CN112581474B - Industrial component visual edge detection method based on sinusoidal scanning - Google Patents
Industrial component visual edge detection method based on sinusoidal scanning Download PDFInfo
- Publication number
- CN112581474B CN112581474B CN202110196444.7A CN202110196444A CN112581474B CN 112581474 B CN112581474 B CN 112581474B CN 202110196444 A CN202110196444 A CN 202110196444A CN 112581474 B CN112581474 B CN 112581474B
- Authority
- CN
- China
- Prior art keywords
- edge
- sine curve
- point
- image
- value
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a sinusoidal scanning-based industrial component visual edge detection method, which comprises the following steps of collecting a gray image: collecting a gray level image by using an industrial camera; and (3) performing Kirsch edge detection: performing gradient detection on the image, calculating gradients in 8 directions by using a Kirsch edge detection operator, and selecting a maximum value as a result; threshold segmentation: after obtaining the gradient value of the image, carrying out threshold segmentation on the gradient image, and selecting a specific threshold to separate two side regions of the edge; sinusoidal scanning: sending out a sine curve from a set starting point, and searching for an edge along the sine curve; screening edge points to obtain an edge point structure: and taking a point on the sine curve as a circle center, and determining the edge as the edge after the number of edge pixels exceeds a certain threshold value in a circular range. Under the condition of surface light, the adverse factors of complex textures caused by illumination conditions, workpiece materials and workpiece height change structures are overcome, and high-precision dimension measurement of multiple items of complex workpieces is completed.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a sinusoidal scanning-based industrial component visual edge detection method.
Background
Most of the existing edge detection technologies acquire the outline of a workpiece based on ground light, so that the workpiece and the ground light have obvious brightness difference in a picture acquired by a camera, and the outline of the workpiece is acquired through threshold segmentation and edge extraction (such as canny edge extraction). In this case, the edge can be determined as long as the value of the pixel point is changed from the surrounding points, but there are many limitations, and most obviously, the edge of the workpiece surface cannot be detected. If surface defect inspection and dimensional measurement are required simultaneously to save a lot of inspection time, edge extraction of surface light is an important issue. Under the condition of surface light, the edge detection is greatly influenced by complex textures caused by illumination conditions, workpiece materials and workpiece height change structures.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method for detecting the visual edge of the industrial part based on the sinusoidal scanning overcomes the adverse factors of complex textures caused by illumination conditions, workpiece materials and workpiece height change structures under the condition of surface light, and completes high-precision size measurement of multiple items of complex workpieces.
The technical scheme adopted by the invention for solving the technical problems is as follows: a sinusoidal scanning-based industrial component visual edge detection method comprises the following specific steps:
the method comprises the following steps of firstly, collecting a gray level image: collecting a gray level image by using an industrial camera;
the second step, Kirsch edge detection: performing gradient detection on the image, calculating gradients in 8 directions by using a Kirsch edge detection operator, and selecting a maximum value as a result, thereby obtaining a gradient value of the image;
the third step, threshold segmentation: after obtaining the gradient value of the image, carrying out threshold segmentation on the gradient image, and selecting a specific threshold to separate two side regions of the edge;
fourth step, sinusoidal scanning: sending out a sine curve from a set starting point, and searching for an edge along the sine curve;
fifthly, screening the edge points to obtain an edge point structure: and taking a point on the sine curve as a circle center, and determining the edge as the edge after the number of edge pixels exceeds a certain threshold value in a circular range.
Further specifically, in the above technical solution, in the fourth step, the specific step of finding the edge is as follows:
step 1, setting sine curve scanning parameters;
step 2, selecting a sine curve starting point;
and 3, emitting a sine curve along a set direction.
Further specifically, in the above technical solution, in the fifth step, the specific steps of screening the edge points are as follows:
step 1, taking a point on a sine curve as a circle center;
step 2, setting a radius according to an edge effect;
and 3, judging that the white pixels in the circular range are larger than the threshold value as edges.
More specifically, in the above technical solution, for the image of the surrounding or semi-surrounding structure, an edge detection method of sinusoids with different parameters is used.
Further specifically, in the above-described technical solution, for an image having only one side, a method of emitting a sinusoidal curve in one direction from a set position is used.
