Cylindrical profile feature detection method and system based on linear array camera

Abstract

The invention provides a cylindrical profile feature detection method and system based on a linear array camera. The method comprises the following steps: S1, adjusting the position of a detection system and positioning an X-axis measurement position; s2, line frequency calculation and data acquisition of the linear array camera are carried out; s3, splicing a plurality of images; s4, resolution calibration and distortion compensation are carried out; and S5, carrying out cylinder inclination compensation calculation. The method is suitable for detecting the outer surfaces of cylinders with different diameters and heights, has high detection precision and high detection efficiency, and can realize dimensional tolerance and geometric tolerance detection of geometric features such as holes and lines of the outer surfaces of the cylinders and appearance quality detection of surface scratches and the like through algorithm analysis of pictures.

Description

Technical Field

[0001] The present invention relates to the technical field of detection of cylindrical outer surface features. Specifically, it relates to a method and system for detecting cylindrical outer surface features based on a linear array camera. Background Art

[0002] Cylindrical structural components are widely used in the industrial field, such as various tanks and cabins, mostly with thin-walled structures. Features such as holes and grooves are distributed on their outer surfaces for bracket connection and equipment installation. To meet the subsequent assembly requirements, it is necessary to detect the hole diameter, axial distance between holes, circumferential angle between holes, length and width of the slots, and surface scratches on the outer surface. When optically scanning and measuring high-reflectivity small holes, the boundary is blurred, and it is difficult to solve the hole position. When using a three-coordinate contact measurement, it is difficult to measure thin-walled small holes and the measurement time is long. In the past, the axial relative position between holes of such components was mainly measured using a height gauge, and the hole diameter, slot length and width, etc. were measured using a caliper. The circumferential angle between holes was indirectly measured by the method of measuring the chord length with a caliper, and when the holes were not in the same section, the circumferential angle could not be calculated through the chord length. The existing platform caliper measurement method has problems such as poor detection accuracy and reliability, low detection coverage and efficiency, and low inspection informatization level.

[0003] The Chinese patent document with the publication number CN 108426537 B discloses a method and system for rapid full-field detection of an in-situ grinding wheel based on a linear array camera. It uses a bilinear light source for illumination, controls the linear array camera to align with the position to be detected on the grinding wheel through a three-axis translation mechanism, controls the linear array camera to obtain the clearest image to achieve autofocus, obtains the full-field surface two-dimensional topography of the grinding wheel, calculates two-dimensional information such as the area, particle size, spacing between abrasive grains, density distribution of abrasive grains on the grinding wheel, and coordinates of single abrasive grains in the whole field, and determines whether the abrasive grains fall off and the coordinates of the fallen abrasive grains. The invention can complete operations such as autofocus, calibration, image acquisition, analysis and statistics during the working process of the grinding wheel, and realizes the in-situ detection of the grinding wheel. The disadvantage is that the system is applicable to the in-situ detection of the grinding wheel, but it cannot realize the axial splicing of the cylindrical image and high-precision calibration, and the imaging accuracy is relatively low.

[0004] The Chinese patent document with the publication number CN 114280083 B discloses a detection method for industrial X-ray nondestructive testing of large-sized flat castings based on automatic CNC programming using a linear array camera. Utilizing the high-speed refreshing and continuous shooting characteristics of the linear array camera, it quickly captures continuous and complete surface images of the large-sized flat casting to be detected, automatically calculates and guides CNC programming, and controls the X-ray detection platform and X-ray acquisition equipment to capture X-ray images. This invention uses the images obtained by the linear array camera to achieve automatic detection programming for industrial X-ray nondestructive testing of large-sized flat castings, improving the detection efficiency. The drawback is that the detection using the linear array camera is only applicable to guiding programming and cannot meet the requirements for detecting the cylindrical outer surface features in terms of accuracy, detection coverage, etc.

[0005] The Chinese patent document with the publication number CN 212133588 U discloses an angle detection platform for cylindrical objects based on a linear array camera, including a mounting table, a conveyor belt, a handling mechanism, a linear array camera detection mechanism, and a light source, which can quickly correct the angle of the cylinder and facilitate subsequent processing operations. The drawback is the low positioning accuracy and the inability to achieve image stitching outside the single camera field of view. Summary of the Invention

[0006] Aiming at the defects in the prior art, the purpose of the present invention is to provide a detection method and system for cylindrical outer surface features based on a linear array camera.

[0007] The detection method for cylindrical outer surface features based on a linear array camera provided by the present invention includes:

[0008] Step S1: Adjust the position of the detection system and locate the measurement position on the X-axis;

[0009] Step S2: Calculate the line frequency of the linear array camera and collect data;

[0010] Step S3: Stitch multiple images;

[0011] Step S4: Calibrate the resolution and compensate for distortion;

[0012] Step S5: Calculate the cylindrical tilt compensation.

[0013] Preferably, the step S1 includes:

[0014] Before measurement, adjust the axis of the linear array camera lens, the axis of the ranging sensor measurement, the moving trajectory of the center of the precision turntable on the X-axis linear module, and the moving trajectory of the center of the linear array camera on the Z-axis linear module, so that these four are on the same plane;

[0015] The tunnel light source is located between the X-axis linear module and the camera lens. The generatrix of the cylinder forms an image on the camera through the gap of the tunnel light source, and the distance from the ranging sensor to the center of the turntable is calibrated. The lens can be a telecentric lens or a standard lens. After starting the tunnel light source, move the X-axis linear module to make the generatrix of the cylinder clearly form an image on the CCD of the camera, and then lock the X-axis.

[0016] Preferably, the step S2 includes:

[0017] The line frequency F of the line array camera refers to the number of image rows sampled by the camera per second. By setting the line frequency parameter, the image resolution in the height direction and the circumferential direction of the cylinder is made consistent. According to the image resolution sampled in the height direction being K h , the turntable rotation speed n, and the cylinder diameter D, calculate and determine the matching relationship between the line frequency F of the line array camera and the turntable rotation speed n. The calculation formula is as follows:

[0018]

[0019] Move the Z-axis linear module to make the object field of view of the camera lens cover the lowest edge of the cabin, and then lock the Z-axis. The turntable rotates to trigger the camera. The turntable rotates 360°, and the image of the outer profile of the cylinder at the current Z-axis height is collected. When the turntable is rotating for measurement, trigger the ranging sensor at the set angular interval of the turntable C-axis rotation to measure the distance from the ranging sensor to the surface of the cylinder.

[0020] Judge whether the entire cabin has been photographed. When the photographing is completed, the camera stops photographing and enters the subsequent processing. When the photographing is not completed, the Z-axis linear module moves upward to the second measurement position. The object fields of view of the lens at the second measurement position and the first measurement position, that is, the Z-axis acquisition heights corresponding to the images, have an overlapping area. After locking the Z-axis at the second measurement position, the turntable rotates to trigger the camera. The turntable rotates 360°, and the image of the outer profile of the cylinder at the Z-axis height of the second measurement position and the distance measurement of the ranging sensor are completed. Repeat the above steps to complete multi-position measurement until the entire cabin is photographed.

[0021] The step S3 includes: After the Z-axis measurement is completed, use the coordinate values of the Z-axis linear module at each measurement position to complete the stitching of multiple images in the Z-axis direction. When the coordinate values of the Z-axis linear module cannot accurately provide the values, use the common feature points between adjacent measurement position images to achieve image stitching. When stitching, find the corresponding multiple feature pixel points in adjacent images respectively, calculate the rotation and translation matrix for coordinate stitching, and achieve the stitching of adjacent images. After stitching, a full-size complete image of the outer profile of the cylinder is obtained.

[0022] Preferably, the step S4 includes: vertically installing the linear scale on the outer surface of the cylinder, with the scale body parallel to the generatrix of the cylinder; moving the X-axis linear module to make the reading line of the linear scale clearly imaged on the camera CCD, and locking the X-axis; using the method of rotating the turntable to trigger the camera to sample the image of the linear scale; selecting the central area of the linear scale image, i.e., the central area of the lens, and calculating the corresponding object height for each pixel point, which is the pixel resolution in the height direction of the cabin; taking the center point of the linear scale image as the coordinate origin, the pixel difference between the center of the linear scale scale line and the center point of the image as the X value, and the corresponding standard length of the linear scale as the Y value, and selecting a total of m groups of data (x1, y1), (x2, y2)... (x m , y m ) from the center of the image to the edge of the picture, and using the Lagrange interpolation method to calculate the distortion compensation function between the image and the physical length for image distortion compensation at each Z-axis measurement position.

[0023] Preferably, the step S5 includes: taking the center of the turntable as the origin, determining the positive direction of the Z-axis with the installation platform of the lower bottom surface of the cylinder, and the positive direction of the X-axis from the origin to the starting zero point of the C-axis of the turntable, to establish a cylindrical tilt compensation coordinate system; in this coordinate system, using the distance from the ranging sensor to the center of the turntable, the distance from the ranging sensor to the surface of the cylinder, and the rotation angle of the C-axis of the turntable, calculate the coordinates of each point on the surface of the cylinder, fit each measurement point on the surface of the cylinder, and obtain the X and Y coordinate values of the center of the circle of this cross-section. Taking the distance from the ranging sensor to the installation surface of the lower bottom surface of the cylinder as the Z-axis coordinate value, obtain the coordinates of the center of the circle of each measurement position cross-section, and use the coordinate values of the center of the circle of each measurement position cross-section to fit and obtain the three-dimensional equation of the axis, which is the cylindrical tilt compensation equation; when measuring the outer profile holes of the cylinder, use this equation to calculate the radial deviation of each hole, obtain the measured radius R', and the ratio R' / R of the measured radius R' to the theoretical radius R(D / 2) is the included angle scaling ratio coefficient, where R' takes the average value of the measured radii of two holes, and D is the diameter of the cylinder.

[0024] According to the cylindrical outer profile feature detection system based on a linear array camera provided by the present invention, it includes:

[0025] Module M1: Adjust the position of the detection system and locate the measurement position of the X-axis;

[0026] Module M2: Calculate the line frequency of the linear array camera and collect data;

[0027] Module M3: Stitch multiple images;

[0028] Module M4: Calibrate the resolution and compensate for distortion;

[0029] Module M5: Calculate the cylindrical tilt compensation.

[0030] Preferably, the module M1 includes:

[0031] Before measurement, adjust the axis of the linear array camera lens, the measurement axis of the distance measurement sensor, the moving trajectory of the center of the precision turntable on the X-axis linear module, and the moving trajectory of the center of the linear array camera on the Z-axis linear module so that the four are on the same plane;

[0032] The tunnel light source is located between the X-axis linear module and the camera lens, and the cylindrical generatrix passes through the gap of the tunnel light source and is imaged on the camera. Calibrate the distance from the distance measurement sensor to the turntable center; The lens uses a telecentric lens or a standard lens. After starting the tunnel light source, move the X-axis linear module so that the cylindrical generatrix can be clearly imaged on the camera CCD, and lock the X-axis.

[0033] Preferably, the module M2 includes:

[0034] The line frequency F of the linear array camera refers to the number of image rows sampled by the camera per second. By setting the line frequency parameter, the image resolution in the height direction and the circumferential direction of the cylinder is made consistent; According to the image resolution sampled in the height direction is K h 、the rotation speed n of the turntable, and the cylinder diameter D, calculate and determine the matching relationship between the line frequency F of the linear array camera and the rotation speed n of the turntable. The calculation formula is as follows:

[0035]

[0036] Move the Z-axis linear module so that the object field of view of the camera lens covers the lowest edge of the cabin, and lock the Z-axis; The turntable rotates to trigger the camera, and the turntable rotates 360° to complete the image acquisition of the cylindrical outer surface at the current Z-axis height; When the turntable rotates for measurement, trigger the distance measurement sensor at the set angular interval of the turntable C-axis rotation to measure the distance from the distance measurement sensor to the cylindrical surface;

[0037] Judge whether the entire cabin has been photographed. When the photographing is completed, the camera photographing ends and enters subsequent processing; When the photographing is not completed, the Z-axis linear module moves upward to the second measurement position. The object fields of view of the lens at the second measurement position and the first measurement position, that is, the Z-axis acquisition height of the corresponding images, have an overlapping area. After locking the Z-axis at the second measurement position, the turntable rotates to trigger the camera, and the turntable rotates 360° to complete the image acquisition of the cylindrical outer surface at the Z-axis height of the second measurement position and the distance measurement of the distance measurement sensor; Repeat the above operations to complete multi-position measurement until the entire cabin is photographed;

[0038] The module M3 includes: After the Z-axis measurement is completed, use the coordinate values of the Z-axis linear module at each measurement position to complete the splicing of multiple images in the Z-axis direction; When the coordinate values of the Z-axis linear module cannot be accurately provided, use the common feature points between adjacent measurement position images to achieve image splicing; When splicing, find the corresponding multiple feature pixel points in adjacent images respectively, calculate the rotation and translation matrix for coordinate splicing, and realize the splicing of adjacent images; After splicing, obtain the full-size complete image of the cylindrical outer surface.

[0039] Preferably, the module M4 includes: vertically installing a linear scale on the outer surface of a cylinder, with the scale body parallel to the generatrix of the cylinder; moving the X-axis linear module to enable the reading line of the linear scale to be clearly imaged on the camera CCD, and locking the X-axis; using the method of rotating the turntable to trigger the camera to perform image sampling on the linear scale; selecting the central area of the linear scale image, i.e., the central area of the lens, and calculating the corresponding object height for each pixel point, which is the pixel resolution in the height direction of the cabin; taking the center point of the linear scale image as the coordinate origin, the pixel difference between the center of the linear scale graduation and the center point of the image as the X value, and the corresponding standard length of the linear scale as the Y value, and selecting a total of m groups of data (x1, y1), (x2, y2) …… (x m , y m ), and using the Lagrange interpolation method to calculate the distortion compensation function between the image and the physical length for image distortion compensation at each Z-axis measurement position.

[0040] Preferably, the module M5 includes: taking the center of the turntable as the origin, determining the positive direction of the Z-axis with the installation platform on the lower bottom surface of the cylinder, and the positive direction of the X-axis from the origin to the starting zero point of the turntable C-axis, to establish a cylindrical tilt compensation coordinate system; in this coordinate system, using the distance from the distance measuring sensor to the center of the turntable, the distance from the distance measuring sensor to the surface of the cylinder, and the rotation angle of the turntable C-axis, calculate the coordinates of each point on the surface of the cylinder, fit each measurement point on the surface of the cylinder, and obtain the X and Y coordinate values of the center of the circle of this cross-section. Taking the distance from the distance measuring sensor to the installation surface of the lower bottom surface of the cylinder as the Z-axis coordinate value, obtain the coordinates of the center of the circle of each cross-section at each measurement position, and use the coordinate values of the center of the circle of each cross-section at each measurement position to fit and obtain the three-dimensional equation of the axis, and this equation is the cylindrical tilt compensation equation; when measuring the outer profile holes of the cylinder, use this equation to calculate the radial deviation of each hole, obtain the measured radius R', and the ratio R' / R of the measured radius R' to the theoretical radius R(D / 2) is the included angle scaling coefficient, where R' takes the average value of the measured radii of two holes, and D is the diameter of the cylinder.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] (1) The present invention adopts a non-contact visual detection method, and the system operation and measurement evaluation method can be solidified by a program to achieve high-precision image acquisition of the outer profile features of the cylinder. Through algorithm analysis of the pictures, it is possible to detect the geometric feature positions such as holes and lines on the outer profile of the cylinder, as well as the appearance quality detection such as surface scratches, improving the inspection coverage rate, reducing the influence of human factors, and improving the detection reliability;

[0043] (2) The present invention can detect the outer profiles of cylinders with different diameters and heights by moving the X-axis and Z-axis, with strong versatility, changing the previous manual holding method using platforms, height gauges, micrometers, and calipers together, reducing the personnel requirements, and improving the detection efficiency;

[0044] (3) The present invention realizes the high-precision detection of the circumferential and axial relative positions of small holes on the thin-walled cylindrical outer surface, solves the problems that the traditional measurement means of such small holes cannot be detected, the three-coordinate contact measurement is difficult, and the scanning accuracy of small holes is low, and provides a solution for the measurement of such hole positions;

[0045] (4) The present invention uses a linear scale to calibrate the system resolution and compensates for lens distortion, and uses sensors such as a distance measuring sensor to solve the cylindrical tilt compensation equation from the measured values. The operation is simple and the calibration accuracy is high, ensuring the final detection accuracy of the system. Description of the Drawings

[0046] By reading the following detailed description of the non-limiting embodiments with reference to the accompanying drawings, other features, objects, and advantages of the present invention will become more apparent:

[0047] Figure 1 It is a schematic diagram of the detection process;

[0048] Figure 2 It is a schematic diagram of the composition of the detection system. Detailed Embodiments

[0049] The present invention will be described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several changes and improvements can still be made. These all belong to the protection scope of the present invention.

[0050] Embodiment

[0051] The present invention provides a method for detecting the features of the cylindrical outer surface based on a linear array camera, including:

[0052] Step S1, system adjustment and positioning of the X-axis measurement position. Before measurement, use an articulated arm measurement system, a micrometer, etc. to adjust the assembly accuracy of the system hardware, so that the moving trajectories of the axis of the linear array camera lens, the measurement axis of the distance measuring sensor, the center of the precision turntable on the X-axis linear module, and the moving trajectory of the center of the linear array camera on the Z-axis linear module are in the same plane. The tunnel light source is located between the X-axis linear module and the camera lens, and the cylindrical generatrix can be imaged on the camera through the gap of the tunnel light source. Place a standard cylinder on the turntable and calibrate the distance from the distance measuring sensor to the center of the turntable. The lens uses a telecentric lens. The cylindrical cabin is connected to the turntable through a special measurement clamping tooling. After starting the tunnel light source, move the X-axis linear module forward and backward with a step distance of 0.1 mm, so that the cylindrical generatrix can be clearly imaged on the camera CCD. After finding the clearest imaging position, lock the X-axis and adjust the light source intensity to determine the light source intensity when the imaging is the clearest.

[0053] Step S2: Camera line frequency calculation and data acquisition. The line frequency F (r / s) of the line array camera refers to the number of image rows that the camera can sample per second. By setting the line frequency parameter, the image resolutions in the height direction and circumferential direction of the cylinder are made consistent. Given that the image resolution in the height direction is K h (mm / pixles), the rotational speed n (° / s) of the turntable, the diameter D (mm) of the cylinder and other key parameters, the matching relationship between the line frequency F of the line array camera and the rotational speed n of the turntable is calculated and determined. The calculation formula is as follows:

[0054]

[0055] Assume that the outer diameter of the cylinder (such as the cabin body, shaft body, etc.) is 300 mm, the rotational speed of the turntable is 36 ° / s, and the image resolution in the height direction is 0.01 mm / pixles. Then the line frequency of the line array camera is F = 9424.78 r / s.

[0056] Move the Z-axis linear module to make the object field of view of the camera lens cover the lowest edge of the cabin body, and lock the Z-axis. Rotate the turntable to trigger the camera. The turntable rotates 360°, and the image of the outer surface of the cylinder at the current Z-axis height is acquired. When the turntable rotates for measurement, trigger the distance measuring sensor at the set angular interval of the turntable's C-axis rotation, and measure the distance from the distance measuring sensor to the cylinder surface. Assume that the distance measuring sensor is triggered to measure once every 30° of the turntable's C-axis rotation, then 12 distance values are measured.

[0057] Judge whether the entire cabin body has been photographed. When the photographing is completed, the camera shooting ends and enters subsequent processing; when the photographing is not completed, the Z-axis linear module moves upward to the second measurement position. It is necessary to ensure that there is an overlapping area between the object field of view of the lens (the Z-axis acquisition height corresponding to the image) at the second measurement position and the first measurement position. After locking the Z-axis at the second measurement position, rotate the turntable to trigger the camera. The turntable rotates 360°, and the image of the outer surface of the cylinder at the Z-axis height of the second position and the distance measurement of the distance measuring sensor are completed. Repeat the above steps to complete multi-position measurement until the entire cabin is photographed. Assume that the height of the cylinder cabin is 500 mm and the object field of view of the camera for a single shot is 200 mm, then the axial direction needs to be photographed 3 times.

[0058] Step S3: Stitching of multiple images. After the Z-axis measurement is completed, use the coordinate values of the Z-axis linear module at each measurement position to complete the stitching of multiple images in the Z-axis direction; when the coordinate values of the Z-axis linear module cannot be accurately provided, the common feature points between adjacent measurement position images can be used to achieve image stitching. When stitching, find the corresponding multiple feature pixel points in adjacent images respectively, calculate the rotation and translation matrix for coordinate stitching, and realize the stitching of adjacent images. After stitching, a full-size complete image of the outer surface of the cylinder is obtained. If a grating scale is installed on the Z-axis of the system, use the reading of the grating scale to cut off some adjacent overlapping parts of each frame to realize the up and down stitching of the images.

[0059] Step S4: Resolution calibration and distortion compensation 4. Vertically install the linear scale on the outer surface of the cylinder with the scale body parallel to the cylinder generatrix. Move the X-axis linear module to make the reading line of the linear scale clearly imaged on the camera CCD, and lock the X-axis. Use the method of rotating the turntable to trigger the camera to sample the image of the linear scale. Select the central area of the linear scale image (the central area of the lens), and calculate the corresponding object height for each pixel point, which is the pixel resolution in the height direction of the cabin. Taking the center point of the linear scale image as the coordinate origin, the pixel difference from the center of the scale line to the center point of the image as the X value, and the corresponding standard length of the linear scale as the Y value, select a total of m groups of data (x1, y1), (x2, y2)... (x m , y m ), and use the Lagrange interpolation method to calculate the distortion compensation function between the image and the actual length. Assume that the standard length interval of the linear scale is selected as 1 mm, the distance from the outermost edge of the field of view to the center is 90 mm, and the corresponding pixel differences from the center of the scale line to the center of the image are 50, 120, 220,... 7500, then 90 groups of data (50, 1), (120, 2), (220, 3),... (750, 90) are obtained. According to these 90 groups of data, calculate the Lagrange interpolation compensation function for the distortion compensation between the image and the actual length.

[0060] The Lagrange interpolation polynomial L(x) can be expressed as:

[0061]

[0062] where, l j (x) is the Lagrange basis polynomial, defined as:

[0063]

[0064] Each l j (x) is a polynomial of degree n with respect to x and satisfies the condition l j (x k ) = δ jk , where δ jk is the Kronecker delta, which is equal to 1 when j = k and 0 otherwise.

[0065] Therefore, for each j = 0, 1... n, the specific form of l j (x) is:

[0066]

[0067] When this polynomial is available, it can be used to estimate the actual length L(x) at any image position x, that is, the distortion compensation function.

[0068] Step S5: Cylindrical tilt compensation calculation. Taking the center of the turntable as the origin, the positive direction of the Z-axis is determined by the installation platform of the lower bottom surface of the cylinder, and the positive direction of the X-axis is from the origin to the starting zero point of the C-axis of the turntable. A cylindrical tilt compensation coordinate system is established. In this coordinate system, using the distance from the distance measuring sensor to the center of the turntable, the distance from the distance measuring sensor to the surface of the cylinder, and the rotation angle of the C-axis of the turntable, the coordinates of the points on the surface of the cylinder are calculated. By fitting the measured points on the surface of the cylinder, the X and Y coordinate values (a1, b1) of the center of the circle of this cross-section are obtained. If the height of this measuring surface from the installation surface of the lower bottom surface of the cylinder is c1, then the coordinates of the center of the circle of the cross-section at this measuring position are (a1, b1, c1). In this way, the coordinate values of the centers of the circles of the cross-sections at f measuring positions are respectively (a1, b2, c1), (a2, b2, c2) …… (a f , b f , c f ), and using f circular coordinates, a three-dimensional equation of the axis is fitted, and this equation is the cylindrical tilt compensation equation. When measuring the outer profile holes of the cylinder, this equation can be used to calculate the radial deviation of each hole, and the measured radius R' is obtained. The ratio R' / R of the measured radius R' to the theoretical radius R (D / 2) is the included angle scaling coefficient. Among them, R' can take the average value of the measured radii of two holes.

[0069] A straight line in space can be defined by two points or a point and a direction vector. Suppose we have calculated the direction vector of the cylindrical axis as d, and a point p on the axis is known, then the parametric equation of the axis can be expressed as: r(t) = p + td, where t is a parameter and r(t) is the coordinate of any point on the axis.

[0070] As Figure 1 , it is a schematic diagram of the detection process. The system realizes measurement focusing by horizontally moving the X-axis, realizes axial measurement at multiple positions by moving the Z-axis, realizes full-size image acquisition of the outer profile surface by stitching multiple images of one week of the Z-axis, and uses the distance value measured by the distance measuring sensor to calculate the cylindrical tilt compensation equation. After the image is calibrated for resolution and compensated for distortion, image processing algorithms can be used to detect outer profile holes, lines, surface defects, etc.

[0071] As Figure 2, which is a schematic diagram of the composition of the detection system, adopts the structure of "linear array camera + telecentric lens + 3-axis structure", specifically including: linear array camera 1, telecentric lens 2, ranging sensor 3, linear module (Z-axis) 4, camera bracket 5, marble platform 6, tunnel light source 7, precision turntable (C-axis) 8, linear module (X-axis) 9 and the product to be measured 10. Among them, the linear array camera is installed at the rear end of the telecentric lens, and the cylindrical outer surface is imaged on the linear array camera through the telecentric lens; the ranging sensor is used to measure the distance between the lens and the outer surface of the cabin, and the measured value of this distance can be used for cylindrical clamping tilt compensation; the linear module (Z-axis) is used to drive the imaging module composed of the linear array camera and the telecentric lens to move in the vertical direction to realize multi-station image acquisition in the vertical direction; the camera bracket is used to connect the telecentric lens and the linear module (Z-axis); the marble platform serves as the system base to ensure the system stability; the tunnel light source is used for illuminating the cylindrical outer surface during imaging, and the linear array camera can realize image acquisition through the middle gap of the tunnel light source; the precision turntable (C-axis) rotates circumferentially during the imaging of the cylindrical outer surface to realize the acquisition of images in one week; the linear module (X-axis) is used to move the cylinder installed on the turntable to the measurement position (imaging focal length position); the product to be measured is installed on the turntable through the positioning tooling. The system solves the problem of high-precision multi-station imaging of the cylindrical outer surface features, with a simple structure, good process feasibility and stable operation.

[0072] Those skilled in the art know that in addition to implementing the systems, devices and their respective modules provided by the present invention in the form of pure computer-readable program codes, it is entirely possible to make the systems, devices and their respective modules provided by the present invention be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. by logically programming the method steps. Therefore, the systems, devices and their respective modules provided by the present invention can be considered as a kind of hardware components, and the modules included therein for implementing various programs can also be regarded as the structures within the hardware components; the modules for implementing various functions can also be regarded as either software programs for implementing methods or the structures within the hardware components.

[0073] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. Without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for detecting cylindrical surface features based on a linear array camera, characterized in that: include: Step S1: adjusting the position of the detection system and positioning the X-axis measurement position; Step S2: Calculate the line frequency of the linear array camera and collect data; Step S3: stitching multiple images of the collected data; Step S4: performing resolution calibration and distortion compensation on the stitched image; Step S5: Perform cylinder tilt compensation calculation.

2. The cylindrical surface feature detection method based on a line array camera according to claim 1, characterized in that: The step S1 comprises: Before measurement, adjust the linear array camera lens axis, the distance sensor measurement axis, the moving track of the precision turntable center on the X-axis linear module, and the moving track of the linear array camera center on the Z-axis linear module, so that the four items are on the same plane; The tunnel light source is located between the X-axis linear module and the camera lens, and the cylindrical busbar is imaged on the camera through the gap of the tunnel light source, and the distance from the ranging sensor to the center of the turntable is calibrated; the lens uses a telecentric lens or a standard lens. After starting the tunnel light source, move the X-axis linear module so that the cylindrical busbar can be clearly imaged on the camera CCD and lock the X-axis.

3. The cylindrical surface feature detection method based on a linear array camera according to claim 1, characterized in that: The step S2 comprises: The line frequency F of the linear array camera refers to the number of image lines sampled per second by the camera. By setting the line frequency parameters, the image resolution in the height direction and the circumferential direction of the cylinder can be made consistent. The image resolution K in the height direction is h , turntable speed n, cylinder diameter D, calculate and determine the matching relationship between the line array camera line frequency F and the turntable speed n. The calculation formula is as follows: Move the Z-axis linear module so that the object field of view of the camera lens covers the bottom edge of the cabin and lock the Z-axis; the turntable rotates to trigger the camera, and the turntable rotates 360° to complete the image acquisition of the cylindrical outer surface at the current Z-axis height; when the turntable rotates to measure, the distance sensor is triggered according to the set turntable C-axis rotation angle interval to measure the distance from the distance sensor to the cylindrical surface; Determine whether the entire cabin has been photographed. When the photographing is completed, the camera ends the photographing and enters the subsequent processing. When the photographing is not completed, the Z-axis linear module moves upward to the second measurement position, where the lens object field of view of the second measurement position and the first measurement position, i.e., the Z-axis acquisition height of the corresponding image, has an overlapping area. After locking the Z-axis of the second measurement position, the turntable rotates to trigger the camera, and the turntable rotates 360° to complete the cylindrical outer surface image acquisition and ranging sensor distance measurement of the Z-axis height of the second measurement position. Repeat the above steps to complete the multi-position measurement until the photographing of the entire cabin is completed. The step S3 includes: after the Z-axis measurement is completed, the coordinate values ​​of the Z-axis linear module of each measurement position are used to complete the Z-axis direction stitching of multiple images; when the coordinate values ​​of the Z-axis linear module cannot be accurately provided, the common feature points between the images of adjacent measurement positions are used to achieve image stitching; when stitching, corresponding multiple feature pixel points are respectively found in adjacent images, and the rotation and translation matrix of the coordinate stitching is calculated to achieve adjacent image stitching; after the stitching is completed, a full-size complete image of the cylindrical outer surface is obtained.

4. The method for detecting cylindrical outer surface features based on a linear array camera according to claim 1, characterized in that: The step S4 comprises: vertically installing the linear ruler on the outer surface of the cylinder, with the ruler body parallel to the generatrix of the cylinder; moving the X-axis linear module so that the linear ruler reading line can be clearly imaged on the camera CCD, and locking the X-axis; sampling the linear ruler image by using a turntable to rotate and trigger the camera; selecting the central area of ​​the linear ruler image, i.e., the central area of ​​the lens, and calculating the object height corresponding to each pixel point, i.e., the pixel resolution in the cabin height direction; taking the central point of the linear ruler image as the coordinate origin, the pixel difference between the linear ruler engraved line center and the image center point as the X value, and the corresponding linear ruler standard length as the Y value, selecting a total of m groups of data (x1, y1), (x2, y2) ... (x m ,y m ), the Lagrange difference method is used to calculate the distortion compensation function between the image and the actual object length, which is used for image distortion compensation at each Z-axis measurement position.

5. The method for detecting cylindrical outer surface features based on a linear array camera according to claim 1, characterized in that: The step S5 comprises: taking the center of the turntable as the origin, the installation platform of the lower bottom surface of the cylinder determines the positive direction of the Z axis, and the origin points to the starting zero point of the C axis of the turntable as the positive direction of the X axis, and establishing a cylinder tilt compensation coordinate system; in this coordinate system, using the distance from the distance sensor to the center of the turntable, the distance from the distance sensor to the surface of the cylinder, and the rotation angle of the C axis of the turntable, the coordinates of each point on the surface of the cylinder are calculated, fitting each measurement point on the surface of the cylinder to obtain the X and Y coordinate values ​​of the center of the cross-sectional circle, taking the distance from the installation surface of the lower bottom surface of the cylinder as the Z axis coordinate value, obtaining the coordinates of the center of the cross-sectional circle at each measurement position, using the coordinates of the center of the cross-sectional circle at each measurement position, fitting to obtain the three-dimensional equation of the axis, which is the cylinder tilt compensation equation; when measuring the outer surface hole of the cylinder, using this equation, solving the radial deviation of each hole, obtaining the measured radius R', and the ratio of the measured radius to the theoretical radius R(D / 2) R' / R is the angle scaling coefficient, wherein R' takes the average of the measured radii of the two holes, and D is the cylinder diameter.

6. A cylindrical surface feature detection system based on a linear array camera, characterized in that: include: Module M1: perform detection system position adjustment and X-axis measurement position positioning; Module M2: Linear camera line frequency calculation and data collection; Module M3: stitching multiple images of the collected data; Module M4: perform resolution calibration and distortion compensation on the stitched images; Module M5: Perform cylinder tilt compensation calculation.

7. The cylindrical outer surface feature detection system based on a linear array camera according to claim 6, characterized in that: The module M1 comprises: Before measurement, adjust the linear array camera lens axis, the distance sensor measurement axis, the moving track of the precision turntable center on the X-axis linear module, and the moving track of the linear array camera center on the Z-axis linear module, so that the four items are on the same plane; The tunnel light source is located between the X-axis linear module and the camera lens, and the cylindrical busbar is imaged on the camera through the gap of the tunnel light source, and the distance from the ranging sensor to the center of the turntable is calibrated; the lens uses a telecentric lens or a standard lens. After starting the tunnel light source, move the X-axis linear module so that the cylindrical busbar can be clearly imaged on the camera CCD and lock the X-axis.

8. The cylindrical surface feature detection system based on a linear array camera according to claim 6, characterized in that: The module M2 comprises: The line frequency F of the linear array camera refers to the number of image lines sampled per second by the camera. By setting the line frequency parameters, the image resolution in the height direction and the circumferential direction of the cylinder can be made consistent. The image resolution K in the height direction is h , turntable speed n, cylinder diameter D, calculate and determine the matching relationship between the line frequency F of the linear array camera and the turntable speed n. The calculation formula is as follows: Move the Z-axis linear module so that the object field of view of the camera lens covers the bottom edge of the cabin and lock the Z-axis; the turntable rotates to trigger the camera, and the turntable rotates 360° to complete the image acquisition of the cylindrical outer surface at the current Z-axis height; when the turntable rotates to measure, the distance sensor is triggered according to the set turntable C-axis rotation angle interval to measure the distance from the distance sensor to the cylindrical surface; Determine whether the entire cabin has been photographed. When the photographing is completed, the camera ends the photographing and enters the subsequent processing. When the photographing is not completed, the Z-axis linear module moves upward to the second measurement position, where the lens object field of view of the second measurement position and the first measurement position, i.e., the Z-axis acquisition height of the corresponding image, has an overlapping area. After locking the Z-axis of the second measurement position, the turntable rotates to trigger the camera, and the turntable rotates 360° to complete the cylindrical outer surface image acquisition and ranging sensor distance measurement of the Z-axis height of the second measurement position. Repeat the above operations to complete multi-position measurement until the photographing of the entire cabin is completed. The module M3 includes: after the Z-axis measurement is completed, the coordinate values ​​of the Z-axis linear module of each measurement position are used to complete the Z-axis direction stitching of multiple images; when the coordinate values ​​of the Z-axis linear module cannot be accurately provided, the common feature points between the images of adjacent measurement positions are used to achieve image stitching; when stitching, the corresponding multiple feature pixel points are respectively found in adjacent images, and the rotation and translation matrix of the coordinate stitching is calculated to achieve adjacent image stitching; after the stitching is completed, a full-size complete image of the cylindrical outer surface is obtained.

9. The cylindrical outer surface feature detection system based on a linear array camera according to claim 6, characterized in that: The module M4 comprises: vertically installing the linear ruler on the outer surface of the cylinder, with the ruler body parallel to the generatrix of the cylinder; moving the X-axis linear module so that the linear ruler reading line can be clearly imaged on the camera CCD and locking the X-axis; sampling the linear ruler image by using a turntable to rotate and trigger the camera; selecting the central area of ​​the linear ruler image, i.e., the central area of ​​the lens, and calculating the object height corresponding to each pixel point, i.e., the pixel resolution in the cabin height direction; taking the central point of the linear ruler image as the coordinate origin, the pixel difference between the linear ruler engraved line center and the image center point as the X value, and the corresponding linear ruler standard length as the Y value, selecting a total of m groups of data (x1, y1), (x2, y2) ... (x m ,y m ), the Lagrange difference method is used to calculate the distortion compensation function between the image and the actual object length, which is used for image distortion compensation at each Z-axis measurement position.

10. The cylindrical surface feature detection system based on a linear array camera according to claim 6, characterized in that: The module M5 includes: taking the center of the turntable as the origin, the installation platform of the lower bottom surface of the cylinder determines the positive direction of the Z axis, and the origin points to the starting zero point of the C axis of the turntable as the positive direction of the X axis, and establishes a cylinder tilt compensation coordinate system; in this coordinate system, using the distance from the distance sensor to the center of the turntable, the distance from the distance sensor to the surface of the cylinder, and the rotation angle of the C axis of the turntable, the coordinates of each point on the surface of the cylinder are calculated, and each measurement point on the surface of the cylinder is fitted to obtain the X and Y coordinate values ​​of the center of the cross-sectional circle, taking the distance from the distance sensor to the installation surface of the lower bottom surface of the cylinder as the Z axis coordinate value, the coordinates of the center of the cross-sectional circle at each measurement position are obtained, and the coordinates of the center of the cross-sectional circle at each measurement position are used to fit the axis three-dimensional equation, which is the cylinder tilt compensation equation; when measuring the outer surface hole of the cylinder, the equation is used to solve the radial deviation of each hole, and the measured radius R' is obtained, and the ratio of the measured radius to the theoretical radius R(D / 2) R' / R is the angle scaling coefficient, wherein R' takes the average of the measured radii of the two holes, and D is the cylinder diameter.

Images