CN115641326A - Sub-pixel size detection method and system for ceramic antenna PIN needle image - Google Patents
Sub-pixel size detection method and system for ceramic antenna PIN needle image Download PDFInfo
- Publication number
- CN115641326A CN115641326A CN202211398987.8A CN202211398987A CN115641326A CN 115641326 A CN115641326 A CN 115641326A CN 202211398987 A CN202211398987 A CN 202211398987A CN 115641326 A CN115641326 A CN 115641326A
- Authority
- CN
- China
- Prior art keywords
- pixel
- edge
- sub
- image
- pin
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a sub-pixel size detection method for a PIN image of a ceramic antenna. The industrial camera is calibrated by using a Zhang calibration method, ROI (region of interest) areas of PIN images are obtained by using a deep learning framework, image preprocessing is carried out on each ROI area, the preprocessed ROI areas are subjected to image analysis and identification to obtain the pixel width of the PIN in the ROI areas, and the diameter size of the detected PIN in the ROI areas is obtained through conversion according to the pixel width of the PIN. The system and method of the present invention enable accurate measurement of PIN dimensions in a contactless condition.
Description
Technical Field
The disclosure belongs to the technical field of vision measurement, and relates to an image identification processing method and system for a needle-shaped structure, in particular to a sub-pixel size detection method and system for a PIN needle image of a ceramic antenna.
Background
With the rapid development and application of ceramic technology, the low-temperature co-fired ceramic technology rapidly becomes an important means for the processing of a new generation of microwave devices by virtue of the advantages of high assembly density, good frequency characteristics, high batch processing consistency and the like, so that the PIN needle of the ceramic antenna becomes a new generation of navigation antenna. The PIN needle of the GPS ceramic antenna is used as the most important receiving part in the GPS equipment and plays a role of being equivalent to an ear, so that the appearance size of the PIN needle is directly related to the product quality of a GPS whole machine.
At present, most manufacturers visually inspect the products through manual naked eyes and a micrometer and other tools, and the manual detection is slow, poor in reliability and easy to miss detection. Machine vision is a comprehensive discipline integrating digital image processing, mechanical engineering, electronic engineering, optical engineering, software engineering and other technologies, and the core of the machine vision is processing and analyzing acquired images. The dimension measuring method based on machine vision has the characteristics of non-contact, high speed and high precision, and meets the requirements of modern industrial production on the measuring speed and precision of PIN needle parts.
Disclosure of Invention
In order to solve the problems in the background art, the invention aims to provide a sub-pixel size detection method for a PIN image of a ceramic antenna, which is used for quickly and accurately measuring the size of the PIN of the GPS ceramic antenna under a non-contact condition.
In order to achieve the purpose, the technical scheme of the invention comprises the following steps:
1. a sub-pixel size detection method for a PIN needle image of a ceramic antenna comprises the following steps:
1) Calibrating the industrial camera by using a Zhang calibration method;
2) Acquiring an ROI (region of interest) of the PIN image by using a deep learning framework;
3) Carrying out image preprocessing on each ROI;
4) Carrying out image analysis and identification on the preprocessed ROI area to obtain the pixel width of the PIN needle at the ROI area;
5) And converting the diameter size of the PIN to be measured at the ROI area according to the pixel width of the PIN.
The PIN has a plurality of different stepped size shapes.
In specific implementation, the PIN needle is vertically placed in the image and is divided into a plurality of sections in the vertical direction, the diameter sizes of different sections are different, and adjacent sections are in transition connection.
The industrial camera is calibrated by using a Zhang calibration method, which comprises the step of calibrating a distortion matrix, pixel equivalent and an internal and external parameter matrix of the industrial camera by using a checkerboard calibration board with the unit length as the side length of a single lattice. The unit length is set to practically 1mm.
The specification of the checkerboard calibration plate is 16 × 20 grids, the side length of a single grid is 1mm, and the precision is 0.001mm.
The calibration of the distortion matrix refers to the distortion correction of a calibration image to obtain a calibration image and a distortion matrix;
the calibration of the pixel equivalent comprises the following steps: and searching and storing the angular points of the checkerboard in the calibration graph after the deformity correction, respectively obtaining the pixel width between two adjacent angular points in each line of angular points, and then obtaining the pixels contained in unit length by taking the average value as pixel equivalent pixelsPerMetric.