Further specifically, in the above technical solution, when the edge to be detected is located in the middle of the image, the periphery of the image is surrounded by other parts, and the surface part to be detected is not located at the edge of the workpiece, a method of searching a start point of a sine curve in advance and emitting the sine curve from the start point in one direction is used.
The invention has the beneficial effects that: the method for detecting the visual edge of the industrial part based on the sinusoidal scanning searches for the initial point in a sinusoidal searching and screening mode, can detect the surface optical edge of a workpiece with complex textures, and has high edge accuracy and strong anti-interference capability.
Drawings
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is an artwork captured by an industrial camera;
FIG. 2 is a schematic diagram of the Kirsch edge detection operator;
FIG. 2.1 is an enlarged view of a portion of FIG. 1 at A;
FIG. 2.2 is a schematic view of the separation of the annular inner region and the annular region;
FIG. 2.3 is a schematic diagram of the edge detection results;
FIG. 3.1 is an enlarged view of a portion of FIG. 1 at B;
FIG. 3.2 is a schematic view of the transition zone from the edge to the black area;
FIG. 3.3 is a schematic representation after use of the screening area;
FIG. 4.1 is a schematic illustration of sinusoidal edge detection;
FIG. 4.2 is a schematic diagram of single-edge detection;
FIG. 4.3 is a flow chart of single-sided sinusoidal edge detection;
FIG. 5 is an overall flow diagram of the present invention;
FIG. 5.1 is a schematic view of the portion of the surface to be inspected not being at the edge of the workpiece;
FIG. 5.2 is a schematic diagram of a pixel variation curve;
FIG. 5.3 is a flow chart of probing a starting point;
FIG. 5.4 is a graph of the results of the starting point test;
FIG. 5.4.1 is an enlarged fragmentary view at H of FIG. 5.4;
FIG. 5.5 is a graph of edge detection results from detected starting points;
FIG. 5.5.1 is an enlarged view of a portion of FIG. 5.5 at I;
FIG. 5.6 is a flow chart of obtaining edge points;
FIG. 5.7 is a flow chart of the screening points;
FIG. 6 is an edge detection flow diagram;
FIG. 7 is a comparison graph of screening radius selections;
FIG. 8.1 is a graph of real image edge effects;
FIG. 8.2 is a graph of the screening effect of the outer edge detection;
FIG. 8.3 is a graph of the inner edge detection screening effect;
FIG. 9.1 is a schematic image disturbance diagram;
figure 9.2 is a schematic diagram of the screening regional filter interference.
Detailed Description
In order to make the technical problems, technical solutions and advantageous effects solved by the present invention more clearly apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 5, a sinusoidal scanning-based industrial component visual edge detection method specifically includes the following steps:
the method comprises the following steps of firstly, collecting a gray level image: and collecting a gray level image by using an industrial camera.
Referring to fig. 1, the picture taken from the industrial camera is an overall view of the entire workpiece object.
Referring to fig. 2.1 and 3.1, it is necessary to first cut out the part to be measured, where fig. 2.1 is taken at a of fig. 1, and fig. 3.1 is taken at B of fig. 1.
The second step, Kirsch edge detection: gradient detection is performed on the image using a Kirsch edge detection operator (see fig. 2), which has the advantage of calculating gradients in 8 directions and selecting the maximum value as a result, thereby obtaining the gradient value of the image.
The Kirsch edge detection can carry out omnidirectional detection on complex textures, enables the surface textures of the workpiece to be in a high gradient value, and provides a good data base for the subsequent threshold segmentation.
Other edge detection operators, such as: sobel edge detection operator and canny edge detection operator. The Sobel edge detection operator cannot enable the texture of the surface of the object to be in a high gradient value, so that the real edge is distinguished; due to the complexity of texture, rather than a simple binarized contour image, the canny edge detection operator cannot accurately extract edges.
The third step, threshold segmentation: after obtaining the gradient value of the image, carrying out threshold segmentation on the gradient image, and selecting a specific threshold to separate two side regions of the edge.
