two intersection points of the 3D coordinate of the center point and the non-discarded vectors;

[0106] Calculate the normal vector to the calculated surface;

[0107] The previous normal vectors are averaged to obtain a single normal vector that represents the angle at which the repair tool must reach the defect on the carrier.

Claims

1. A method for inspecting paint defects of vehicles by means of an inspection tunnel, characterized in that The method comprises the following steps: calibrating a camera included in an inspection tunnel by calculating intrinsic and extrinsic parameters of the camera; establishing a common reference system -SCE- for the inspection tunnel and linking the camera to the common reference system -SCE-; Calculating the 3D coordinates of at least four reference points of the vehicle using stereo vision based on the common reference system - SCE - to obtain the X, Y, and Z coordinates of each reference point; calculating the X, Y, Z coordinates of each reference point based on a vehicle reference system - SCV - from a computer-aided design - CAD - file and 3D measurements of said vehicle; acquiring, by means of the camera, at least two synchronized 2D images of the vehicle, wherein an identifier -ID- of each synchronized 2D image is associated with the spatial position of the vehicle relative to the inspection tunnel, and applying a defect recognition algorithm for calculating the X and Y coordinates of each defect and of the identifier -ID- based on the common reference system -SCE; combining the synchronized 2D images into a 3D image, wherein defects in the 3D image have X, Y, Z coordinates linked to the common reference system - SCE -; The X, Y, Z coordinates of the defect in the 3D image linked to the vehicle reference system -SCV- are calculated with the help of the following relations: SCV = inverse (MR) x SCE where SCV is a matrix defining the X, Y, Z coordinates linked to the vehicle reference frame -SCV-; MR is a relational matrix, and SCE is a matrix defining the X, Y, Z coordinates linked to the common reference frame -SCE-, Calculating a camera-vehicle vector between the position of the camera and a center point coinciding with the X, Y, Z coordinates of the defect linked to the vehicle reference system -SCV-; generating four additional vectors parallel to the camera-vehicle vector under conditions of equidistant and radial translation of predetermined distances; Calculate the 3D coordinates of the intersection of each additional vector on the carrier to obtain four 3D points (P1x, P1y, P1z) (P2x, P2y, P2z) (P3x, P3y, P3z) (P4x, P4y, P4z); Calculate the distance from each 3D point obtained from the previous step to the center point; discarding vectors on edges or surfaces of a shape line with said carrier; Calculating triangles on the surface of the vehicle, wherein each triangle is calculated by connecting the 3D coordinates of the center point with two intersection points of non-discarded vectors; Calculate the normal vector to the calculated surface; Calculates the average of the previous normal vectors to get a single normal vector.

2. The method for inspecting paint defects of a vehicle by means of an inspection tunnel according to claim 1, characterized in that The step of calibrating the camera further comprises the following sub-steps to calculate the intrinsic parameters: acquiring at least two images of a calibration checkerboard including at least a data matrix code and fiducial marks; Decoding the data matrix code to obtain the square size, number of rows and number of columns of the calibration chessboard; determining a center square of the calibration chessboard based on the data matrix code; Calculate all connections of squares starting from the central square; The optical center, the focal length, at least six radial distortion parameters (K1-K6) and at least two tangential distortion parameters (P1, P2) are calculated based on the connections of the squares, the size of the optics included in the camera and the size of the CCD unit of the camera.

3. The method for inspecting paint defects of a vehicle by means of an inspection tunnel according to claim 1, characterized in that: The step of calibrating the camera further comprises the following sub-steps to calculate the extrinsic parameters: placing a calibration chessboard within the inspection tunnel at a location where the calibration chessboard is visible from at least one camera; The measurements of the calibration chessboard are obtained using a total station by: measuring four fixed points on the structure of the inspection tunnel by means of the total station; iteratively stationing the total station to obtain a common reference system -SCE- relative to the vehicle conveyor in the inspection tunnel; measuring at least twelve auxiliary points located on the calibration chessboard in the common reference system -SCE- by means of the total station; calculating the relationship between the common reference frame - SCE - and the calibration chessboard by using a rigid transformation and estimation; Each camera stores at least one image of the calibration chessboard; A local coordinate system is calculated for each camera and a transformation of the local coordinate system to the common reference system - SCE - is calculated.

4. The method for inspecting paint defects of a vehicle by means of an inspection tunnel according to claim 1, characterized in that: The step of calculating 3D coordinates by stereo vision further includes the following sub-steps: selecting two cameras on each side of the vehicle that have visual access to the four reference points to be measured; selecting a reference point to be calculated on the vehicle taking into account the synchronous movement of the vehicle on the conveyor relative to the inspection tunnel; A recognition pattern is created with the aid of a comparison vector search algorithm for identifying subsequent identical vehicles.

5. An inspection tunnel for paint defects of a vehicle, wherein the inspection tunnel (1) is characterized in that the inspection tunnel (1) comprises: a camera (2) for acquiring an image of the vehicle (7); a conveyor (8) which moves the carrier (7) linearly and longitudinally through the inspection tunnel (1); a position encoder (9) which measures the movement of the carrier (7); a total station (11) for measuring fixed points (12) of the structure of the inspection tunnel; a calibration chessboard (3) on which the calibration pattern (4) is located; A processing and storage device (10) which stores at least the images acquired by the camera, the computer-aided design (CAD) file of the carrier and a defect recognition algorithm; connected to the camera (2), the conveyor (8) and the position encoder (9); and wherein the processing and storage device performs the steps of the method according to any one of claims 1 to 4 when performing an inspection of the carrier for paint defects by means of the inspection tunnel.

6. The paint defect inspection tunnel for a vehicle according to claim 5, characterized in that: The calibration pattern (4) is formed of squares (4c) arranged in a staggered form, and wherein the calibration pattern (4) further comprises a data matrix code (4a) and a fiducial mark (4b).

7. The paint defect inspection tunnel for a vehicle according to claim 5, characterized in that The inspection tunnel (1) further comprises an inverted U-shaped support structure (5) and a front support structure (6) for supporting a camera (2) in the inspection tunnel (1).

8. The paint defect inspection tunnel for a vehicle according to claim 5, characterized in that: The inspection tunnel (1) further comprises a repair tool (24) for repairing paint defects of the carrier (7).

Images