Embodied as a 16 by 20 checkerboard, in which a total of 15 by 18 pixel wide data is obtained.
The step 2) specifically comprises the following steps: the method comprises the steps of collecting a data set in advance for a PIN sample to be measured, pre-training a YooloX-s network by using a labelimg tool marking data set, then taking a PIN image collected by a current camera as input, and taking different measurement areas of the PIN output by a trained YooloX-s network model as ROI areas to obtain coordinate information of the ROI areas so as to distinguish a plurality of ROI targets in the PIN image.
The specific implementation is that different step sizes of the PIN needle are obtained to be different ROI areas, and one ROI area is formed by one area with a uniform diameter size.
The preprocessing of the step 3) comprises format conversion, gray level processing and Gaussian filtering smoothing processing which are sequentially carried out.
The step 4) is specifically as follows:
4.1 Carrying out sub-pixel edge detection based on curve fitting and improved Sobel operator to obtain a sub-pixel edge image;
4.2 Utilizing a probability Hough method to fit in the sub-pixel edge image to obtain parallel edge lines of the PIN needle, and calculating the space pixel width of the parallel edge lines as the pixel width of the PIN needle;
the 4.1) is specifically as follows:
4.1.1 Processing the image after the Gaussian filtering smoothing processing by using an improved Sobel operator to obtain a pixel-level edge:
firstly, acquiring the gradient amplitude and direction of an ROI (region of interest) by using an improved 8-direction Sobel operator set by the following formula, and establishing a gradient map of the ROI;
wherein S is 1 ～S 8 Respectively are direction templates with gradient directions of 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees and 315 degrees;
then, carrying out thinning edge and strong and weak edge connection on the gradient map of the ROI by using non-maximum suppression and double thresholds in a Canny method to obtain a pixel-level edge;
4.1.2 ) extracting sub-pixel edges of the image by processing the pixel-level edges by using a curve fitting method:
traversing edge points on each pixel level edge, selecting 30 edge points on the gradient map along the gradient direction of the current edge point on the pixel level edge, and performing Gaussian curve fitting sub-pixel processing by adopting a Gaussian function of the following formula to obtain a Gaussian curve, wherein the Gaussian function is described as follows:
wherein i =1,2, 3.., 20, x i Is the abscissa position, z, of the ith edge point i Is the gradient magnitude of the ith edge point, z max 、x max And L is the peak value of the gradient amplitude, the abscissa position and the half-width information of the peak value respectively;
then finding the position with the maximum gradient in the Gaussian curve as the edge point on the edge of the sub-pixel by the least square method, namely solving the peak value z of the gradient amplitude in the gradient direction max And peak abscissa position x max And then according to the peak value z of the gradient amplitude max And peak abscissa position x max Solving the vertical coordinate position y of the peak value by combining the gradient direction max Thereby determining the edge point coordinate (x) of the current edge point corresponding to the sub-pixel edge max ,y max )。
4.1.3 The edge points corresponding to all the edge points on the sub-pixel edge are obtained after traversing each edge point on the pixel-level edge, and finally the sub-pixel edge is formed.
The step 4.2) is specifically as follows: the sub-pixel edge is input into a probability Hough method for processing, converting and extracting to obtain each straight line segment, and each straight line segment returns coordinates (x 1, y 1), (x 2, y 2) of two end points; calculating the slope of all the extracted straight line segments, taking two straight line segments with the same or the closest slope as parallel edge lines on two sides of the PIN needle, namely finishing the extraction of the parallel edge lines, and calculating the pixel distance between the two straight line segments as the pixel width of the PIN needle.
Specifically, the step 5) is to convert the pixel width of the PIN needle according to the pixel equivalent pixelPerMetric to obtain the actual diameter size of the measured PIN needle at the ROI area.