Due to the consistency of the texture of the workpiece, namely, the similar gradient strength exists at the part to be measured and the strength difference exists between the two sides of the edge to be extracted, a certain threshold value can be used for separating the two side areas of the edge.
Referring to fig. 2.1, the gradient value of the texture light and shade staggered reaction of the circular ring is larger, and the gradient value of the texture in the circular ring is smaller.
As shown in fig. 2.2, the annular region can be separated from the annular region by selecting a specific threshold.
Fourth step, sinusoidal scanning: sending out a sine curve from the set starting point, and searching for an edge along the sine curve. In the fourth step, see fig. 5.6, the specific steps for finding edges are as follows: step 1, setting sine curve scanning parameters; step 2, selecting a sine curve starting point; and 3, emitting a sine curve along a set direction.
Fifthly, screening the edge points to obtain an edge point structure: and taking a point on the sine curve as a circle center, and determining the edge as the edge after the number of edge pixels exceeds a certain threshold value in a circular range. In the fifth step, see fig. 5.7, the specific steps of screening the edge points are as follows: step 1, taking a point on a sine curve as a circle center; step 2, setting a radius according to an edge effect; and 3, judging that the white pixels in the circular range are larger than the threshold value as edges.
For different images, the adopted sine curve edge extraction method is also different:
in the first case, for images of surrounding or semi-surrounding structures, an edge detection method of sinusoids of different parameters is used (see fig. 4.1).
Referring to fig. 4.1 and fig. 6, the specific steps of obtaining the edge points are as follows:
step 1, setting parameters of a sine curve:a andwherein A represents the amplitude of the vibration,representing the frequency. The selection of a requires that the upper and lower portions intersect the edge portion,the frequency is set according to the number of points needed. The value range of x represents the range in which edge detection is required.
Step 2, selecting a starting point: a point is chosen as the start of the sinusoid and a point on the medial axis of the image can be chosen, the horizontal straight line in the figure being denoted J.
And 4, calculating the proportion of white pixels in the screening area to obtain an edge point result: and when the white pixel proportion of the screening area is greater than the threshold value, acquiring the circle center of the circular screening area meeting the conditions on the sine curve as the result of edge extraction.
If a single sinusoid has fewer points, the starting point may be selected multiple times in step 2 and scanned.
The screening area and the screening conditions are used to remove disturbances inside the circle as in fig. 2.2 and defects in the edge itself. The screening area is a range which is expanded to the periphery by taking a point on the sine curve as a center, and the range can be a square or a circle, and the circle is superior to the square. In particular, when the edges of the arc are curved, the circular area is tangent to the edge with the same effect at different angles, whereas the square shape differs.
The range of the region and the proportion of the white pixels in the region are set to screen the edge points, the proportion is the proportion of the white pixels in fig. 6, and the white points are regions with larger gradient after threshold segmentation and represent the surface of the object. When the white pixel is larger than the set ratio in the area of a certain point of the sine curve, the point is determined as an edge point. At this time, the result of the edge detection can avoid the influence of the small area in the ring and the interference of the unevenness of the ring edge, and the detection result (see fig. 2.3).
The selection of the radius of the circular screening area is mainly determined by the thickness and roughness of the edge and the interference in the middle of the line scan. As shown in fig. 7, in the case of thin and completely smooth edge, a smaller radius needs to be selected, and the screening threshold is set at 50%, i.e. when the sinusoidal line is in contact with the edge, the edge point result is screened out. A larger radius may result in the screening range at the edge failing to meet 50% requirements. However, when the edge is rough or noisy, it is often caused by the fuzzy edge of the real image or the existence of a chamfer (chamfer means to cut the edge of the workpiece into a certain slope), as shown in fig. 8.1. In this case, if we need to find the outer edge, as in fig. 8.2, the threshold of the screening needs to be changed to 30%; if an inner edge is to be found, as in fig. 8.3, the threshold for the screening needs to be changed to 70%. Finally, if there is some interference on the sinusoidal path, as in the case of fig. 9.1, white circles the interference part, and with reasonable threshold and radius adjustment, the screening area can filter these interference areas, as in fig. 9.2.
