KR20240049958A - Multi-camera lens defect screening method and system using stitching condition - Google Patents

Abstract

A method and system for selecting defective multicamera lenses using stitching conditions are disclosed. A system for selecting lens defects of a multi-camera using stitching conditions according to an embodiment of the present invention includes an image acquisition unit that acquires an inspection pattern from a camera image captured by the multi-camera and corrects distortion; A stitching application unit that generates a composite image by combining distortion-corrected camera images using a stitching technique; an intersection inspection unit that compares an intersection of a pattern image included in the composite image with an intersection of the inspection pattern; a straight line inspection unit that derives a straight line from the synthesized image using a designated image processing technique and checks the slope of the straight line; and a defect selection unit that outputs a result that there is a lens defect if the number of intersection points is different as a result of comparison in the intersection inspection unit or if the slope of the straight line is greater than a reference value as a result of inspection in the straight line inspection unit.

Description

Multi-camera lens defect screening method and system using stitching condition {Multi-camera lens defect screening method and system using stitching condition}

The present invention relates to a method and system for selecting defective multi-camera lenses, and more specifically, to a method and system for selecting defective multi-camera lenses using stitching conditions.

As technology advances, the supply of cameras increases and multi-cameras become commercialized, camera image combining (stitching or panorama) technology using them has also been developed. Accordingly, inspection results for each camera for defects in raw materials such as foreign matter adhesion (stains) and scratches have become necessary to provide reliable camera products.

In the case of mobile phone cameras, defects may occur due to defects in the module itself or insufficient management by the user. If there is a foreign object, normal images are not output, which causes the quality of the multi-camera to deteriorate.

Additionally, in the case of lens defects due to foreign substances or stains, the stitching function may not operate properly or the image composition results may be distorted, making it impossible to obtain normal results.

Conventionally, when defects such as stains, scratches, or adhesion of foreign substances occurred during multi-camera shooting, a method of visually checking was used. In addition, most of them had the problem of low detection efficiency because they could not identify the criteria for automatic determination of the causes of scratches and stains on the lens.

Korean Patent No. 10-2184950 (registered on November 25, 2020) - System for panoramic imaging

The present invention provides the user with camera status information about defects caused by adsorption of foreign substances on the lens or stains such as fingerprints through the stitching function of a multi-camera, establishes a standard for foreign matter tracking inspection using the interpolation method between cameras and stitching results, and The purpose is to provide a method and system for selecting defective multicamera lenses using stitching conditions that can provide reliable data.

Other objects of the present invention will become clearer through the preferred embodiments described below.

According to one aspect of the present invention, there is a system for selecting lens defects of a multi-camera using stitching conditions, comprising: an image acquisition unit that acquires an inspection pattern from a camera image captured by the multi-camera and corrects distortion; A stitching application unit that generates a composite image by combining distortion-corrected camera images using a stitching technique; an intersection inspection unit that compares an intersection of a pattern image included in the composite image with an intersection of the inspection pattern; a straight line inspection unit that derives a straight line from the synthesized image using a designated image processing technique and checks the slope of the straight line; and a defect sorting unit that outputs a result indicating that there is a lens defect when the number of intersections is different as a result of comparison with the intersection point inspection unit or when the slope of the straight line is greater than a reference value as a result of the inspection by the straight line inspection unit. A multi-camera lens defect screening system is provided. .

When distortion occurs in the camera image, the image acquisition unit may perform distortion correction after interpolating the brightness of the peripheral area.

The straight line inspection unit extracts corner points using the Si-Tomas detection algorithm, converts and interpolates the relative difference ratio per distance based on the corner points, and performs straight component matching through an edge detection technique based on the interpolated composite image. Proceeding, the tilt amount according to the slope of the straight line can be converted.

The defect selection unit extracts the color value of a pixel that is confirmed to be defective, derives characteristic values based on the ratio difference according to brightness and the tilt amount, and can classify the lens defect as a foreign matter, stain, or normal.

The defect selection unit classifies the lens defects according to the following equation,

Here, C is the Si-Tomas detection coordinate, D is the relative distance between Si-Tomas detection coordinates, and R is the interpolation result,

The detection coordinates (C) are selected based on the case where there are coordinates within the Si-Tomas detection coordinate interpolation section, the matching point (M) is set based on the case where there is an edge line (E) through this, and the set matching point (M ), check the tilt amount (T) as the difference between Si-Tomasi detection coordinates (C),

Extract the color and brightness values (R, G, B, Y values) of pixels that are confirmed to be defective, and characterize the characteristic values (A) based on the ratio difference (color difference) (U) and tilt amount (T) according to brightness. Derives

If the characteristic value (A) exceeds 1, the defect screening unit determines it as a foreign matter, if the characteristic value (A) exceeds 1, it determines it as a stain, and if it is 0.5 or less, it can be judged as normal.

Meanwhile, according to another aspect of the present invention, there is a method of selecting a lens defect of a multi-camera using stitching conditions, comprising: acquiring a camera image of an inspection pattern captured by the multi-camera in an image acquisition unit; performing distortion correction on the camera image; Generating a composite image by combining distortion-corrected camera images in a stitching application unit using a stitching technique; Comparing, in an intersection inspection unit, an intersection of a pattern image included in the composite image and an intersection of the inspection pattern; In a straight line inspection unit, deriving a straight line using a designated image processing technique for the composite image and inspecting the slope of the straight line; and a defect screening step in which the defect screening unit outputs a result as a lens defect if the number of intersections is different as a result of comparison with the intersection inspection unit or the slope of the straight line is greater than a reference value as a result of inspection in the straight line inspection unit. A screening method is provided.

In the distortion correction performing step, when distortion occurs in the camera image, distortion correction may be performed after interpolation of peripheral brightness is performed.

The step of checking the slope of the straight line includes extracting a corner point using a Si-Tomasi detection algorithm; Converting and interpolating the relative difference ratio per distance based on the corner point; The method may include converting a tilt amount according to the inclination of the straight line by performing straight component matching using an edge detection technique based on the interpolated composite image.

The defect selection step includes extracting color values of pixels determined to be defective; It may include the step of deriving characteristic values based on the ratio difference according to brightness and the tilt amount and classifying the lens defect as foreign matter, stain, or normal.

In the defect selection step, the lens defects are classified according to the following equation,

Here, C is the Si-Tomas detection coordinate, D is the relative distance between Si-Tomas detection coordinates, and R is the interpolation result,

The detection coordinates (C) are selected based on the case where there are coordinates within the Si-Tomas detection coordinate interpolation section, the matching point (M) is set based on the case where there is an edge line (E) through this, and the set matching point (M ), check the tilt amount (T) as the difference between Si-Tomasi detection coordinates (C),

Extract the color and brightness values (R, G, B, Y values) of pixels that are confirmed to be defective, and characterize the characteristic values (A) based on the ratio difference (color difference) (U) and tilt amount (T) according to brightness. Derives

If the characteristic value (A) exceeds 1, the defect screening unit determines it as a foreign matter, if the characteristic value (A) exceeds 1, it determines it as a stain, and if it is 0.5 or less, it can be judged as normal.

Other aspects, features and advantages in addition to those described above will become apparent from the following drawings, claims and detailed description of the invention.

According to an embodiment of the present invention, through the stitching function of a multi-camera, camera status information is notified to the user about defects caused by foreign matter adsorption on the lens or stains such as fingerprints, and foreign matter tracking inspection is performed using an interpolation method between cameras and stitching results. It has the effect of establishing standards and providing reliable data.

1 is a block diagram of a multi-camera lens defect screening system using a stitching function according to an embodiment of the present invention;
Figure 2 is a flowchart of a method for selecting defective multi-camera lenses using a stitching function according to an embodiment of the present invention;
Figure 3 is an example of a multi-camera distortion correction process and failure case,
Figure 4 is an example of corner point extraction and straight line extraction performed in the straight line inspection unit;
Figure 5 is a conceptual diagram of calculating the tilt amount,
Figure 6 is a graph comparing the transmittance on the lens curve of a normal lens (OK) and a defective lens (NG).

Since the present invention can make various changes and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all changes, equivalents, and substitutes included in the spirit and technical scope of the present invention.

When a component is said to be “connected” or “connected” to another component, it is understood that it may be directly connected or connected to the other component, but that other components may exist in between. It should be. On the other hand, when it is mentioned that a component is “directly connected” or “directly connected” to another component, it should be understood that there are no other components in between.

Terms such as first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another.

The terms used herein are only used to describe specific embodiments and are not intended to limit the invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as “comprise” or “have” are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are not intended to indicate the presence of one or more other features. It should be understood that it does not exclude in advance the possibility of the existence or addition of elements, numbers, steps, operations, components, parts, or combinations thereof.

In addition, the components of the embodiments described with reference to each drawing are not limited to the corresponding embodiments, and may be implemented to be included in other embodiments within the scope of maintaining the technical spirit of the present invention, and may also be included in separate embodiments. Even if the description is omitted, it is natural that a plurality of embodiments may be re-implemented as a single integrated embodiment.

In addition, when describing with reference to the accompanying drawings, identical or related reference numbers will be assigned to identical or related elements regardless of the drawing symbols, and overlapping descriptions thereof will be omitted. In describing the present invention, if it is determined that a detailed description of related known technologies may unnecessarily obscure the gist of the present invention, the detailed description will be omitted.

Figure 1 is a block diagram of a multi-camera lens defect sorting system using a stitching function according to an embodiment of the present invention, and Figure 2 is a flow chart of a multi-camera lens defect sorting method using a stitching function according to an embodiment of the present invention. , Figure 3 is an example of a multi-camera distortion correction process and a failure case, Figure 4 is an example of corner point extraction and straight line extraction performed in the straight line inspection unit, Figure 5 is a conceptual diagram of calculating the tilt amount, and Figure 6 is a normal This is a comparison graph of the transmittance on the curved surface of the lens (OK) and the defective lens (NG).

A multi-camera lens defect screening system and method using a stitching function according to an embodiment of the present invention uses a tilt amount for a composite image in which images taken from each single camera constituting the multi-camera are combined using a stitching function to select a lens defect. It is characterized by being able to distinguish defects.

Referring to FIG. 1, the multi-camera lens defect screening system 100 using the stitching function according to this embodiment applies the stitching function through distortion correction to the image acquired from the multi-camera 50 to be inspected. Create a composite image and apply a designated image processing technique to the composite image to calculate the amount of tilt compared to the straight component to determine whether the lens is defective and whether the camera is defective among the plurality of cameras 51 and 52 constituting the multicamera 50. selects.

The multi-camera 50 may include two or more cameras 51 and 52.

The multi-camera 50 may be implemented as a type in which an optical zoom camera, a wide-angle camera, a TOF camera, etc. are provided as a single module in one user terminal, such as a smartphone. Alternatively, the multi-camera 50 may refer to a grouping of two or more image acquisition devices (eg, CCTV) disposed at locations spaced apart from each other.

Hereinafter, the description will be made assuming that two cameras (the first camera 51 and the second camera 52) are included in the multi-camera 50. This is for convenience of understanding and explanation of the invention, and of course, three or more cameras may be included in a multi-camera.

Information about the relative positional relationship between the first camera 51 and the second camera 52 included in the multi-camera 50 may be predetermined or obtained as camera parameters.

The multi-camera lens defect screening system 100 may include an image acquisition unit 110, a stitching application unit 120, an intersection inspection unit 130, a straight line inspection unit 140, and a defect selection unit 150.

The image acquisition unit 110 acquires the image captured by the multi-camera 50. The multi-camera lens defect screening system 100 is connected to the multi-camera 50 by wire or wirelessly, and can acquire images (camera images) captured by the multi-camera 50 in real time, periodically, or in response to requests. You can.

The camera image acquired by the image acquisition unit 110 may include a first image captured by the first camera 51 and a second image captured by the second camera 52. The first image and the second image may be images taken from different viewpoints of the same subject depending on the location and direction of each camera.

The stitching application unit 120 may correct distortion of the camera image using a distortion correction coefficient and generate a composite image, which is a panoramic image, by combining the distortion-corrected images.

In this embodiment, since a distortion correction coefficient is required for stitching application, a pre-designated pattern (for example, a grid pattern with regular intervals) can be primarily photographed to obtain this. In this case, the entire pattern is captured in each of the first and second images, and if the entire pattern is not captured in any one image based on the center point of the image, the camera(s) belonging to the multi-camera 50 Reshooting can be performed by changing at least one of the position and direction of .

If normal results are not obtained in the distortion verification process for the pattern image during the distortion correction coefficient calculation process, a brightness interpolation process may be performed. This is because differences in brightness may occur depending on the location where the photo was taken. When brightness interpolation for the peripheral brightness deviation is performed on one or more of the first image and the second image, the distortion of the pattern image can be removed.

Interpolation for the peripheral brightness deviation can be performed as shown in Equation 1 below.

[Equation 1]

Here, i, j represent a matrix, P represents the pixel position as an array, and n is the number of pixels to be grouped for peripheral averaging in the pixel unit, which is the number of peripheral averaging that can be set by the user. And L is the average value of the matrix.

The stitching application unit 120 can generate a composite image using the distortion correction result with brightness interpolation, and verify the multi-camera 50 at a later stage.

When applying stitching, an image processing technique according to the SURF algorithm can be applied, and a matching point for combining can be determined between two camera images that are combined with each other.

In the process of verifying the distortion, the straight line inspection unit 140 can check whether a straight line is correct or incorrect in a point-to-point relationship and whether there is a difference in slope after connecting the lines. In this case, X, which is the intersection point of the straight lines, may be a matching point component.

The intersection inspection unit 130 can perform inspection of these intersections.

[Equation 2]

Here, B is the black detection pixel point, Max is the maximum size value, and X is the intersection point.

In the case of P pixels, it is a matrix, and in the case of the Max value, the maximum value expressed in P, that is, if the pattern is white/black, the value of 255 can be expressed as the Max value. In other words, Equation 2 tracks the intersection of patterns. In order to extract the value of X, the value X = 1 must be satisfied, and it can be checked whether the pattern intersection of

If the intersection count (Pattern Count) in the actual pattern and the intersection count (X) in the image do not match, the lens may be classified as defective. In this case, the defect selection unit 150 may determine that the lens is defective and output the result. If the lens defect is caused by a foreign substance, cleaning can be performed and rematching can be performed.

If the intersection test is passed, the straight line tester 140 can classify lens defects depending on whether the point-to-point connection line has a slope.

Corner points can be detected in a composite image by applying a corner detection technique based on the camera parameters of the multi-camera that has been determined not to be defective. For example, the Shi & Tomasi detection technique may be applied as a corner detection technique. The Si-Tomas detection technique is a corner detection technique that adds a threshold to the Harris corner algorithm.

Based on the corner point, the relative difference ratio per distance can be converted and interpolated (slope: Y = ax + b).

[Equation 3]

Here, C is the Si-Tomasi detection coordinate, which is output as a matrix result and increases from 1 to n orders from the upper left. D is the relative distance between Si-Tomasi detection coordinates, and R is the interpolation result, meaning that the difference in distance is calculated within a 5% error.

Based on the interpolated image, straight component (straight line) matching is performed using an edge detection algorithm, and the tilt amount (tilt ratio) is converted.

[Equation 4]

The detection coordinates (C) are selected based on the case where there is a coordinate within the Si-Tomas detection coordinate interpolation interval (R), and the matching point (M) is set based on the case where there is an edge line through this. E is the edge extraction value (0, 255).

Based on the set matching point (M), the tilt amount (T) can be confirmed by the difference between the Si-Tomasi detection coordinates (C) (see FIG. 5).

If it is confirmed that the defect is due to a ratio difference, the defect can be distinguished according to the tilt ratio and color value.

[Equation 5]

Here, R, G, and B are color values, Y is the brightness value, U is the color difference, and A is the characteristic result.

According to Equation 5, the color value and brightness value (R, G, B, Y value) of the pixel determined to be defective are extracted. And the characteristic value (A) is derived based on the ratio difference (color difference) (U) and tilt amount (T) according to brightness.

The defect selection unit 150 may determine it to be a foreign body if the characteristic result value corresponding to the ratio value is greater than 1. Additionally, you can select stains from 0.5 to 1 and request cleaning. If it is 0.5 or less, it is judged as normal, but the user's judgment can be determined based on the lens difference. The lens difference is the rate of change for each lens, and may have different results depending on the angle and location of the lens.

As a result of regular analysis, numerical values below 0.5 are normal, but due to the lens's deviation, it can be determined that there is a problem with normal interpolation.

In order to select lens defects of the multi-camera 50, calculation may be made based on the relative ratio of the lens curved surface. As shown in Figure 6, when comparing two normal products and two defective samples, it can be seen that there is a difference in transmittance on the curved surface of the lens.

However, as data such as the user terminal's auto white balance (AWB) and lens shading correction (LSC) are input, analysis of the type of defect may become difficult.

Therefore, extracting corner points and inspecting them based on the amount of distortion, as in this embodiment, may be more efficient in analyzing defect types.

Hereinafter, with reference to FIG. 2, a multi-camera lens defect selection method performed in the multi-camera lens defect selection system will be described.

In order to apply stitching conditions, a pre-designated pattern may be provided in the photographing area. The pattern is a grid pattern in which white and black squares alternate and repeat, and the spacing between each grid can be constant.

The image acquisition unit 110 may acquire a camera image of a pattern from the multi-camera 50 (step S200). The camera image may be a plurality of single camera images each captured by each camera (the first camera 51 and the second camera 52) of the multi-camera 50.

Distortion occurs in camera images even when the same pattern is photographed due to the characteristics (position, direction, etc.) of each camera of the multi-camera 50, so camera distortion correction can be performed for each location (step S205). In this way, the distortion correction coefficient for each camera image can be calculated through distortion correction using camera parameters.

When performing distortion correction, it is applied based on the ratio value of the entire image. Since the distortion correction value is based on the entire range, if the tilt result is reflected, the correction result will be distorted. Referring to Figure 3, camera distortion correction for each location and examples of distortion correction failure are shown.

If distortion correction is performed normally but tilt occurs, the difference is small or unrecognizable to the naked eye. In fact, if there is a stain on the top of the lens, it is expressed as a slight difference in the amount of light transmission. Therefore, it cannot be said that normal results are output even though distortion correction has been performed.

Therefore, when a distortion correction defect occurs, image correction may be performed on the camera image using shading data.

The stitching application unit 120 may perform distortion correction on the pattern image and apply stitching to generate a composite image that is a panoramic image (step S210). Here, when applying stitching, an image synthesis technique according to the SURF algorithm can be used.

The amount of deviation according to the matching point is calculated using the SURF algorithm, and if the amount of deviation from other cameras is greater than the standard value, it is possible to determine whether the camera is defective.

The intersection inspection unit 130 may perform the inspection by counting the number of intersections in the pattern image output from the composite image and comparing the number of intersections with the number of intersections in the actual pattern. If the number of intersection points is not the same, the defect selection unit 150 may output an inspection result indicating that there is a lens defect.

If it is determined to be normal, the straight line inspection unit 140 extracts corner points from the synthesized image using the Si-Tomasi detection algorithm (step S215). Referring to Figure 4, the results of performing Si-Tomas detection and Sobel and Huffman transformation on an example image are shown.

In addition to extracting corner points, analysis of the color channel of the synthetic image can also be performed. This is to later distinguish defects using the color value along with the tilt amount.

The straight line inspection unit 140 selects a matching section based on the position of the corner point, converts the relative difference ratio per distance, and interpolates (step S220). Edge detection is performed based on the interpolated image (step S225), and the tilt difference (tilt amount) using the straight component is confirmed (step S230).

If the defect is a defect due to a difference in tilt ratio, the defect selection unit 150 extracts a color value for characteristic separation and can extract detailed results between items based on the curvature according to the value (step S230).

Based on the location where the tilt ratio difference occurred, it is possible to determine which of the cameras (the first camera 51 and the second camera 52) belonging to the multi-camera 50 used in the composite image had a defect in the lens of the camera. It may be possible. Therefore, the number of subjects that need to be checked for lens defects can be reduced by half and the check can be performed more quickly.

According to the multi-camera lens defect screening system and method using stitching conditions according to this embodiment, foreign matter tracking using a multi-camera is possible, and room for improvement of defects according to characteristic separation can be identified.

When a defect occurs due to a problem with the lens, the cause of the problem can be identified quickly, increasing operator convenience.

As resolution increases and the size of the lens angle of view increases, simple inspection may be possible by changing the detection criteria.

The method for selecting defective multi-camera lenses using the above-described stitching conditions can also be implemented in the form of a recording medium containing instructions executable by a computer, such as an application or program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and non-volatile media, removable and non-removable media. Additionally, computer-readable media may include computer storage media. Computer storage media includes both volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules, or other data.

The multi-camera lens defect screening method using the above-mentioned stitching conditions can be executed by an application installed by default on the terminal (this may include programs included in the platform or operating system, etc. installed by default on the terminal), and the user can use the application. It may also be executed by an application (i.e. program) installed directly on the master terminal through an application providing server such as a store server, application, or web server related to the service. In this sense, the multi-camera lens defect screening method using the above-described stitching conditions is implemented as an application (i.e., program) installed by default in the terminal or directly installed by the user and can be recorded on a computer-readable recording medium such as the terminal. You can.

Although the present invention has been described above with reference to preferred embodiments, those skilled in the art can vary the present invention without departing from the spirit and scope of the present invention as set forth in the claims below. You will understand that it can be modified and changed.

50: Multicamera 100: Multicamera lens defect screening system
110: Image acquisition unit 120: Stitching application unit
130: intersection inspection section 140: straight line inspection section
150: Defect sorting unit

Claims (10)

A system for selecting lens defects of a multi-camera using stitching conditions,
an image acquisition unit that acquires a camera image captured by the multi-camera and corrects distortion of the inspection pattern;
A stitching application unit that generates a composite image by combining distortion-corrected camera images using a stitching technique;
an intersection inspection unit that compares an intersection of a pattern image included in the composite image with an intersection of the inspection pattern;
a straight line inspection unit that derives a straight line from the synthesized image using a designated image processing technique and checks the slope of the straight line; and
A multi-camera lens defect screening system comprising a defect sorting unit that outputs a result as if there is a lens defect when the number of intersection points is different as a result of comparison by the intersection point inspection unit or when the slope of the straight line is greater than a reference value as a result of inspection by the straight line inspection unit.
According to paragraph 1,
The image acquisition unit performs distortion correction after interpolating the brightness of the peripheral area when distortion occurs in the camera image.
According to paragraph 1,
The straight line inspection unit extracts corner points using the Si-Tomas detection algorithm, converts and interpolates the relative difference ratio per distance based on the corner points, and performs straight component matching through an edge detection technique based on the interpolated composite image. A multi-camera lens defect screening system characterized by converting the tilt amount according to the slope of the straight line.
According to paragraph 3,
The defect selection unit extracts the color value of the pixel that has been confirmed to be defective, derives characteristic values based on the ratio difference according to brightness and the tilt amount, and determines the lens defect to be classified as foreign matter, stain, or normal. Multi-camera lens defect screening system.
According to paragraph 4,
The defect selection unit classifies the lens defects according to the following equation,


Here, C is the Si-Tomas detection coordinate, D is the relative distance between Si-Tomas detection coordinates, and R is the interpolation result,


The detection coordinates (C) are selected based on the case where there are coordinates within the Si-Tomas detection coordinate interpolation section, the matching point (M) is set based on the case where there is an edge line (E) through this, and the set matching point (M ), check the tilt amount (T) as the difference between Si-Tomasi detection coordinates (C),


Extract the color and brightness values (R, G, B, Y values) of pixels that are confirmed to be defective, and characterize the characteristic values (A) based on the ratio difference (color difference) (U) and tilt amount (T) according to brightness. Derives
A multi-camera lens defect screening system characterized in that the defect selection unit determines a foreign matter if the characteristic value (A) exceeds 1, determines it as a stain if the characteristic value (A) is between 0.5 and 1, and determines it as normal if the characteristic value (A) is less than 0.5.
As a method of selecting lens defects of a multicamera using stitching conditions,
Acquiring a camera image of the inspection pattern captured by the multi-camera in an image acquisition unit;
performing distortion correction on the camera image;
Generating a composite image by combining distortion-corrected camera images in a stitching application unit using a stitching technique;
Comparing, in an intersection inspection unit, an intersection of a pattern image included in the composite image and an intersection of the inspection pattern;
In a straight line inspection unit, deriving a straight line using a designated image processing technique for the composite image and inspecting the slope of the straight line; and
In the defect screening unit, if the number of intersection points is different as a result of comparison with the intersection inspection unit or the slope of the straight line is greater than a reference value as a result of inspection in the straight line inspection unit, a defect screening step of outputting a result as a lens defect is a multi-camera lens defect screening step. method.
According to clause 6,
In the distortion correction performing step, when distortion occurs in the camera image, interpolation of peripheral brightness is performed and then distortion correction is performed.
According to clause 6,
The slope test step of the straight line is,
extracting corner points using a Si-Tomasi detection algorithm;
Converting and interpolating the relative difference ratio per distance based on the corner point;
A multi-camera lens defect screening method comprising the step of performing straight component matching using an edge detection technique based on the interpolated composite image and converting the tilt amount according to the slope of the straight line.
According to clause 8,
The defect selection step includes extracting color values of pixels determined to be defective; A multi-camera lens defect screening method comprising deriving characteristic values based on the ratio difference according to brightness and the tilt amount and classifying the lens defect as foreign matter, stain, or normal.
According to clause 9,
In the defect selection step, the lens defects are classified according to the following equation,


Here, C is the Si-Tomas detection coordinate, D is the relative distance between Si-Tomas detection coordinates, and R is the interpolation result,


The detection coordinates (C) are selected based on the case where there are coordinates within the Si-Tomas detection coordinate interpolation section, the matching point (M) is set based on the case where there is an edge line (E) through this, and the set matching point (M ), check the tilt amount (T) as the difference between Si-Tomasi detection coordinates (C),


Extract the color and brightness values (R, G, B, Y values) of pixels that are confirmed to be defective, and characterize the characteristic values (A) based on the ratio difference (color difference) (U) and tilt amount (T) according to brightness. Derives
A multi-camera lens defect selection method, characterized in that the defect selection unit determines a foreign matter if the characteristic value (A) exceeds 1, determines it as a stain if the characteristic value (A) is between 0.5 and 1, and determines it as normal if the characteristic value (A) is less than 0.5.