2. A sub-pixel size detection system for ceramic antenna PIN for implementing a method, the system applying machine vision in PIN measurement:
the image acquisition and display module is used for acquiring an image of the tested PIN, preprocessing the image and displaying a measurement result;
the ROI extraction module is used for acquiring an ROI (region of interest) of the PIN needle and rejecting image redundant information by the deep learning framework;
the camera calibration module is used for calibrating an internal and external parameter matrix, a distortion matrix and pixel equivalent of the industrial camera by a Zhang calibration method;
the measuring module is used for sub-pixel edge detection, fitting parallel edge lines of the part in a sub-pixel edge image by utilizing a probability Hough method, and calculating the distance between the parallel edge lines;
and the parameter monitoring and storing module is used for tracking the measuring progress in real time, displaying key parameters and locally storing the parameters.
Compared with the prior art, the above one or more technical schemes have the following beneficial effects:
the technical scheme includes that the industrial camera is used for carrying out size measurement of the PIN needle, the traditional image processing and deep learning technology is combined to carry out ROI area detection and size measurement to be measured on the PIN needle image collected by the camera, and measurement data are compared and analyzed according to the standard requirements and acceptance standards of actual production of a factory to obtain a PIN needle size measurement result.
According to the technical scheme, the camera is calibrated by using a high-precision checkerboard through a Zhang Zhengyou calibration method, and the pixel equivalent calibration method based on the checkerboard is provided, so that distortion correction of an image in normal calibration work can be completed, and the pixel equivalent can be determined to complete the measurement task of the invention.
The technical scheme disclosed improves the traditional Sobel operator, constructs the template in 8 directions on the basis of only vertical and horizontal directions, and can detect the edge gradient more accurately. In addition, the sub-pixel edge positioning based on Gaussian curve fitting enables the method to further improve the precision from the aspect of the method, and the method has the advantages of simplicity, effectiveness, low equipment cost and the like.
The technical scheme disclosed by the invention realizes the fitting of the buses on the two sides of the PIN needle by utilizing probability Hough transformation, and realizes the measurement of the diameter of the PIN needle. Aiming at the problem in the straight line detection, the parallel edge line detection which is more efficient in the diameter measurement task is provided.
Drawings
Fig. 1 is a flowchart of a method for detecting a diameter size of a PIN in an embodiment of the present disclosure.
Fig. 2 is a flowchart of calibrating internal and external parameters of an industrial camera according to an embodiment of the disclosure.
Fig. 3 is a statistical chart of pixel data between adjacent grids of checkerboard pictures before and after correction according to an embodiment of the disclosure.
Fig. 4 is a graph illustrating the segmentation effect of the well-trained Yolox network target in the embodiment of the present disclosure.
FIG. 5 is a flowchart of a method for measuring dimensions in an embodiment of the disclosure.
Fig. 6 is a diagram of improved Sobel operator single edge detection in an embodiment of the present disclosure.
Fig. 7 is a process diagram for capturing sub-pixel edges in an embodiment of the disclosure.
Fig. 8 is a diagram of the fitting effect of the probability Hough method in the embodiment of the present disclosure.
Fig. 9 is a diagram of the final effect in the embodiment of the present disclosure.
Fig. 10 is experimental comparison data of a pixel-level detection system and a sub-pixel-level detection system in an embodiment of the disclosure.
Fig. 11 is a software interface of a PIN size measuring upper computer in the embodiment of the present disclosure.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
The present invention will be described in detail with reference to the specific embodiments shown in the drawings, but the embodiments are not limited to the present invention, and structural, methodological, or functional changes made by those skilled in the art according to the embodiments are included in the scope of the present invention.
One embodiment of the present invention as shown in fig. 1 comprises:
in a specific implementation example, the most classical labeling method of Zhang Zhengyou is selected for the camera calibration method. The specification of the checkerboard was determined as 16 by 20 cells, with individual cells having a side length of 1mm and an accuracy of 0.001mm.
The calibration process is shown in fig. 2, and an internal parameter matrix and an external parameter matrix of the camera are obtained by taking 20 checkerboard pictures at different angles and running a calibration program. According to the internal parameter matrix and the external parameter matrix, a world coordinate system (X) can be completed w ,Y w ,Z w ) To camera coordinate system (X) c ,Y c ,Z c ) And the image coordinate system is changed to (X) p ,Y p ,Z p ) A transformation into a pixel coordinate system (u, v) in the following relationship:
and the distortion correction can be carried out on the image acquired by the camera by the obtained internal and external parameter matrixes.