In addition, there is typically a transition due to light and workpiece edges (see fig. 3.1). The transition zone from the edge to the black area (see area C in fig. 3.2) is jagged, but after the use of the screening area it can be seen from the results in fig. 3.3 that the choice of edge points does not fluctuate too much, selecting the correct edge substantially, without being affected by jagging.
Therefore, edge detection under various different conditions can be met by reasonably setting the radius and the screening proportion of the screening area.
In the second case, for an image with only one edge, a method of emitting a sinusoid in one direction from a set position is used (see fig. 4.2 and 4.3), thereby obtaining a set of edge points on a straight line.
The specific steps for obtaining the edge points are as follows:
step 1, setting parameters of a sine curve:the amount of the component (A) in (1),andwherein A represents the amplitude of the vibration,the frequency is represented by a frequency-dependent variable,indicating the phase. The selection of a requires that the upper and lower portions intersect the edge portion,the frequency is set according to the number of points needed. The value range of x represents the range in which edge detection is required.
Step 2, selecting a starting point and a direction: a starting straight line is chosen outwardly of the edge, as indicated by the horizontal line M in fig. 4.2. The curve direction is chosen from left to right.
And 4, calculating the proportion of white pixels in the screening area to obtain an edge point result: the screening area is represented by O, the circle center moves along the sine curve, the white pixel proportion of the screening area is calculated, and the circle center point meeting the screening condition is marked as the edge point on the sine curve.
Finally, after the sinusoidal scanning is completed at all the positions in the range, the edge point set on the corresponding range can be obtained. The basic steps are similar to the previous case, but are replaced by unilateral point taking.
In the third case, in some complex cases, the edge portion to be detected is not at the edge of the workpiece, there is no possibility of starting scanning from the edge of the picture (see fig. 5.1, 5.4 and 5.5), in the case where the edge to be detected is in the middle of the image, there is surrounding of other parts around, and the surface portion to be detected is not at the edge of the workpiece, using a method of searching in advance for the start point of the sinusoid, and emitting the sinusoid from the start point in one direction. If necessary, the edges in the region D in figure 5.1.
The specific steps for obtaining the edge points are as follows, and the flow is shown in fig. 5.3:
step 1, setting parameters of a sine curve:the amount of the component (A) in (1),andwherein A represents the amplitude of the vibration,the frequency is represented by a frequency-dependent variable,indicating the phase. The selection of a requires that the upper and lower portions intersect the edge portion,the frequency is set according to the number of points needed. The value range of x represents the range in need of edge detection, here, the region D in fig. 5.1;
step 2, selecting a starting point: the method for searching the initial point of the sine curve comprises the following steps: first, select the edge interval, such as the range of the length of D region in fig. 5.1, from top to bottom with 2 × 2 region mean values, step size is one pixel, and take the pixel values of the range of the edge portion to be detected in fig. 5.1 by column (take 2 × 2 pixel mean values in the direction perpendicular to the edge). The 2 x 2 area mean is used for smoothing pixels, reducing the influence of individual abrupt pixels in the texture, and finally obtaining the change situation of each column of pixel values (see fig. 5.2, the ordinate represents the pixel value, and the abscissa represents the coordinate from top to bottom). As can be seen from fig. 5.2, the fluctuation of the pixel value has a very significant variation over the edge that needs to be detected. In this case, the position of the black portion is acquired using the size of the pixel value as a starting point of the edge extraction. However, in some cases, white anomalies with high brightness may appear in the black parts (see area F in fig. 5.1), and in order to avoid such occurrences, a method of averaging multiple points around is adopted, such as threshold screening after 5 points are averaged. The effect of the 5 points is to remove the effect of individual outliers within the local area (2 x 2 region). Since the starting point must be in the black background part, the method of accurately locating the starting point is to find the midpoint of the transitions of the pixel values from high to low and from low to high as the starting point, specifically, the middle of the two transitions of the pixel values from high to low (white to black) and from low to high (black to white), i.e., at E of fig. 5.2, i.e., at the black gap of the H region in fig. 5.4 (see fig. 5.4.1), the starting point is drawn by the white point.
Step 4, obtaining an edge point result: as in the second case, the screening area is a white circle, and the center of the circle moves along the sine curve until the center point satisfying the screening condition is marked as the edge point on the sine curve, and the result is shown as the white edge point in the I area in fig. 5.5 (see fig. 5.5.1).
Under the condition that the edge quality is normal, the edge points can filter fluctuation on most textures, the overall fluctuation is within 3 pixels, and the multipoint fitting is closer to the true value. In addition, the method avoids the use of morphological operations such as swelling corrosion to avoid loss of accuracy.
It should be noted that: the start and end points of the sinusoid and the change of the direction of motion (e.g. from the top to the bottom of the sinusoid), the shape of the screening area (e.g. the screening area becomes a semicircle) and the screening conditions (e.g. the screening conditions become the number of white dots on the edge of the circle instead of the dots within the range) all count the same approach.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention are equivalent to or changed within the technical scope of the present invention.
Claims (3)
1. A sinusoidal scanning-based industrial component visual edge detection method is characterized by comprising the following specific steps:
the method comprises the following steps of firstly, collecting a gray level image: collecting a gray level image by using an industrial camera;
the second step, Kirsch edge detection: performing gradient detection on the image, calculating gradients in 8 directions by using a Kirsch edge detection operator, and selecting a maximum value as a result, thereby obtaining a gradient value of the image;
the third step, threshold segmentation: after obtaining the gradient value of the image, carrying out threshold segmentation on the gradient image, and selecting a specific threshold to separate two side regions of the edge;
fourth step, sinusoidal scanning: sending out a sine curve from a set starting point, and searching for an edge along the sine curve;
fifthly, screening the edge points to obtain an edge point structure: taking a point on the sine curve as a circle center, and determining the number of edge pixels as an edge after the number of the edge pixels exceeds a certain threshold value in a circular range;
under the condition that the edge to be detected is positioned in the middle of the image, the surrounding is surrounded by other parts, the surface part to be detected is not positioned at the edge of the workpiece, a method for searching a sine curve starting point in advance and emitting the sine curve from the starting point along one direction is used, and the specific steps for acquiring the edge point at the moment are as follows:
step 1, setting parameters of a sine curve:a in (1),Andwherein A represents the amplitude of the vibration,the frequency is represented by a frequency-dependent variable,indicating the phase, a is chosen such that the upper and lower portions intersect the edge portion,setting frequency according to the number of the required points, wherein the value range of x represents the range of the required edge detection;
step 2, selecting a starting point: the method for searching the initial point of the sine curve comprises the following steps: firstly, selecting an edge interval, using 2 x 2 area mean values from top to bottom, taking the step length as a pixel, and taking out the pixel value of the range of the edge part to be detected according to columns; using the 2 x 2 region average value to smooth the pixels, reducing the influence of individual mutation pixels in the texture, and finally obtaining the change condition of each row of pixel values; the fluctuation of the pixel value has a very obvious change on the edge to be detected, in this case, the position of the black part is obtained by using the size of the pixel value as the starting point of the edge extraction; in some cases, white abnormality with high brightness occurs in a black part, and a method of averaging multiple points around is adopted;
step 3, sine curve edge extraction: scanning from the starting point to the left and searching for the edge of the area;
step 4, obtaining an edge point result: the screening area is a white circle, the center of the circle moves along the sine curve until the center point meeting the screening condition is marked as the edge point on the sine curve.
2. The method for detecting the visual edge of the industrial part based on the sinusoidal scanning as claimed in claim 1, wherein the method comprises the following steps: in the fourth step, the specific steps of finding the edge are as follows:
step 1, setting sine curve scanning parameters;
step 2, selecting a sine curve starting point;
and 3, emitting a sine curve along a set direction.
3. The method for detecting the visual edge of the industrial part based on the sinusoidal scanning as claimed in claim 1, wherein the method comprises the following steps: in the fifth step, the specific steps of screening the edge points are as follows:
step 1, taking a point on a sine curve as a circle center;
step 2, setting a radius according to an edge effect;
and 3, judging that the white pixels in the circular range are larger than the threshold value as edges.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110196444.7A CN112581474B (en) | 2021-02-22 | 2021-02-22 | Industrial component visual edge detection method based on sinusoidal scanning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110196444.7A CN112581474B (en) | 2021-02-22 | 2021-02-22 | Industrial component visual edge detection method based on sinusoidal scanning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112581474A CN112581474A (en) | 2021-03-30 |
CN112581474B true CN112581474B (en) | 2021-05-18 |
Family
ID=75113940
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110196444.7A Active CN112581474B (en) | 2021-02-22 | 2021-02-22 | Industrial component visual edge detection method based on sinusoidal scanning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112581474B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115082462B (en) * | 2022-08-22 | 2022-11-08 | 山东海鑫达石油机械有限公司 | Method and system for detecting appearance quality of fluid conveying pipe |
CN115760884B (en) * | 2023-01-06 | 2023-04-14 | 山东恩信特种车辆制造有限公司 | Semitrailer surface welding slag optimization segmentation method based on image processing |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104200196A (en) * | 2014-08-12 | 2014-12-10 | 侯志勇 | Method for automatically identifying position of guide pin in X-ray perspective image |
CN110412257A (en) * | 2019-07-22 | 2019-11-05 | 深圳市预防宝科技有限公司 | A kind of combination is manually demarcated and the indicator paper block localization method of astral ray algorithm |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7953288B2 (en) * | 2008-03-18 | 2011-05-31 | Xerox Corporation | Edge transition detection in a high-addressable error diffusion environment |
CN108022233A (en) * | 2016-10-28 | 2018-05-11 | 沈阳高精数控智能技术股份有限公司 | A kind of edge of work extracting method based on modified Canny operators |
CN107784300A (en) * | 2017-11-30 | 2018-03-09 | 西安科锐盛创新科技有限公司 | Anti- eye closing photographic method and its system |
-
2021
- 2021-02-22 CN CN202110196444.7A patent/CN112581474B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104200196A (en) * | 2014-08-12 | 2014-12-10 | 侯志勇 | Method for automatically identifying position of guide pin in X-ray perspective image |
CN110412257A (en) * | 2019-07-22 | 2019-11-05 | 深圳市预防宝科技有限公司 | A kind of combination is manually demarcated and the indicator paper block localization method of astral ray algorithm |
Also Published As
Publication number | Publication date |
---|---|
CN112581474A (en) | 2021-03-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112581474B (en) | Industrial component visual edge detection method based on sinusoidal scanning | |
CN105334219B (en) | A kind of bottle mouth defect detection method of residual analysis dynamic threshold segmentation | |
CN114140679B (en) | Defect fusion method, device, recognition system and storage medium | |
US6876445B2 (en) | Method for analyzing defect data and inspection apparatus and review system | |
CN110858395B (en) | Method for detecting defects of dirty filaments of coiled filaments | |
TWI532116B (en) | Method and system for wafer registration | |
CN102779279A (en) | Image identification method of stirring friction welding arc stripe interval | |
CN111353993A (en) | Thread angle measuring method based on machine vision | |
CN113340909B (en) | Glue line defect detection method based on machine vision | |
CN114119534A (en) | Three-dimensional point cloud-based weld defect detection method | |
CN114677376B (en) | Method and system for testing defects in production of woven belts based on illumination pretreatment | |
KR19980032065A (en) | Image processing method and apparatus | |
CN110717909A (en) | Metal surface scratch detection method and device | |
CN115620061A (en) | Hardware part defect detection method and system based on image recognition technology | |
CN109585325B (en) | Method for detecting crystal edge washing boundary | |
CN109544513A (en) | A kind of steel pipe end surface defect extraction knowledge method for distinguishing | |
CN115330770B (en) | Cloth area type defect identification method | |
JP2011008482A (en) | Defect detection method, defect detection device and defect detection program | |
CN110857921A (en) | Method for detecting broken filament defects of packaged filaments | |
JP3456096B2 (en) | Shape inspection method | |
CN114937003A (en) | Multi-type defect detection system and method for glass panel | |
JP6114559B2 (en) | Automatic unevenness detector for flat panel display | |
CN113222880A (en) | Analytical detection method | |
JP5033076B2 (en) | Appearance inspection method and apparatus | |
JP2710685B2 (en) | Defect detection method by visual inspection |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |