WO2020133538A1

WO2020133538A1 - Circular feature detection method, processing system and apparatus having storage function

Info

Publication number: WO2020133538A1
Application number: PCT/CN2018/125877
Authority: WO
Inventors: 李洪杰
Original assignee: 深圳配天智能技术研究院有限公司
Priority date: 2018-12-29
Filing date: 2018-12-29
Publication date: 2020-07-02
Also published as: CN111801709B; CN111801709A

Abstract

A circular feature detection method, a processing system, and an apparatus having a storage function. The detection method comprises: receiving an image to be processed, said image comprising a circular feature (S101); extracting a region of interest from said image, and obtaining a first parameter corresponding to the region of interest, the circular feature being located within the region of interest, the region of interest being a circle (S102); using the first parameter to detect the circular feature, to obtain a detection parameter corresponding to the circular feature (S103); using the detection parameter to adjust the first parameter, and returning to the step of using the first parameter to detect the circular feature, recording the obtained detection parameter of each detection, until the distance between first center coordinates and detected center coordinates is less than or equal to a preset threshold (S104); and determining final parameters of the circular feature according to a plurality of recorded detection parameters and a first preset policy, the final parameters comprising the center coordinates and the radius of the circular feature (S105). In the described manner, the detection method improves the accuracy of a detection result.

Description

Circular feature detection method, processing system and device with storage function

Technical field

The present application relates to the field of image processing technology, and in particular, to a circular feature detection method, a processing system, and a device with a storage function.

Background technique

When using machine vision technology to detect the image to be processed, it is not necessary to detect the entire image to be processed, but only to detect a part of the image to be processed, for example, only to detect the edge of the image to be processed.

For example, when the edge is a circular feature, the area to be processed can be outlined on the image to be processed in the form of boxes, circles, ellipses, irregular polygons, etc. This area is called the area of interest (ROI). By performing circular feature detection on the region of interest, the circular feature detection time can be shortened. However, due to the randomness of the region of interest, the results of the circular feature detection will be quite different, and the accuracy and stability of the results cannot be guaranteed.

Summary of the invention

The present application provides a circular feature detection method, processing system, and device with a storage function, which can improve the accuracy and stability of circular feature detection results.

In order to solve the above technical problems, a technical solution adopted by the present application is: to provide a circular feature detection method, the circular feature detection method includes receiving an image to be processed, the image to be processed includes a circular feature; Extract the region of interest from the image to be processed and obtain the first parameter corresponding to the region of interest, wherein the circular feature is located in the region of interest, the region of interest is a circle, and the first The parameters include a first center-of-circle coordinate, an inner diameter, and an outer diameter corresponding to the region of interest; the first feature is used to detect the circular feature to obtain a detection parameter corresponding to the circular feature, the detection The parameters include the detection center coordinates and the detection radius; use the detection parameters to adjust the first parameter, and return to the step of using the first parameters to detect the circular feature, and record the results obtained after each detection Detecting parameters until the distance between the first center-of-circle coordinates and the detected center-of-center coordinates is less than or equal to a preset threshold; determining the final parameters of the circular feature based on the recorded multiple detection parameters and the first predetermined strategy The final parameter includes the center coordinates and radius of the circular feature.

In order to solve the above technical problems, another technical solution adopted by the present application is to provide a circular feature detection method. The circular feature detection method includes: receiving an image to be processed, the image to be processed includes a circular feature; Extracting the region of interest by the image to be processed and obtaining the first parameter corresponding to the region of interest, wherein the circular feature is located in the region of interest, the region of interest is a circle, the The first parameter includes the first center coordinates, inner diameter, and outer diameter corresponding to the region of interest; the first parameter is used to detect the circular feature to obtain a detection parameter corresponding to the circular feature, so The detection parameters include detecting the coordinates of the center of the circle and the detection radius; adjusting the first parameter using the detection parameter, and returning to the step of detecting the circular feature using the first parameter, recording each detection The detection parameters obtained, and the value of the iteration parameter is updated after each detection until the value of the iteration parameter exceeds a preset range; the circular feature is determined according to the recorded multiple detection parameters and the first predetermined strategy Final parameters, the final parameters include the center coordinates and radius of the circular feature.

In order to solve the above technical problems, another technical solution adopted by the present application is to provide a circular feature detection and processing system. The circular feature detection and processing system includes a processor, a memory, a transceiver, and a display. The transceiver is used for Receive the image to be processed, and transmit the image to be processed to the processor, the display is used to display the detection result, the memory stores program instructions, and the program instructions can be loaded by the processor and execute the above The circular feature detection method described in any embodiment.

In order to solve the above technical problems, another technical solution adopted by the present application is to provide a device with a storage function on which program data is stored, and when the program data is executed by a processor, the program described in any of the above embodiments is implemented. Steps in the circular feature detection method.

The beneficial effects of the present application are: different from the situation in the prior art, the circular feature detection method provided by the present application includes: using the first parameter corresponding to the region of interest to obtain the detection parameter corresponding to the circular feature after detection; and using The detection parameter adjusts the first parameter, and returns to the step of detecting the circular feature using the first parameter, and records the detection parameter obtained after each detection, until the distance between the first center coordinates and the detected center coordinates is less than or equal to the preset threshold Determine the final parameters of the circular feature according to the recorded multiple detection parameters and the first predetermined strategy, and the final parameters include the center coordinates and radius of the circular feature. On the one hand, the above-mentioned iterative method can make the region of interest closer to the circular feature, and the center of the region of interest closer to the circular feature, so as to reduce the randomness of the region of interest and improve the circular feature The accuracy of the detection results; on the other hand, the method of using multiple detection parameters and the first predetermined strategy to obtain the final parameters of the circular feature in the above method can give a more average result and further improve the detection result of the circular feature Accuracy and stability, in an application scenario, the circle detection results can reach a few pixels.

[Description of the drawings]

In order to more clearly explain the technical solutions in the embodiments of the present invention, the drawings required in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without paying any creative work, other drawings can also be obtained based on these drawings, in which:

FIG. 1 is a schematic flowchart of an embodiment of a circular feature detection method of this application;

2 is a schematic flowchart of an implementation manner of step S103 in FIG. 1;

3a is a schematic structural diagram of an embodiment of an image of a region of interest in a rectangular coordinate system;

3b is a schematic structural view of an embodiment of converting an image of a region of interest in a rectangular coordinate system in FIG. 3a into a first image in a polar coordinate system;

4 is a schematic flowchart of an implementation manner of step S203 in FIG. 2;

FIG. 5 is a schematic flowchart of an embodiment of screening multiple second pole diameters in step S206 in FIG. 2 to reserve at most one second pole diameter;

FIG. 6 is a schematic flowchart of an embodiment of screening multiple third pole diameters in step S207 in FIG. 2 to reserve at most one third pole diameter;

7 is a schematic flowchart of another embodiment of a circular feature detection method of this application;

8 is a schematic structural diagram of an embodiment of a circular feature detection and processing system of the present application;

9 is a schematic structural diagram of an embodiment of a device with a storage function according to this application.

【detailed description】

The technical solutions in the embodiments of the present invention will be described clearly and completely in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by a person of ordinary skill in the art without making creative efforts fall within the protection scope of the present invention.

Please refer to FIG. 1. FIG. 1 is a schematic flowchart of an embodiment of a circular feature detection method according to the present application. The detection method includes:

S101: Receive an image to be processed, the image to be processed includes a circular feature.

Specifically, the image to be processed may be provided by an imaging device with a shooting function such as a camera, and the image to be processed may be a black-and-white image or a color image. The image to be processed contains circular features, such as round holes, cylinders, etc. of the workpiece. In this embodiment, the circular features may be the outer edge of the workpiece or the edge of the inner hollowed-out area of the workpiece. In practical applications, circular features include noise-free standard circular features, noisy standard circular features, and non-standard circular features (that is, irregular circular features), and the processor can receive pending processing provided by the camera device Image, and then process it.

S102: Extract the region of interest of the image to be processed, and obtain the first parameter corresponding to the region of interest, wherein the circular feature is located in the region of interest, and the region of interest is a circle.

Specifically, in an application scenario, a variety of PEEK tools (eg, surface PEEK, line PEEK, and point PEEK) developed in the existing machine vision software sherlock can be used to select a region of interest (ROI). The user can input the coordinates of the center of the circle, the diameter of the outer circle and the diameter of the inner circle on the PEEK tool interface. The first parameter includes the first circle center coordinate, inner diameter, and outer diameter corresponding to the region of interest.

S103: Detect the circular feature using the first parameter to obtain the detection parameter corresponding to the circular feature.

Specifically, the detection parameters include detection center coordinates corresponding to the circular features and detection radius.

In an application scenario, as shown in FIG. 2, FIG. 2 is a schematic flowchart of an implementation manner of step S103 in FIG. 1. The above step S103 specifically includes:

S201: Convert the image of the region of interest in the rectangular coordinate system to the first image in the polar coordinate system, where the image of the region of interest contains circular features.

Specifically, the abscissa of the first image in the polar coordinate system is a polar diameter, and the ordinate is a polar angle. In this embodiment, the region of interest is a ring, the inner diameter of the ring is R ₁ , and the outer diameter is R ₂ , and the circular feature in the image to be detected is located within the ring. As shown in FIG. 3, FIG. 3a is a schematic structural diagram of an embodiment of an image of a region of interest in a rectangular coordinate system, and FIG. 3b is a diagram of an image of a region of interest in a rectangular coordinate system in FIG. A schematic diagram of an embodiment of an image and an embodiment. The rectangular coordinate system in FIG. 3a and the polar coordinate system in FIG. 3b have the same origin and positive direction. The origin and positive direction can be determined by the user. For example, as shown in FIG. 3b, the abscissa of this figure is the polar diameter r, and the ordinate is the polar angle θ; the origin is the upper left corner of the figure, the positive direction of the abscissa is right, and the ordinate is downward.

Assuming that in a rectangular coordinate system (or, image coordinate system), the coordinates of each point on the circular feature is I(x, y), converted to a polar coordinate system, the coordinates of each point is I'(r, θ). The outer diameter of the area of interest is R ₂ and the inner diameter is R ₁ , then I'(r,θ) satisfies the following relationship:

I'(r,θ)=I(rcosθ,rsinθ), where: r∈[0, R ₂ -R ₁ ], θ∈[0,359].

S202: Obtain a one-dimensional histogram corresponding to the first image, and a first gradient image corresponding to the first image and a second gradient image corresponding to the one-dimensional histogram.

In an application scenario, the above step S202 specifically includes: performing a first-order derivation on the first image to obtain the corresponding first gradient image; and obtaining a row of vector images by calculating the average value of the pixel gray levels in columns for the first image. The row vector image is a one-dimensional histogram; the first-order derivation is performed on the one-dimensional histogram to obtain a second gradient image. Of course, in other application scenarios, it is also possible to perform second-order derivation on the first image or the one-dimensional histogram. The second gradient image in the above step S202 is used to find the abrupt part of the image, and then obtain its edge information.

S203: Obtain corresponding first polar diameters at multiple extreme values in the second gradient image.

Specifically, in an application scenario, please refer to FIG. 4, which is a schematic flowchart of an implementation manner of step S203 in FIG. 2. The above step S203 specifically includes:

S301: Obtain the corresponding gradient value at the current position in the second gradient image in a predetermined order.

Specifically, the predetermined order may be from left to right or from right to left, or other prescribed order, which is not limited in this application.

S302: Determine whether the gradient value is an extreme value.

Specifically, the extreme value includes a maximum value and a minimum value; the above step S302 specifically includes: obtaining the first gradient value and the second gradient value corresponding to two pixel positions adjacent to the left and right at the current position; Whether the gradient value is greater than or equal to the first gradient value, and whether the gradient value at the current position is greater than or equal to the second gradient value, and if so, the gradient value at the current position is the maximum value; otherwise, it is further determined whether the gradient value at the current position Less than or equal to the first gradient value, and whether the gradient value at the current position is less than or equal to the second gradient value, and if so, the gradient value at the current position is a minimum value; otherwise, the gradient value at the current position is not an extreme value.

For example, suppose the polar diameter at the current position is r ₁ and the gradient value is T ₁ ; the polar diameters at the two adjacent pixel positions at the left and right are r ₁ -1, r ₁ +1, and r ₁ -1 and r ₁ +1 gradient values corresponding to T ₂ and T _3, when T ₁ ≥T _2, and T ₁ ≥T _3, the current value of the gradient polar radius r ₁ corresponding to a maximum value T _1; T ₁ ≤T ₂ when and T ₁ ≤T _3, the polar radius r ₁ corresponding to the current gradient value T ₁ is the minimum value. Otherwise, the current polar radius r ₁ corresponding to the gradient value T _1, not extremum.

S303: If yes, obtain the corresponding first pole diameter at the current position; otherwise, go directly to step S304;

S304: Determine whether the current position is the last position in the second gradient image, and if so, end, otherwise obtain the corresponding gradient value at the next position of the current position and return to step S302.

Specifically, when the corresponding polar diameter at the current position in the second gradient image is obtained from left to right, it can be determined whether the corresponding polar diameter value at the current position is smaller than the maximum polar diameter value in the second gradient image, If it is less, it is determined that the current position is not the last position in the second gradient image; if it is equal, it is determined that the current position is the last position in the second gradient image.

Of course, in other application scenarios, the above step S203 may also be implemented in other ways. For example, the step of "determining whether the current position is the last position in the second gradient image" in the above step S304 may also be located in "determining" in step S302 Before the step of whether the corresponding gradient value at the current position is an extreme value".

S204: Obtain a maximum value point within a predetermined range of each first polar diameter in the first gradient image, and a polar diameter and a polar angle corresponding to the maximum value point are defined as a first local polar diameter and a first polar angle.

Specifically, the second gradient image in step S203 is used to indicate where the edge information appears, and only the position range can be known, and the precise positioning cannot be obtained; and through the above step S204, the precise positioning can be performed from the first gradient image. The predetermined range may be artificially predetermined, for example, it may be 3 or 5 pixels wide, and the maximum point or the minimum point within the predetermined range and the corresponding value of the maximum point or the minimum point within the predetermined range may be obtained from the first gradient image The first local polar diameter and the first polar angle. For example, taking the first image in FIG. 3b as the first gradient image temporarily, r _{1 in the} figure is the first polar diameter obtained from the second gradient image, the middle of the two dotted lines is the predetermined range, and the point P in the figure is The highest point within the predetermined range.

S205: Obtain a second polar angle and a third polar angle corresponding to each first polar angle in the first gradient image, where the first polar angle, the second polar angle, and the third polar angle satisfy the first preset condition.

Specifically, in an application scenario, satisfying the first preset condition between the first polar angle, the second polar angle, and the third polar angle includes: the absolute value of the difference between the second polar angle and the first polar angle is 180° The absolute value of the difference between the third polar angle and the first polar angle is 90°. The three points that are not collinear can uniquely define a circle.

S206: Obtain multiple second polar diameters corresponding to the second polar angle from the first gradient image, and filter the multiple second polar diameters to retain at most one second polar diameter.

Specifically, in an application scenario, the process of obtaining multiple second polar diameters corresponding to the second polar angle from the first gradient image in step S206 includes: obtaining the corresponding second polar angle from the first gradient image For the row vector, search from the leftmost or rightmost side of the row vector to obtain multiple extreme points (the extreme points are edge points) appearing on the row vector, and then obtain multiple second extreme diameters.

In this embodiment, since the absolute value of the difference between the second polar angle and the first polar angle is 180°, when the region of interest and the circular feature are not at the same center, the point corresponding to the first polar angle is the circular feature The point closest to or farthest from the center of the circle of interest; the point corresponding to the second polar angle is the point farthest or closest to the center of the circle of interest. These two points exactly constitute the diameter of the circle. A point can also define a circle. The process of screening multiple second pole diameters in step S206 to retain at most one second pole diameter includes: matching the current second pole diameter and second pole angle with the first local pole diameter and first pole angle found previously Perform fitting to obtain a second fitted circle, and then determine whether the second fitted circle actually exists. If the extreme point corresponding to the second polar diameter and the second polar angle is a point on a circular feature, then the second The conjoint circle is real, and the second polar diameter remains; if the extreme point corresponding to the second polar diameter and the second polar angle is not a point on a circular feature, but a noise point or an interference point, then this second pseudo There is no roundness, and the second polar diameter is screened out. The above process will filter out multiple second polar diameters and only one will be left.

Please refer to FIG. 5. FIG. 5 is a schematic flowchart of an embodiment in which a plurality of second pole diameters are screened in step S206 in FIG. 2 to reserve at most one second pole diameter. In step S206, a plurality of second polar diameters are screened to retain at most one second polar diameter, that is, the method for specifically determining whether the second fitting circle actually exists includes:

S401: Fit all second polar diameters and second polar angles to the first local polar diameters and first polar angles, respectively, to obtain a plurality of second fitting circles, and a plurality of second fitting circles corresponding to the first Four parameters.

Specifically, in an application scenario, the first polar angle, the first local polar diameter, the second polar angle, and the second polar diameter can be converted into corresponding first coordinate values and second coordinate values in a rectangular coordinate system; The first coordinate value and the second coordinate value obtain a second fitted circle, and a fourth parameter corresponding to the second fitted circle in the rectangular coordinate system. The above-mentioned fourth parameter includes the coordinates and radius of the center of the circle corresponding to the second fitted circle in the rectangular coordinate system.

S402: Use the fourth parameter corresponding to the second fitted circle to obtain a first partial image within a first predetermined range from the second fitted circle from the region of interest.

Specifically, the first predetermined range may be set by the user, and may generally be 1-10 (for example, 1, 3, 5, 7, 10, etc.) pixel widths;

S403: Obtain a third gradient image corresponding to the first partial image.

Specifically, in an application scenario, the first partial image can be converted into a second image in a polar coordinate system; the second image can be averaged by the pixel gray scale in columns to obtain a corresponding one of the second image Dimensional histogram, and first-order derivation of the one-dimensional histogram to obtain its corresponding third gradient image.

S404: Obtain the first contrast of the second fitted circle through the third gradient image.

Specifically, the first contrast can be obtained by any one of the prior art. For example, the first contrast is the maximum value of the absolute value of gradient values in the third gradient image.

S405: Use the first contrast to score the second fitted circle, the first contrast is proportional to the first score.

Specifically, in an application scenario, the first contrast can be multiplied by a scale factor to obtain the first score; of course, in other application scenarios, the first score can also be obtained in other ways, which is not limited in this application .

S406: Obtain a second polar diameter corresponding to the second fitting circle whose first score exceeds the first threshold and whose score is the highest.

Specifically, the first threshold can be set by the user.

S207: Obtain multiple third pole diameters corresponding to the third pole angle from the first gradient image, and filter the multiple third pole diameters to retain at most one third pole diameter.

Specifically, in an application scenario, the process of obtaining multiple third polar diameters corresponding to the third polar angle from the first gradient image in step S207 includes: obtaining the third polar angle corresponding to the third polar angle in the first gradient image For the row vector, search from the leftmost or rightmost side of the row vector to obtain multiple extreme points (the extreme points are edge points) appearing on the row vector, and then obtain multiple third extreme diameters.

In this embodiment, since the absolute value of the difference between the third polar angle and the first polar angle is 90°, the process of screening multiple third polar diameters in step S207 to retain at most one third polar diameter includes: Fit the current third polar diameter and third polar angle to the first local polar diameter, first polar angle, second polar diameter, and second polar angle found previously to obtain a third fitted circle, and then determine the third Whether the fitted circle really exists, if the extreme point corresponding to the third pole diameter and the third pole angle is the point on the circular feature, then the third fitted circle is real, and the third pole diameter remains; if The extreme point corresponding to the third polar diameter and the third polar angle is not a point on the circular feature, but is a noise point or an interference point. Then this third fitting circle does not exist, and the third polar diameter is filtered out. The above process will screen out multiple third pole diameters, leaving at most one.

Please refer to FIG. 6. FIG. 6 is a schematic flowchart of an embodiment of screening a plurality of third pole diameters in step S207 in FIG. 2 to reserve at most one third pole diameter. In step S207, a plurality of third pole diameters are screened to reserve at most one third pole diameter, that is, the method for specifically determining whether the third fitting circle actually exists includes:

S501: Fit all third polar diameters and third polar angles to the first local polar diameters, first polar angles, and the second polar diameters and second polar angles screened and retained after step S206 to obtain multiple Three fitted circles, and fifth parameters corresponding to multiple third fitted circles.

Specifically, in an application scenario, the current third polar diameter, the third polar angle and the first local polar diameter, the first polar angle, the second polar diameter retained after the screening in step S206, and the second polar angle can be converted into Corresponding first coordinate value, second coordinate value, and third coordinate value in a rectangular coordinate system; according to the first coordinate value, second coordinate value, and third coordinate value, a third fitted circle is obtained, and the third fitted circle is in The corresponding fifth parameter in the Cartesian coordinate system. The above-mentioned fifth parameter includes the coordinate and radius of the center of the circle corresponding to the third fitted circle in the rectangular coordinate system.

S502: Using the fifth parameter corresponding to the third fitted circle, obtain a second partial image within a second predetermined range from the third fitted circle from the region of interest.

Specifically, the second predetermined range may be set by the user, and may generally be 1-10 (for example, 1, 3, 5, 7, 10, etc.) pixel widths;

S503: Obtain a fourth gradient image corresponding to the second partial image.

Specifically, in an application scenario, the second partial image can be converted into a third image in a polar coordinate system; the third image can be averaged by the pixel gray scale in columns to obtain a corresponding one of the third image Dimensional histogram, and first-order derivation of the one-dimensional histogram to obtain its corresponding fourth gradient image.

S504: Obtain the second contrast of the third fitted circle from the fourth gradient image.

Specifically, the second contrast can be obtained by any method in the prior art. For example, the second contrast is the maximum value of the absolute value of the gradient value in the fourth gradient image.

S505: Use the second contrast to score the third fitted circle, and the second contrast is proportional to the second score.

Specifically, in an application scenario, the second contrast can be multiplied by a scale factor to obtain a second score; of course, in other application scenarios, the second score can also be obtained in other ways, which is not limited in this application .

S506: Obtain a third polar diameter corresponding to the third fitting circle whose second score exceeds the second threshold and whose second score is the highest.

Specifically, the second threshold can be set by the user.

So far, after the above step S207, one first local pole diameter corresponds to one second pole diameter and one third pole diameter.

S208: obtaining a plurality of first fitting circles corresponding to the selected first polar angle, first local polar diameter, second polar angle, second polar diameter, third polar angle, and third polar diameter after screening, and Multiple third parameters corresponding to multiple first fitting circles.

Specifically, in an application scenario, the above step S208 specifically includes: obtaining a third fitting circle corresponding to all the first polar angles and the first local polar diameter retaining the third polar diameter, and the first fitting circle is the third Fitting a circle, the third parameter is the fifth parameter. The above-mentioned third parameters include the coordinates and radius of the center of the circle corresponding to the first fitted circle in the rectangular coordinate system.

S209: Use a plurality of first fitted circles, a plurality of third parameters, and a second predetermined strategy to obtain detection parameters.

Specifically, in an application scenario, the above step S209 specifically includes: obtaining a circle with the highest score, or a circle with the largest radius, or a circle with the smallest radius among the plurality of first fitting circles, and the third parameter corresponding to the circle As a detection parameter.

S104: Adjust the first parameter using the detection parameter, and return to the step of detecting the circular feature using the first parameter, and record the detection parameter obtained after each detection until the distance between the first center coordinate and the detected center coordinate is less than or equal to Preset threshold.

Specifically, in an application scenario, using the detection parameter to adjust the first parameter of the region of interest in step S104 above specifically includes: using the center coordinates of the detection in the detection parameters obtained from the latest detection as the first center coordinates of the region of interest To adjust the position of the region of interest. The purpose of this step is equivalent to adjusting the position of the region of interest so that the center of the region of interest coincides with the center of the circular feature as much as possible. Of course, in other application scenarios, while adjusting the position of the region of interest, the size of the region of interest can be further adjusted, for example, the inner diameter and the outer diameter corresponding to the region of interest can be adjusted according to the radius in the detection parameter obtained from the last detection To reduce the area of interest. It should be noted that when adjusting the size of the region of interest, it is necessary to ensure that the circular features are within the range of the region of interest.

In addition, in this embodiment, after the first parameter is adjusted by using the detection parameter, the detection method provided by the present application further includes: obtaining the distance between the first circle center coordinate and the detection circle center coordinate, and judging whether the distance is less than or equal to a preset threshold ; If yes, go to the subsequent step S105; otherwise, return to step S103. Among them, the preset threshold mentioned above can be set by the user.

S105: Determine the final parameters of the circular feature according to the recorded multiple detection parameters and the first predetermined strategy, and the final parameters include the center coordinates and radius of the circular feature.

Specifically, in an application scenario, the above step S105 includes: using the average value of multiple detection parameters as the final parameter of the circular feature, for example, the center coordinates of the circular feature are equal to the average of all detected center coordinates, and the radius of the center feature Equal to the average of all detection radii.

Please refer to FIG. 7, which is a schematic flowchart of another embodiment of a circular feature detection method according to the present application. The detection method includes:

S601: Receive an image to be processed, the image to be processed contains circular features. Specifically, this step is the same as step S101 in the foregoing embodiment, and will not be repeated here.

S602: Extract the region of interest of the image to be processed, and obtain the first parameter corresponding to the region of interest, where the circular feature is located in the region of interest, and the region of interest is a circle, and the first parameter includes the first parameter corresponding to the region of interest A circle center coordinate, inner diameter and outer diameter. Specifically, this step is the same as step S102 in the foregoing embodiment, and will not be repeated here.

S603: Detect the circular feature by using the first parameter to obtain a detection parameter corresponding to the circular feature. The detection parameter includes the detection center coordinate and the detection radius. Specifically, this step is the same as step S103 in the foregoing embodiment, and details are not described herein again.

S604: Adjust the first parameter using the detection parameter, and return to the step of detecting the circular feature using the first parameter, record the detection parameter obtained after each detection, and update the value of the iteration parameter after each detection until the iteration The value of the parameter is outside the preset range.

Specifically, the first parameter of the region of interest adjusted by the detection parameter in step S604 is the same as the related content in step S104 in the foregoing embodiment, and details are not described herein again.

In this embodiment, the iteration parameter is the number of times that the first parameter is used to detect the circular feature. If the value of the iteration parameter exceeds the preset range, the number of detections is greater than or equal to the preset value, or less than or equal to the preset value. The initial value of the iteration parameter can also be an arbitrary positive integer. The iteration parameter can be increased by a set value after each detection until the value of the iteration parameter is greater than or equal to the preset value; of course, the iteration parameter can also be subtracted by one after each detection Set the value until the value of the iteration parameter is less than or equal to the preset value.

In other embodiments, the iteration parameter may also be an angle value on the circular feature that uses the first parameter to detect the circular feature. The value of the iteration parameter that exceeds the preset range is that the angle value is greater than or equal to the preset angle, or less than Equal to the preset angle. The initial value of the iteration parameter can be an arbitrary angle, the iteration parameter can be increased by a set value after each detection until the value of the iteration parameter is greater than or equal to the preset angle; of course, the iteration parameter can also be subtracted by a setting after each detection Value until the value of the iteration parameter is less than or equal to the preset angle. For example, the initial value of the iteration parameter may be 0°, the set value may be 45°, and the preset angle may be 360°.

In another embodiment, after the first parameter is adjusted using the detection parameter, the detection method provided by the present application further includes: judging whether the value of the current iteration parameter exceeds the preset range; if yes, proceed to the subsequent step S605; otherwise, return In step S603, the value of the iteration parameter is updated (for example, a set value is added or subtracted).

S605: Determine the final parameters of the circular feature according to the recorded multiple detection parameters and the first predetermined strategy, and the final parameters include the center coordinates and radius of the circular feature. Specifically, this step is the same as step S105 in the foregoing embodiment, and will not be repeated here.

In another embodiment, the detection methods shown in FIGS. 1 and 7 of the present application can also be combined. For example, the method in FIG. 1 can be performed first. The method in FIG. 1 is equivalent to the process of roughly searching for a circular feature. At first, the circular feature and the region of interest may not be concentric or severely eccentric. The method in FIG. 1 can make the sense The distance between the first center coordinates corresponding to the region of interest and the detection center coordinates of the circular feature is less than or equal to the preset threshold, thereby making the circular feature substantially concentric with the moved region of interest; then, the method in FIG. 7 can be performed. The method in 7 is equivalent to the process of finely searching for circular features. Iterative parameters control the number of circular feature detections, thereby ensuring the accuracy and stability of circular features.

In practical applications, circular features include noise-free standard circular features, noisy standard circular features, and non-standard circular features (ie, irregular circular features). For the noise-free standard circular feature, because it has high-quality edge information, it can accurately calculate the edge detection parameters after two or three iterations, and the accuracy can reach 0.03 pixels; for the noisy standard For circular features, due to the presence of noise, there is usually a case of oscillation and non-convergence in the detection of circular features. Using multiple iterations and then averaging can well avoid the influence of noise and ensure the accuracy and stability of the final parameters For non-standard circular features, it is theoretically impossible to give its parameters (because it is a non-circular feature, so there is no center, no radius and other parameters), but by comprehensively detecting and calculating it, it can be given Its average center coordinates and radius parameters are very useful for practical applications, because many of the circular features in the actual picture are non-standard circular features. In addition, the iterative idea in the circular feature detection method provided by this application can also be used in the detection process of other specific shapes, such as straight lines, etc., which will not be described in detail in this application.

Please refer to FIG. 8, which is a schematic structural diagram of an embodiment of a circular feature detection and processing system of the present application. The circular feature detection and processing system 1 includes a processor 10, a memory 12, a transceiver 14, and a display 16. The transceiver 14 is used to receive an image to be processed, and transmits the image to be processed to the processor 10, and the display 16 is used to display detection As a result, of course, the display 16 can also display the image to be processed and the region of interest, and the memory 12 stores program instructions, which can be loaded by the processor 10 and execute the circular feature detection method in any of the above embodiments.

Please refer to FIG. 9. FIG. 9 is a schematic structural diagram of an embodiment of a device with a storage function according to the present application. The device 2 with a storage function stores program data 20. When the program data 20 is executed by a processor, any of the above embodiments is implemented. Steps in the circular feature detection method.

In a word, different from the situation in the prior art, the circular feature detection method provided by the present application includes: using the first parameter corresponding to the region of interest to obtain the detection parameter corresponding to the circular feature after detection; and using the detection parameter to adjust the first Parameters, and return to the step of detecting the circular features using the first parameter, recording the detection parameters obtained after each detection, until the distance between the first center coordinates and the detected center coordinates is less than or equal to the preset threshold; according to the number of records The detection parameters and the first predetermined strategy determine the final parameters of the circular feature. The final parameters include the center coordinates and radius of the circular feature. On the one hand, the above-mentioned iterative method can make the region of interest closer to the circular feature, and the center of the region of interest closer to the circular feature, so as to reduce the randomness of the region of interest and improve the circular feature The accuracy of the detection results; on the other hand, the method of using multiple detection parameters and the first predetermined strategy to obtain the final parameters of the circular feature in the above method can give a more average result and further improve the detection result of the circular feature Accuracy and stability, in an application scenario, the circle detection results can reach a few pixels.

Claims

A circular feature detection method, characterized in that the circular feature detection method includes:

Receiving an image to be processed, the image to be processed includes a circular feature;

Extract the region of interest from the image to be processed, and obtain the first parameter corresponding to the region of interest, wherein the circular feature is located in the region of interest, and the region of interest is a circle, so The first parameter includes the first circle center coordinate, inner diameter, and outer diameter corresponding to the region of interest;

Detecting the circular feature by using the first parameter to obtain a detection parameter corresponding to the circular feature, the detection parameter including a detection center coordinate and a detection radius;

Adjust the first parameter using the detection parameter, and return to the step of detecting the circular feature using the first parameter, and record the detection parameter obtained after each detection until the first circle center The distance between the coordinates and the coordinates of the detected circle center is less than or equal to a preset threshold;

The final parameters of the circular feature are determined according to the multiple recorded detection parameters and the first predetermined strategy, and the final parameters include the center coordinates and radius of the circular feature.
The detection method according to claim 1, wherein the adjusting the first parameter using the detection parameter comprises:

The detected center coordinates of the detection parameters obtained in the latest detection are used as the first center coordinates of the region of interest to adjust the position of the region of interest.
The detection method according to claim 2, wherein the adjusting the first parameter using the detection parameter further comprises:

The inner diameter and the outer diameter corresponding to the region of interest are adjusted according to the detection radius in the detection parameter obtained in the latest detection, so that the range of the region of interest is reduced.
The detection method according to claim 1, wherein the detecting the circular feature using the first parameter to obtain a detection parameter corresponding to the circular feature includes:

Converting the image of the region of interest in the rectangular coordinate system into the first image in the polar coordinate system, wherein the image of the region of interest contains the circular feature, and the abscissa of the first image is polar Diameter, the ordinate is the polar angle;

Obtaining a one-dimensional histogram corresponding to the first image, a first gradient image corresponding to the first image, and a second gradient image corresponding to the one-dimensional histogram;

Obtaining a first polar diameter corresponding to multiple extreme values in the second gradient image;

Obtaining a maximum value point within a predetermined range of each first polar diameter in the first gradient image, and a polar diameter and a polar angle corresponding to the maximum value point are defined as a first local polar diameter and a first polar angle;

Obtaining a second polar angle and a third polar angle corresponding to the first polar angle in the first gradient image, wherein the first polar angle, the second polar angle, and the third polar angle satisfy The first preset condition;

Obtaining a plurality of second polar diameters corresponding to the second polar angle in the first gradient image, and filtering the plurality of second polar diameters to retain at most one second polar diameter;

Obtaining a plurality of third polar diameters corresponding to the third polar angle in the first gradient image, and filtering the plurality of third polar diameters to retain at most one third polar diameter;

Obtain multiple sets of multiple first corresponding to the first polar angle, the first local polar diameter, the second polar angle, the second polar diameter, the third polar angle, and the third polar diameter A fitting circle, and a plurality of third parameters corresponding to the plurality of first fitting circles;

The detection parameters are obtained by using a plurality of the first fitted circles, a plurality of third parameters, and a second predetermined strategy.
The detection method according to claim 4, wherein the obtaining the first polar diameter corresponding to a plurality of extreme values in the second gradient image includes:

Obtaining the corresponding gradient value at the current position in the second gradient image in a predetermined order;

Determine whether the gradient value is an extreme value;

If yes, obtain the corresponding first pole diameter at the current position and determine whether the current position is the last position in the second gradient image; if so, end the obtaining of multiple poles in the second gradient image The step corresponding to the first polar diameter at the value; otherwise, obtaining the corresponding gradient value at the next position of the current position, and returning to the step of determining whether the gradient value is an extreme value;

Otherwise, it is determined whether the current position is the last position in the second gradient image, and if so, the step of obtaining corresponding first polar diameters at multiple extreme values in the second gradient image is ended; otherwise, Obtain the corresponding gradient value at the next position of the current position, and return to the step of determining whether the gradient value is an extreme value.
The detection method according to claim 5, wherein the extreme value includes a maximum value and a minimum value, and the judging whether the gradient value is an extreme value includes:

Obtain the first gradient value and the second gradient value corresponding to the two adjacent pixel positions at the current position;

Judging whether the gradient value at the current position is greater than or equal to the first gradient value, and whether the gradient value at the current position is greater than or equal to the second gradient value;

If yes, the gradient value at the current position is a maximum value;

Otherwise, it is further determined whether the gradient value at the current position is less than or equal to the first gradient value, and whether the gradient value at the current position is less than or equal to the second gradient value;

If yes, the gradient value at the current position is a minimum value;

Otherwise, the gradient value at the current position is not an extreme value.
The detection method according to claim 4, wherein the first polar angle, the second polar angle, and the third polar angle satisfy a first preset condition, including:

The absolute value of the difference between the second polar angle and the first polar angle is 180°, and the absolute value of the difference between the third polar angle and the first polar angle is 90°.
The detection method according to claim 7, wherein the screening of the plurality of second pole diameters to retain at most one second pole diameter includes:

Fitting all the second polar diameters and second polar angles to the first local polar diameters and the first polar angles, respectively, to obtain a plurality of second fitting circles, and a plurality of second fitting circles corresponding The fourth parameter, the fourth parameter includes the coordinates and radius of the center of the circle corresponding to the second fitted circle;

Using the fourth parameter corresponding to the second fitted circle to obtain a first partial image within a first predetermined range from the second fitted circle from the region of interest;

Obtain a third gradient image corresponding to the first partial image;

Obtaining the first contrast of the second fitted circle through the third gradient image;

Scoring the second fitted circle using the first contrast, the first contrast is proportional to the first score;

A second polar diameter corresponding to the second fitted circle with the first score exceeding the first threshold and the highest first score is obtained.
The detection method according to claim 8, wherein the screening of the plurality of third pole diameters to retain at most one third pole diameter includes:

Fitting all third polar diameters and third polar angles to the first local polar diameters, the first polar angles, the second polar diameters and the second polar angles selected by screening to obtain Multiple third fitting circles, and fifth parameters corresponding to multiple third fitting circles;

Using the fifth parameter corresponding to the third fitted circle to obtain a second partial image within a second predetermined range from the third fitted circle from the region of interest;

Obtaining a fourth gradient image corresponding to the second partial image;

Obtaining the second contrast of the third fitted circle through the fourth gradient image;

Scoring the third fitted circle using the second contrast, the second contrast is proportional to the second score;

A third polar diameter corresponding to a third fitted circle with the second score exceeding the second threshold and the highest second score is obtained.
The detection method according to claim 9, wherein the obtained multiple sets of first polar angle, first local polar diameter, second polar angle, second polar diameter, third polar angle, third polar diameter correspond The plurality of first fitted circles and the plurality of third parameters corresponding to the plurality of first fitted circles include:

Obtaining the third fitted circle corresponding to all the first polar angles and the first local polar diameters retaining the third polar diameter, the first fitted circle being the third fitted circle, so The third parameter is the fifth parameter.
The detection method according to claim 10, wherein the use of a plurality of first fitted circles, a plurality of third parameters, and a second predetermined strategy to obtain the detection parameters includes:

Obtain the circle with the highest second score, the circle with the largest radius, or the circle with the smallest radius among the plurality of first fitting circles, and use the third parameter corresponding to the circle as the detection parameter.
The detection method according to claim 8, wherein all the second polar diameters and second polar angles are fitted to the first local polar diameters and the first polar angles, respectively, to obtain multiple A second fitted circle, and the fourth parameters corresponding to a plurality of second fitted circles, including:

Converting the second polar diameter, the second polar angle, the first polar angle, and the first local polar diameter in the polar coordinate system to the corresponding first coordinate value and second coordinate value in the rectangular coordinate system;

Obtain the second fitted circle according to the first coordinate value and the second coordinate value, and the fourth parameter corresponding to the second fitted circle in the rectangular coordinate system.
The detection method according to claim 4, characterized in that

The obtaining a one-dimensional histogram corresponding to the first image includes: averaging pixel gray scales in columns of the first image to obtain the one-dimensional histogram;

The obtaining the first gradient image corresponding to the first image includes: performing first-order derivation on the first image to obtain the first gradient image;

The obtaining the second gradient image corresponding to the one-dimensional histogram includes: performing first-order derivation on the one-dimensional histogram to obtain the second gradient image.
The detection method according to claim 1, wherein the determining the final parameter of the circular feature based on the plurality of recorded detection parameters and the first predetermined strategy includes:

The average value of the multiple detection parameters is used as the final parameter of the circular feature.
A circular feature detection method, characterized in that the circular feature detection method includes:

Receiving an image to be processed, the image to be processed includes a circular feature;

Extract the region of interest from the image to be processed, and obtain the first parameter corresponding to the region of interest, wherein the circular feature is located in the region of interest, and the region of interest is a circle, so The first parameter includes the first circle center coordinate, inner diameter, and outer diameter corresponding to the region of interest;

Detecting the circular feature by using the first parameter to obtain a detection parameter corresponding to the circular feature, the detection parameter including a detection center coordinate and a detection radius;

Adjust the first parameter using the detection parameter, and return to the step of detecting the circular feature using the first parameter, record the detection parameter obtained after each detection, and after each detection Update the value of the iteration parameter until the value of the iteration parameter exceeds a preset range;

The final parameters of the circular feature are determined according to the multiple recorded detection parameters and the first predetermined strategy, and the final parameters include the center coordinates and radius of the circular feature.
The detection method according to claim 15, characterized in that

The iteration parameter is the number of times that the first parameter is used to detect the circular feature, and the value of the iteration parameter exceeds a preset range when the number of detections is greater than or equal to a preset value.
The detection method according to claim 15, wherein the adjusting the first parameter by using the detection parameter comprises: using the detection center coordinates of the detection parameters obtained in the most recent detection as the Coordinate of the first circle center corresponding to the region of interest to adjust the position of the region of interest.
The detection method according to claim 15, wherein the detecting the circular feature using the first parameter to obtain a detection parameter corresponding to the circular feature includes:

Converting the image of the region of interest in the rectangular coordinate system into the first image in the polar coordinate system, wherein the image of the region of interest contains the circular feature, and the abscissa of the first image is polar Diameter, the ordinate is the polar angle;

Obtaining a one-dimensional histogram corresponding to the first image, a first gradient image corresponding to the first image, and a second gradient image corresponding to the one-dimensional histogram;

Obtaining a first polar diameter corresponding to multiple extreme values in the second gradient image;

Obtaining a maximum value point within a predetermined range of each first polar diameter in the first gradient image, and a polar diameter and a polar angle corresponding to the maximum value point are defined as a first local polar diameter and a first polar angle;

Obtaining a second polar angle and a third polar angle corresponding to the first polar angle in the first gradient image, wherein the first polar angle, the second polar angle, and the third polar angle satisfy The first preset condition;

Obtaining a plurality of second polar diameters corresponding to the second polar angle in the first gradient image, and filtering the plurality of second polar diameters to retain at most one second polar diameter;

Obtaining a plurality of third polar diameters corresponding to the third polar angle in the first gradient image, and filtering the plurality of third polar diameters to retain at most one third polar diameter;

Obtain multiple sets of multiple first corresponding to the first polar angle, the first local polar diameter, the second polar angle, the second polar diameter, the third polar angle, and the third polar diameter A fitting circle, and a plurality of third parameters corresponding to the plurality of first fitting circles;

The detection parameters are obtained by using a plurality of the first fitted circles, a plurality of third parameters, and a second predetermined strategy.
A circular feature detection and processing system, characterized in that the circular feature detection and processing system includes a processor, a memory, a transceiver, and a display, and the transceiver is used to receive an image to be processed and transmit the image to be processed To the processor, the display is used to display the detection result, and the memory stores program instructions, which can be loaded by the processor and execute any one of claims 1-14 or claim 15 -18 Any one of the circular feature detection methods.
A device with a storage function, on which program data is stored, characterized in that, when the program data is executed by a processor, it is implemented according to any one of claims 1-14 or any one of claims 15-18 Steps in the circular feature detection method.