On the basis of a conventional Zhang calibration method, in order to cater to the measurement task of the invention, a checkerboard-based pixel equivalent calibration method is provided. Searching for coordinates of chessboard angular points in the collected chessboard pattern and storing the coordinates into a Corners matrix; respectively obtaining: pixel width between two adjacent corners in each row of corners:
wherein d is pixel [i]Represents the pixel distance from the (i + 1) th corner point to the (i) th corner point,
the abscissa representing the ith corner point,
indicating the ordinate of the i-th corner point.
A total of 15 x 18 data were obtained in a checkerboard, and the pixel equivalents pixel equivalent pixel permametric, which were actually contained in 1mm, were obtained by averaging. The statistics of the pixel data between adjacent lattices in the pre-and post-correction pictures are shown in fig. 3 (a) to 3 (b). It can be seen that the fluctuation range of the pixels after correction is significantly reduced, while the statistical mean before correction is 139.05, the variance is 0.570, and the standard deviation is 0.756, and the mean after correction is 135.65, the variance is 0.073, and the standard deviation is 0.272. Thus, pixel equivalent pixelspermeric =135.65 was determined.
The method comprises the steps of collecting a data set in advance for a PIN sample to be measured, labeling the data set by using a labellimg tool, taking a PIN sample image as input, and training the PIN image sample by using a YoloX-s network. The target segmentation effect obtained by inputting the image acquired by the current camera into the YoloX-s network for prediction is shown in fig. 4 (a) -4 (b).
After the object segmentation is performed on the measurement area of the PIN, the size measurement method is performed on the obtained object, and the detailed flow of the method is shown in fig. 5.
Image preprocessing and sub-pixel edge detection based on curve fitting and improved Sobel operator: converting an input image format into an image format suitable for opencv, and performing graying processing and Gaussian filtering on the image; acquiring the gradient amplitude and the direction of the preprocessed image by using an improved 8-direction Sobel operator shown in formula (1);
wherein S is 1 ～S 8 Respectively, are directional templates with gradient directions of 0 °,45 °,90 °,135 °,180 °,225 °,270 °,315 °.
Then, the classical non-maximum suppression and dual threshold in the Canny method are applied to refine the edges and connect strong and weak edges to obtain the pixel level edges, as shown in fig. 6.
Selecting 30 points on the edge of a single pixel along the gradient direction of the target edge point to perform Gaussian curve fitting sub-pixel calculation, wherein the Gaussian function is described as follows:
wherein i =1,2, 3.., 20, x i Is the abscissa position of the ith point, z i Gradient amplitude of the ith point, z max ，x max And L are gradient amplitude peak, peak position and half-width information, respectively. (2) taking natural logarithms on two sides of the formula, and converting into:
in matrix form:
wherein w i ＝lnz i ，
The above equation is abbreviated as W = XF, and according to the least squares principle, the least squares solution of the matrix F is:
F＝(X T X) -1 X T W (5)
the peak z of the gradient amplitude in the gradient direction can be obtained from the equations (5) and (3) max And peak abscissa position x max Then, the vertical coordinate position y of the peak value is obtained according to the gradient direction max I.e. the sub-pixel edge can be determinedEdge coordinate (x) max ,y max )。
The sub-pixel edge map can be roughly and intuitively displayed by magnifying the sub-pixel coordinates by ten times and rounding, then the pixel sub-pixel edge map and the sub-pixel level sub-pixel edge map are zoomed to the same size, and the process of obtaining the sub-pixel edge by taking the edge part is shown in fig. 7.
And (4) inputting the sub-pixel edge into probability Hough transformation processing to extract and obtain each straight line segment.
Calculating the slope k in the extracted straight line, and reserving and outputting each pair of straight lines with the slope error smaller than error =0.01, namely completing the extraction of the parallel edge lines, wherein the fitting effect graph of the probability Hough method is shown in FIG. 8.
Each group of parallel edge lines returns four end point coordinates (x) of two Line segments Line1 and Line2 1 ,y 1 )，(x 2 ,y 2 )，(x 3 ,y 3 )，(x 4 ,y 4 ) And the end point of one Line segment Line1 is coordinated into a straight Line equation form: ax + By + C =0.
Wherein:
then according to a dot-line spacing formula:
respectively calculating the pixel distance N3 from (x 3, y 3) to Line1 and the pixel distance N4 from (x 4, y 4) to Line1, then averaging to obtain the pixel distance N between the extracted parallel edge lines, and calculating the actual diameter size d = N/pixel PerMetric mm of the part of the PIN needle according to the pixel equivalent pixel PerMetric.
The final effect graph of the overall detection is shown in fig. 9.
In order to prove the advantages and effects of the invention, the embodiment also utilizes the edge detection algorithm of the common pixel level to detect the PIN size by combining the idea of the invention, and carries out a comparison experiment with the system. The comparative data is shown in fig. 10, where fig. 10 (a) is experimental data of a pixel-level detection system, and fig. 10 (b) is experimental data of a sub-pixel-level detection system. According to the comparison experiment data, the error result obtained by comparing the measured value of the PIN needle diameter of the pixel-level detection system with the measured value of the micrometer screw with the precision of 0.001mm is within 0.02mm, and the error of the sub-pixel-level detection system is improved to be within 0.006 mm.
The embodiment also discloses a sub-pixel size detection system for the ceramic antenna PIN needle, wherein the software interface of the PIN needle size measurement upper computer is shown in fig. 11. The system comprises:
the image acquisition and display module is used for acquiring an image of the tested PIN, preprocessing the image and displaying a measurement result;
the ROI extraction module is used for acquiring ROI regional rejection image redundant information of the PIN needle by the deep learning framework;
the camera calibration module is used for calibrating an internal and external parameter matrix, a distortion matrix and pixel equivalent of the industrial camera by a Zhang calibration method;
the measuring module is used for sub-pixel edge detection, fitting parallel edge lines of a part in a sub-pixel edge image of a region to be detected by utilizing a probability Hough method, and calculating distance information between the parallel edge lines;
and the parameter monitoring and storing module is used for tracking the measuring progress in real time, displaying key parameters and locally storing the parameters.
Through the above detailed implementation, the visual inspection of the size of the PIN needle of the GPS ceramic antenna can be conveniently carried out, the precision of the visual inspection reaches 0.006mm, and the detection requirement is met. The method is suitable for the on-line high-precision detection of the PIN size of the GPS ceramic antenna and has important application value. The stability of visual detection of the PIN needle of the GPS ceramic antenna is improved, the manual detection cost is saved, and the detection efficiency and accuracy are improved.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
Although the embodiments of the present disclosure have been described with reference to the accompanying drawings, the scope of the present disclosure is not limited thereto
By way of limitation, it should be apparent to those skilled in the art that the present disclosure may be practiced without these specific details
Various modifications or alterations that require inventive activity may be made while remaining within the scope of the present disclosure.
Claims (10)
1. A sub-pixel size detection method for a PIN image of a ceramic antenna, which is characterized by comprising the following steps:
1) Calibrating the industrial camera by using a Zhang calibration method;
2) Acquiring an ROI (region of interest) of the PIN image by using a deep learning frame;
3) Carrying out image preprocessing on each ROI;
4) Carrying out image analysis and identification on the preprocessed ROI area to obtain the pixel width of the PIN needle at the ROI area;
5) And converting the diameter size of the PIN needle at the ROI according to the pixel width of the PIN needle.
2. The sub-pixel size detection method for the PIN needle image of the ceramic antenna as claimed in claim 1, wherein: the PIN has a plurality of different stepped size shapes.
3. The sub-pixel size detection method for the ceramic antenna PIN image as claimed in claim 1, wherein: the industrial camera is calibrated by using a Zhang calibration method, which comprises the step of calibrating a distortion matrix, pixel equivalent and an internal and external parameter matrix of the industrial camera by using a checkerboard calibration board.
4. The sub-pixel size detection method for the PIN needle image of the ceramic antenna as claimed in claim 3, wherein: the calibration of the distortion matrix refers to the distortion correction of a calibration image to obtain a calibration image and a distortion matrix;
the calibration of the pixel equivalent comprises the following steps: and searching and storing the angular points of the checkerboard in the calibration graph after the deformity correction, respectively obtaining the pixel width between two adjacent angular points in each line of angular points, and then obtaining pixels contained in unit length by taking the average value as pixel equivalent pixelPerMetric.
5. The sub-pixel size detection method for the ceramic antenna PIN image as claimed in claim 1, wherein: the step 2) specifically comprises the following steps: the method comprises the steps of collecting a data set in advance for a PIN sample, pre-training a YooloX-s network by using a labelimg tool labeling data set, then inputting a PIN image collected by a current camera, and obtaining coordinate information of an ROI (region of interest) by using different measurement regions of the PIN output by a trained YooloX-s network model as the ROI.
6. The sub-pixel size detection method for the PIN needle image of the ceramic antenna as claimed in claim 1, wherein: the step 4) is specifically as follows:
4.1 Carrying out sub-pixel edge detection based on curve fitting and improved Sobel operator to obtain a sub-pixel edge image;
4.2 Utilizing a probability Hough method to fit in the sub-pixel edge image to obtain parallel edge lines of the PIN needle, and calculating the space pixel width of the parallel edge lines as the pixel width of the PIN needle.
7. The sub-pixel size detection method for the PIN needle image of the ceramic antenna as claimed in claim 6, wherein: the 4.1) is specifically as follows:
4.1.1 Processing the image after the Gaussian filter smoothing processing by using an improved Sobel operator to obtain a pixel-level edge:
firstly, acquiring the gradient amplitude and direction of an ROI (region of interest) by using an improved 8-direction Sobel operator set by the following formula, and establishing a gradient map of the ROI;
wherein S is 1 ～S 8 Respectively are directional templates with gradient directions of 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees and 315 degrees;
then, carrying out thinning edge and strong and weak edge connection on the gradient map of the ROI by using non-maximum suppression and double thresholds in a Canny method to obtain a pixel-level edge;
4.1.2 ) extracting sub-pixel edges of the image by processing the pixel-level edges by using a curve fitting method:
traversing edge points on each pixel-level edge, selecting 30 edge points on the gradient map along the gradient direction of the current edge point on the pixel-level edge, and performing Gaussian curve fitting sub-pixel processing by adopting a Gaussian function of the following formula to obtain a Gaussian curve:
wherein i =1,2, 3.., 20, x i Is the abscissa position, z, of the ith edge point i Is the gradient magnitude of the ith edge point, z max 、x max And L is the gradient amplitude peak value, the peak value abscissa position and half width information respectively;
then, the position with the maximum gradient in the Gaussian curve is found as the edge point on the edge of the sub-pixel, namely, the gradient amplitude peak value z in the gradient direction is solved max And peak abscissa position x max And then according to the peak value z of the gradient amplitude max And peak abscissa position x max Solving the vertical coordinate position y of the peak value by combining the gradient direction max Thereby determining that the current edge point corresponds to the sub-pixelEdge point coordinates (x) on the edge max ,y max )。
4.1.3 Go through each edge point on the traversal pixel level edge to obtain the edge points of all the edge points corresponding to the sub-pixel edge, and finally form the sub-pixel edge.
8. The method for detecting the sub-pixel size of the PIN image of the ceramic antenna as claimed in claim 6, wherein: the step 4.2) is specifically as follows: inputting the edge of the sub-pixel into a probability Hough method for processing and extracting to obtain each straight-line segment, wherein each straight-line segment returns coordinates (x 1, y 1), (x 2, y 2) of two end points; calculating the slope of all the extracted straight line segments, taking two straight line segments with the same or similar slopes as parallel edge lines on two sides of the PIN needle, and calculating the pixel interval between the two straight line segments as the pixel width of the PIN needle.
9. The sub-pixel size detection method for the PIN needle image of the ceramic antenna as claimed in claim 1, wherein: specifically, the step 5) is to convert the pixel width of the PIN needle according to the pixel equivalent pixelpermetric to obtain the actual diameter size of the PIN needle at the ROI area.
10. A sub-pixel size detection system for implementing the method of any one of claims 1-9, comprising:
the image acquisition and display module is used for acquiring an image of the tested PIN, preprocessing the image and displaying a measurement result;
the ROI extraction module is used for acquiring an ROI area of the PIN needle and eliminating image redundant information by the deep learning framework;
the camera calibration module is used for calibrating an internal and external parameter matrix, a distortion matrix and pixel equivalent of the industrial camera by a Zhang calibration method;
the measuring module is used for sub-pixel edge detection, fitting parallel edge lines of the part in a sub-pixel edge image by utilizing a probability Hough method, and calculating the distance between the parallel edge lines;
and the parameter monitoring and storing module is used for tracking the measuring progress in real time, displaying key parameters and locally storing the parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211398987.8A CN115641326A (en) | 2022-11-09 | 2022-11-09 | Sub-pixel size detection method and system for ceramic antenna PIN needle image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211398987.8A CN115641326A (en) | 2022-11-09 | 2022-11-09 | Sub-pixel size detection method and system for ceramic antenna PIN needle image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115641326A true CN115641326A (en) | 2023-01-24 |
Family
ID=84949745
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211398987.8A Pending CN115641326A (en) | 2022-11-09 | 2022-11-09 | Sub-pixel size detection method and system for ceramic antenna PIN needle image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115641326A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117670916A (en) * | 2024-01-31 | 2024-03-08 | 南京华视智能科技股份有限公司 | Deep learning-based edge detection method for coating machine |
-
2022
- 2022-11-09 CN CN202211398987.8A patent/CN115641326A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117670916A (en) * | 2024-01-31 | 2024-03-08 | 南京华视智能科技股份有限公司 | Deep learning-based edge detection method for coating machine |
| CN117670916B (en) * | 2024-01-31 | 2024-04-12 | 南京华视智能科技股份有限公司 | Coating edge detection method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11551341B2 (en) | Method and device for automatically drawing structural cracks and precisely measuring widths thereof | |
| CN112651968B (en) | Wood board deformation and pit detection method based on depth information | |
| JP2017053819A (en) | Crack detection method and detection program of concrete | |
| CN103149087B (en) | Follow-up window and digital image-based non-contact real-time strain measurement method | |
| CN112132886A (en) | Circular hole center rapid positioning and roundness detection method for aviation part | |
| CN112991287B (en) | Automatic indentation measurement method based on full convolution neural network | |
| CN104700395A (en) | Method and system for detecting appearance crack of structure | |
| CN109409290A (en) | A kind of thermometer calibrating reading automatic recognition system and method | |
| CN112489042A (en) | Metal product printing defect and surface damage detection method based on super-resolution reconstruction | |
| CN115641326A (en) | Sub-pixel size detection method and system for ceramic antenna PIN needle image | |
| CN106529548A (en) | Sub-pixel level multi-scale Harris corner detection algorithm | |
| CN117173125A (en) | Panoramic view-based defect point location display method, device and storage medium | |
| CN111582270A (en) | Identification tracking method based on high-precision bridge region visual target feature points | |
| CN105354855A (en) | High-rise structure appearance detection device and method | |
| CN113705564B (en) | Pointer type instrument identification reading method | |
| CN118015004A (en) | Laser cutting scanning system and method | |
| CN113947583A (en) | Weld joint nondestructive testing method based on deep learning two-dimensional time sequence image | |
| CN111476792B (en) | Extraction method of strip steel image contour | |
| CN115689929A (en) | Aperture measurement method and device based on distortion compensation and adaptive mean value blurring | |
| CN115063681A (en) | Highway pavement disease detection method | |
| CN118537339B (en) | Method and system for evaluating surface quality of part based on sander | |
| CN110349174B (en) | Sliding rail multi-parameter measurement method and measurement device | |
| CN113592877B (en) | Method and device for identifying red line exceeding of pumped storage power station | |
| Yanmin et al. | A method of vehicle dashboard image location based on machine vision | |
| CN116580159B (en) | Three-dimensional modeling system based on static scanning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |