CN115100116A - Plate defect detection method based on three-dimensional point cloud - Google Patents
Plate defect detection method based on three-dimensional point cloud Download PDFInfo
- Publication number
- CN115100116A CN115100116A CN202210588295.3A CN202210588295A CN115100116A CN 115100116 A CN115100116 A CN 115100116A CN 202210588295 A CN202210588295 A CN 202210588295A CN 115100116 A CN115100116 A CN 115100116A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- dimensional point
- cloud data
- plate
- points
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 56
- 230000007547 defect Effects 0.000 title claims abstract description 39
- 238000001914 filtration Methods 0.000 claims abstract description 30
- 238000000034 method Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 45
- 239000013598 vector Substances 0.000 claims description 19
- 238000006243 chemical reaction Methods 0.000 claims description 15
- 230000009466 transformation Effects 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 3
- 238000003062 neural network model Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
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
-
- G06T5/70—
-
- 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
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- 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
Abstract
The invention discloses a three-dimensional point cloud-based plate defect detection method, which relates to the field of industrial defect detection, and is characterized in that the method comprises the steps of filtering three-dimensional point cloud data of a plate acquired in a production field, converting coordinates, fitting a plane and the like, measuring related parameters, comparing the measured related parameters with a set standard value and a set threshold value, and judging whether defects exist in the aspects of size and flatness.
Description
Technical Field
The invention relates to the field of industrial defect detection, in particular to a plate defect detection method based on three-dimensional point cloud.
Background
Along with the development of 3D machine vision technology, product defect intelligent detection technique based on 3D machine vision will be future development trend, and panel is the regular stamping workpiece of common shape, quality problems such as deformation, rake angle often can appear, and in actual production process, to size and roughness aspect have millimeter level precision requirement usually, only rely on the unable precision requirement that satisfies of the traditional detection method of workman's visual inspection.
CN 114187267a proposes a method for detecting defects of a stamping part based on machine vision, which utilizes LOG operators of different scales to perform edge detection on an image of the stamping part to obtain edge images of multiple stamping parts of different scales, and matches actual coordinate positions of edge pixel points of the stamping part with standard positions of standard edge pixel points in a template image to detect defects of the stamping part. The workpiece surface defect detection method based on deep learning proposed by CN 111415329a acquires workpiece images under different backgrounds and illumination conditions, constructs a neural network model after preprocessing to obtain feature maps of different layers, and outputs position information and categories of workpiece surface defects after feature pyramid fusion prediction.
However, the method of edge detection needs template matching with a standard image, and the detection result depends on the image quality of the template to a certain extent and has no independent detection property; the detection method based on deep learning needs to collect a workpiece image data set in advance, can be used for detecting an actual product after being trained by using a neural network model, and has larger error in detection accuracy when data of related defect images are less.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a plate defect detection method based on three-dimensional point cloud, which comprises the steps of carrying out operations such as filtering, coordinate conversion, plane fitting and the like on plate three-dimensional point cloud data acquired in a production field, measuring related parameters, comparing the measured related parameters with a set standard value and a set threshold value, and judging whether defects exist in the aspects of size and flatness or not.
The technical scheme adopted by the invention is as follows:
a plate defect detection method based on three-dimensional point cloud comprises the following steps:
s1: setting a selection area, and performing filtering operation on the three-dimensional point cloud data of the plate by adopting a filtering algorithm to remove data outside the selection area and noise of the data in the selection area;
s2: utilizing the filtered three-dimensional point cloud data, segmenting the three-dimensional point cloud data into a plurality of parts according to semantics, and screening out the plate parts according to semantic features to obtain the three-dimensional point cloud data of the plate to be measured;
s3: converting the three-dimensional point cloud data of the plate to be detected into a three-dimensional point cloud matrix, calculating a characteristic value and a characteristic vector of the three-dimensional point cloud matrix, creating a transformation matrix according to the characteristic vector, and performing coordinate conversion on the three-dimensional point cloud data of the plate;
s4: detecting size defects; calculating the length, width and height of the plate according to set precision by using a size detection algorithm, comparing the calculated length, width and height with set standard values of the actual length, width and height of the plate, and judging that size defects exist if the set standard values exceed threshold requirements;
s5: extracting a tilt angle detection area; cutting the three-dimensional point cloud data of the plate by using a cutting algorithm to obtain the three-dimensional point cloud data of a rake angle detection area;
s6: fitting a plane model equation of a rake angle detection area; fitting a plane model where three-dimensional point cloud data of the warped foot detection area are located by using a plane fitting algorithm, and solving a plane model equation;
s7: detecting a warped foot defect; and calculating the average distance from each point in the tilt detection area to the plane model equation in the S6, comparing the obtained average distance value with a set standard value of the plate warping degree, and judging that the tilt defect exists if the average distance value exceeds the threshold value requirement.
The filtering algorithm is used for filtering the original three-dimensional point cloud data by adopting conditional filtering and radius filtering in sequence, and comprises the following steps of:
s1.1, performing conditional filtering, namely setting range conditions for X, Y, Z coordinates of the three-dimensional point cloud data, and filtering according to the range conditions to remove useless points outside the set range;
s1.2, radius filtering, namely, randomly selecting one point in three-dimensional point cloud data after condition filtering, selecting a sphere range taking the point as a sphere center and a set value as a radius, counting the number of points falling in the sphere range by calculating the distance from points around the sphere center to the sphere center, and keeping the selected point when the number is larger than or equal to a set threshold value; and when the number is smaller than the set threshold value, removing the selected points, traversing all the points in the three-dimensional point cloud data, and leaving the points to be the three-dimensional point cloud data after the noise is removed.
The S3 specifically includes the following steps:
s3.1: converting the three-dimensional point cloud data into an N-row 3-column matrix M according to X, Y, Z coordinates of each point, wherein the coordinates of each point correspond to three values in one row of the matrix M; average of all elements in M
With each value M in the matrix M ij Subtract mean value
Obtaining a zero-averaged matrix M', as shown in formula (1) and formula (2):
s3.2: calculating a covariance matrix C as shown in equation (3):
wherein (M') T A rank-conversion matrix representing matrix M';
s3.3: establishing an equation set shown in the formula (4), and calculating an eigenvalue lambda of the covariance matrix C and a corresponding eigenvector v;
wherein λ 1 ,λ 2 ,λ 3 Is the first eigenvalue, the second eigenvalue, the third eigenvalue, v of the matrix C 1 ,v 2 ,v 3 Are each lambda 1 ,λ 2 ,λ 3 The corresponding feature vector;
s3.4: the feature vector v 1 ，v 2 ，v 3 Respectively setting direction vectors of X, Y, Z axes under a coordinate system;
s3.5: and creating a transformation matrix according to the feature vector, wherein the transformation matrix is defined as shown in a formula (5):
wherein, the first and the second end of the pipe are connected with each other,
each row is respectively corresponding to a direction vector on X, Y, Z coordinate axes as a rotation matrix; w is a 1 ,w 2 ,w 3 Respectively represent the offset amount to be generated in the X, Y, Z axis direction after transformation, and the default setting of 0 represents no offset; u represents the scaling factor of the transform, which is set to 1 by default to represent no scaling;
s3.6: and carrying out coordinate conversion on the three-dimensional point cloud data of the plate according to the transformation matrix, wherein the three coordinate axis directions after the conversion are respectively as follows: the X axis is horizontally rightward, the Y axis is vertically upward, and the Z axis is vertically outward.
The size detection algorithm specifically comprises the following steps: traversing three-dimensional coordinates of all points in the three-dimensional point cloud data, calculating a minimum cuboid bounding box which contains all the points and has edges parallel to coordinate axes according to the range of the coordinate values, calculating the edge length of the minimum cuboid bounding box and sequencing to obtain the length, width and height of the three-dimensional point cloud data of the plate.
The cutting algorithm specifically comprises the following steps:
s5.1: root of herbaceous plantsObtaining two-dimensional coordinates P of X axis and Y axis of four angular points of the plate according to the vertex information of the cuboid bounding box 1 (x 1 ,y 1 )，P 2 (x 2 ,y 2 )，P 3 (x 3 ,y 3 )，P 4 (x 4 ,y 4 )；
S5.2: according to the actual frame width L of the plate, solving two-dimensional coordinates P 'of an X axis and a Y axis of four corner points of a cutting area' 1 (x 1 +L,y 1 -L)，P′ 2 (x 2 +L,y 2 +L)，P′ 3 (x 3 -L,y 3 +L)，P′ 4 (x 4 -L,y 4 -L)；
S5.3: from P' 1 、P′ 2 、P′ 3 、P′ 4 And forming a rectangular area by the four points, performing conditional filtering on the three-dimensional point cloud data of the plate in the rectangular area, removing the three-dimensional point cloud data in the rectangular area, and leaving the three-dimensional point cloud data as a warp angle detection area.
The plane fitting algorithm specifically comprises the following steps:
s6.1: randomly selecting coordinates Q of three points in three-dimensional point cloud data of a rake angle detection area 1 (x 1 ,y 1 ,z 1 )，Q 2 (x 2 ,y 2 ,z 2 )，Q 3 (x 3 ,y 3 ,z 3 )；
S6.2: according to Q 1 ,Q 2 ,Q 3 The coordinates of the three points are established into a (6) equation set, the value of the unknown coefficient A, B, C, D is calculated, and the plane model equation is determined to be that Ax + By + Cz + D is 0:
s6.3: let (x) i ,y i ,z i ) Calculating the average distance from all K points in the three-dimensional point cloud data of the tilt angle detection area to the point-plane of the plane model equation obtained in S6.2 for the coordinates of any point in the tilt angle detection area
Wherein the distance d between the point and the surface i And average distance
Comprises the following steps:
s6.4: repeating S6.2-S6.3 until reaching the set iteration number, and selecting the average distance of the point and the surface
And the minimum plane equation model is the plane model equation finally fitted.
Advantageous technical effects
Compared with the prior art, the plate defect detection method based on the three-dimensional point cloud provided by the invention has the advantages of high efficiency, high precision and low labor, and can realize the defect detection in the aspects of size and flatness only according to the three-dimensional point cloud data acquired on site without using a standard data sample as a template and training a preset neural network model.
Drawings
Fig. 1 is a flowchart of a plate defect detection method based on three-dimensional point cloud according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating an embodiment of the present invention providing original three-dimensional point cloud data to be subjected to defect detection;
FIG. 3 is a three-dimensional point cloud data of a plate to be detected according to an embodiment of the present invention;
fig. 4 is three-dimensional point cloud data of a plate material after coordinate conversion according to an embodiment of the present invention.
Detailed Description
An embodiment of the present invention will be described in detail below with reference to the accompanying drawings.
The embodiment provides a plate defect detection method based on three-dimensional point cloud, as shown in fig. 1, comprising the following steps:
in this embodiment, the original three-dimensional point cloud data to be subjected to defect detection is shown in fig. 2;
s1: removing useless points and noise in the three-dimensional point cloud data by adopting a filtering algorithm to obtain filtered three-dimensional point cloud data, and comprising the following steps of:
s1.1, conditional filtering, namely setting range conditions for X, Y, Z coordinates of the three-dimensional point cloud data, filtering according to the range conditions, and removing useless points outside the set range.
S1.2, radius filtering, namely, randomly selecting one point in three-dimensional point cloud data after condition filtering, selecting a sphere range taking the point as a sphere center and a set value as a radius, counting the number of points falling in the sphere range by calculating the distance from points around the sphere center to the sphere center, and keeping the selected point when the number is larger than or equal to a set threshold value; and when the number is smaller than the set threshold value, removing the selected points, traversing all the points in the three-dimensional point cloud data, and leaving the points to be the three-dimensional point cloud data after the noise is removed.
S2: dividing the three-dimensional point cloud data into a plurality of parts according to semantics, screening out a plate part according to semantic information characteristics to obtain the three-dimensional point cloud data of the plate to be measured, wherein the three-dimensional point cloud data is shown in figure 3;
s3: carrying out PCA coordinate conversion on the three-dimensional point cloud data of the plate to be detected, and comprising the following steps:
s3.1: and converting the three-dimensional point cloud data into an N-row 3-column matrix M according to X, Y, Z coordinates of each point, wherein the coordinates of each point correspond to three values in one row of the matrix M. Average of all elements in M
With each value M in the matrix M ij Subtract mean value
Obtaining a matrix M 'after zero equalization'As shown in formulas (1) and (2):
s3.2: calculating a covariance matrix C, as shown in formula (3), where (M') T A rank-conversion matrix representing matrix M';
s3.3: establishing an equation set shown in the formula (4), and calculating an eigenvalue lambda of the covariance matrix C and a corresponding eigenvector v;
find lambda 1 ,λ 2 ,λ 3 Is the first eigenvalue, the second eigenvalue, the third eigenvalue, v of the matrix C 1 ,v 2 ,v 3 Are each lambda 1 ,λ 2 ,λ 3 The corresponding feature vector.
S3.4: the feature vector v 1 ，v 2 ，v 3 Respectively setting direction vectors of X, Y, Z axes under a coordinate system;
s3.5: and creating a transformation matrix according to the feature vector, wherein the transformation matrix is defined as shown in the formula (5):
wherein the content of the first and second substances,
for the rotation matrix, each row corresponds to X, YDirection vector on the Z coordinate axis; w is a 1 ,w 2 ,w 3 Each represents an offset amount to be generated in the X, Y, Z-axis direction after transformation, and setting it to 0 by default represents no offset; u represents the scaling factor of the transform and by default setting it to 1 represents no scaling.
S3.6: and carrying out coordinate conversion on the three-dimensional point cloud data of the plate according to the transformation matrix, wherein the three coordinate axis directions after the conversion are respectively as follows: the X axis is horizontally rightward, the Y axis is vertically upward, the Z axis is vertically outward, and the three-dimensional point cloud data of the plate after coordinate conversion is shown in figure 4.
S4: traversing three-dimensional coordinates of all points in the three-dimensional point cloud data, calculating a minimum cuboid bounding box which comprises all the points and has the edge parallel to a coordinate axis according to the range of the coordinate values, calculating the edge length of the minimum cuboid bounding box and sequencing to obtain the length, width and height of the three-dimensional point cloud data of the plate, comparing the obtained length, width and height with set standard values of the length, width and height of the plate, and judging that a size defect exists if the length, width and height exceed the threshold requirement;
s5: the method for extracting the tilt angle detection area comprises the following steps:
s5.1: obtaining two-dimensional coordinates P of X axis and Y axis of four corner points of the plate according to the vertex information of the cuboid bounding box 1 (x 1 ,y 1 )，P 2 (x 2 ,y 2 )，P 3 (x 3 ,y 3 )，P 4 (x 4 ,y 4 )；
S5.2: according to the actual frame width L of the plate, solving two-dimensional coordinates P 'of an X axis and a Y axis of four corner points of a cutting area' 1 (x 1 +L,y 1 -L)，P′ 2 (x 2 +L,y 2 +L)，P′ 3 (x 3 -L,y 3 +L)，P′ 4 (x 4 -L,y 4 -L)；
S5.3: from P' 1 、P′ 2 、P′ 3 、P′ 4 And forming a rectangular area by the four points, performing conditional filtering on the three-dimensional point cloud data of the plate in the rectangular area, removing the three-dimensional point cloud data in the rectangular area, and leaving the three-dimensional point cloud data as a warp angle detection area.
S6: the method for fitting the plane model of the tilt detection area comprises the following steps:
s6.1: randomly selecting coordinates Q of three points in three-dimensional point cloud data of a rake angle detection area 1 (x 1 ,y 1 ,z 1 )，Q 2 (x 2 ,y 2 ,z 2 )，Q 3 (x 3 ,y 3 ,z 3 )；
S6.2: according to Q 1 ,Q 2 ,Q 3 Establishing a (6) equation set By coordinates of the three points, solving the value of an unknown coefficient A, B, C, D, and determining that a plane model equation is that Ax + By + Cz + D is 0;
s6.3: let (x) i ,y i ,z i ) Calculating the average distance from all K points in the three-dimensional point cloud data of the warp detection area to the point-plane of the planar model equation obtained in S6.2 for the coordinates of any point in the warp detection area
Wherein the distance d between the point and the plane i And average distance
Comprises the following steps:
s6.4: repeating S6.1-S6.3 until reaching the set iteration number, and selecting the average distance of the point and the surface
The least planar equation model, i.e. the plane to which the final fit is madeEquation of surface model, will be at this time
Set as the minimum point-plane average distance d min 。
S7: the average distance d of the minimum point and the minimum surface can be obtained from S6.4 min D is mixing min And comparing the warp degree with a set standard value of the warp degree of the plate, and judging that the warp defect exists if the warp degree exceeds the threshold requirement.
Claims (7)
1. A plate defect detection method based on three-dimensional point cloud is characterized by comprising the following steps: the method comprises the following steps:
s1: setting a selection area, and performing filtering operation on the three-dimensional point cloud data of the plate by adopting a filtering algorithm to remove data outside the selection area and noise of the data in the selection area;
s2: utilizing the filtered three-dimensional point cloud data, segmenting the three-dimensional point cloud data into a plurality of parts according to semantics, and screening out the plate parts according to semantic features to obtain the three-dimensional point cloud data of the plate to be measured;
s3: converting the three-dimensional point cloud data of the plate to be detected into a three-dimensional point cloud matrix, calculating a characteristic value and a characteristic vector of the three-dimensional point cloud matrix, creating a transformation matrix according to the characteristic vector, and performing coordinate conversion on the three-dimensional point cloud data of the plate;
s4: detecting size defects; calculating the length, width and height of the plate according to set precision by using a size detection algorithm, comparing the calculated length, width and height with set standard values of the actual length, width and height of the plate, and judging that size defects exist if the set standard values exceed threshold requirements;
s5: extracting a tilt angle detection area; cutting the three-dimensional point cloud data of the plate by using a cutting algorithm to obtain the three-dimensional point cloud data of a rake angle detection area;
s6: fitting a plane model equation of a rake angle detection area; fitting a plane model where three-dimensional point cloud data of the warped foot detection area are located by using a plane fitting algorithm, and solving a plane model equation;
s7: detecting a warped foot defect; and calculating the average distance from each point in the tilt detection area to the plane model equation in the S6, comparing the obtained average distance value with a set standard value of the plate warping degree, and judging that the tilt defect exists if the average distance value exceeds the threshold value requirement.
2. The plate defect detection method based on the three-dimensional point cloud as claimed in claim 1, wherein:
the filtering algorithm is used for filtering the original three-dimensional point cloud data by adopting conditional filtering and radius filtering in sequence, and comprises the following steps of:
s1.1, performing conditional filtering, namely setting range conditions for X, Y, Z coordinates of the three-dimensional point cloud data, filtering according to the range conditions, and removing useless points outside the set range;
s1.2, radius filtering, namely, randomly selecting one point in three-dimensional point cloud data after condition filtering, selecting a sphere range taking the point as a sphere center and a set value as a radius, counting the number of points falling in the sphere range by calculating the distance from points around the sphere center to the sphere center, and keeping the selected point when the number is larger than or equal to a set threshold value; and when the number is smaller than the set threshold value, removing the selected points, traversing all the points in the three-dimensional point cloud data, and leaving the points to be the three-dimensional point cloud data after the noise is removed.
3. The plate defect detection method based on the three-dimensional point cloud as claimed in claim 1, wherein:
the S3 specifically includes the following steps:
s3.1: converting the three-dimensional point cloud data into an N-row 3-column matrix M according to X, Y, Z coordinates of each point, wherein the coordinates of each point correspond to three values in one row of the matrix M; average of all elements in M
With each value M in the matrix M ij Subtract mean value
Obtaining a zero-averaged matrix M', as shown in formula (1) and formula (2):
s3.2: calculating a covariance matrix C as shown in equation (3):
wherein (M') T A rank-conversion matrix representing matrix M';
s3.3: establishing an equation set shown in the formula (4), and calculating an eigenvalue lambda of the covariance matrix C and a corresponding eigenvector v;
wherein λ 1 ,λ 2 ,λ 3 Is the first eigenvalue, the second eigenvalue, the third eigenvalue, v of the matrix C 1 ,v 2 ,v 3 Are each lambda 1 ,λ 2 ,λ 3 The corresponding feature vector;
s3.4: feature vector v 1 ，v 2 ，v 3 Respectively setting direction vectors of X, Y, Z axes under a coordinate system;
s3.5: and creating a transformation matrix according to the feature vector, wherein the transformation matrix is defined as shown in a formula (5):
wherein, the first and the second end of the pipe are connected with each other,
each row is respectively corresponding to a direction vector on X, Y, Z coordinate axes as a rotation matrix; w is a 1 ,w 2 ,w 3 Respectively represent the offset to be generated in the direction of X, Y, Z axis after transformation; u represents a scaling coefficient of the transform;
s3.6: and carrying out coordinate conversion on the three-dimensional point cloud data of the plate according to the transformation matrix, wherein the three coordinate axis directions after the conversion are respectively as follows: the X axis is horizontally rightward, the Y axis is vertically upward, and the Z axis is vertically outward.
4. The plate defect detection method based on the three-dimensional point cloud as claimed in claim 3, wherein:
said w 1 ,w 2 ,w 3 Default to 0 represents no offset; the u represents the scaling factor of the transform and by default setting it to 1 represents no scaling.
5. The plate defect detection method based on the three-dimensional point cloud as claimed in claim 1, wherein:
the size detection algorithm specifically comprises the following steps: traversing three-dimensional coordinates of all points in the three-dimensional point cloud data, calculating a minimum cuboid bounding box which contains all the points and has edges parallel to coordinate axes according to the range of the coordinate values, calculating the edge length of the minimum cuboid bounding box and sequencing to obtain the length, width and height of the three-dimensional point cloud data of the plate.
6. The plate defect detection method based on the three-dimensional point cloud as claimed in claim 1, wherein:
the cutting algorithm specifically comprises the following steps:
s5.1: obtaining two-dimensional coordinates P of X axis and Y axis of four corner points of the plate according to the vertex information of the cuboid bounding box 1 (x 1 ,y 1 )，P 2 (x 2 ,y 2 )，P 3 (x 3 ,y 3 )，P 4 (x 4 ,y 4 )；
S5.2: according to the actual frame width L of the plate, solving the two dimensions of the X axis and the Y axis of the four corner points of the cutting areaCoordinate P' 1 (x 1 +L,y 1 -L)，P′ 2 (x 2 +L,y 2 +L)，P′ 3 (x 3 -L,y 3 +L)，P′ 4 (x 4 -L,y 4 -L)；
S5.3: from P' 1 、P′ 2 、P′ 3 、P′ 4 And forming a rectangular area by the four points, performing conditional filtering on the three-dimensional point cloud data of the plate in the rectangular area, removing the three-dimensional point cloud data in the rectangular area, and leaving the three-dimensional point cloud data as a warp angle detection area.
7. The plate defect detection method based on the three-dimensional point cloud as claimed in claim 1, wherein:
the plane fitting algorithm specifically comprises the following steps:
s6.1: randomly selecting coordinates Q of three points in three-dimensional point cloud data of a rake angle detection area 1 (x 1 ,y 1 ,z 1 )，Q 2 (x 2 ,y 2 ,z 2 )，Q 3 (x 3 ,y 3 ,z 3 )；
S6.2: according to Q 1 ,Q 2 ,Q 3 The coordinates of the three points are established into a (6) equation set, the value of the unknown coefficient A, B, C, D is calculated, and the plane model equation is determined to be that Ax + By + Cz + D is 0:
s6.3: let (x) i ,y i ,z i ) Calculating the average distance from all K points in the three-dimensional point cloud data of the warp detection area to the point-plane of the planar model equation obtained in S6.2 for the coordinates of any point in the warp detection area
Wherein the distance d between the point and the surface i And average distance
Comprises the following steps:
s6.4: repeating S6.2-S6.3 until reaching the set iteration number, and selecting the average distance of the point and the surface
And the minimum plane equation model is the plane model equation finally fitted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210588295.3A CN115100116A (en) | 2022-05-27 | 2022-05-27 | Plate defect detection method based on three-dimensional point cloud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210588295.3A CN115100116A (en) | 2022-05-27 | 2022-05-27 | Plate defect detection method based on three-dimensional point cloud |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115100116A true CN115100116A (en) | 2022-09-23 |
Family
ID=83288182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210588295.3A Pending CN115100116A (en) | 2022-05-27 | 2022-05-27 | Plate defect detection method based on three-dimensional point cloud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115100116A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116630330A (en) * | 2023-07-26 | 2023-08-22 | 征图新视(江苏)科技股份有限公司 | Triangular mesh plane defect detection method based on edge difference |
| CN117141046A (en) * | 2023-10-30 | 2023-12-01 | 江苏爱箔乐铝箔制品有限公司 | Safety monitoring method and system of aluminum foil cutlery box punch forming machine |
-
2022
- 2022-05-27 CN CN202210588295.3A patent/CN115100116A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116630330A (en) * | 2023-07-26 | 2023-08-22 | 征图新视(江苏)科技股份有限公司 | Triangular mesh plane defect detection method based on edge difference |
| CN117141046A (en) * | 2023-10-30 | 2023-12-01 | 江苏爱箔乐铝箔制品有限公司 | Safety monitoring method and system of aluminum foil cutlery box punch forming machine |
| CN117141046B (en) * | 2023-10-30 | 2023-12-26 | 江苏爱箔乐铝箔制品有限公司 | Safety monitoring method and system of aluminum foil cutlery box punch forming machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111062915B (en) | Real-time steel pipe defect detection method based on improved YOLOv3 model | |
| US10115035B2 (en) | Vision system and analytical method for planar surface segmentation | |
| US6983065B1 (en) | Method for extracting features from an image using oriented filters | |
| CN115100116A (en) | Plate defect detection method based on three-dimensional point cloud | |
| CN112233116B (en) | Concave-convex mark visual detection method based on neighborhood decision and gray level co-occurrence matrix description | |
| CN110060338B (en) | Prefabricated part point cloud identification method based on BIM model | |
| CN115797354B (en) | Method for detecting appearance defects of laser welding seam | |
| CN108830831B (en) | Zinc flotation foam natural speed feature extraction method based on improved SURF matching | |
| CN113470090A (en) | Multi-solid-state laser radar external reference calibration method based on SIFT-SHOT characteristics | |
| CN112613097A (en) | BIM rapid modeling method based on computer vision | |
| CN113538503B (en) | Solar panel defect detection method based on infrared image | |
| CN114627080B (en) | Vehicle stamping accessory defect detection method based on computer vision | |
| CN114119488A (en) | Intelligent size and quality detection method for prefabricated laminated plate facing factory | |
| CN117314901B (en) | Scale-adaptive chip detection neural network system | |
| CN108053485B (en) | Potato three-dimensional modeling method based on contour image | |
| CN115656182A (en) | Sheet material point cloud defect detection method based on tensor voting principal component analysis | |
| CN112561989A (en) | Method for identifying hoisting object in construction scene | |
| CN116596987A (en) | Workpiece three-dimensional size high-precision measurement method based on binocular vision | |
| CN109506569B (en) | Method for monitoring three-dimensional sizes of cubic and columnar crystals in crystallization process based on binocular vision | |
| CN115112098B (en) | Monocular vision one-dimensional two-dimensional measurement method | |
| CN114187269B (en) | Rapid detection method for surface defect edge of small component | |
| CN116125489A (en) | Indoor object three-dimensional detection method, computer equipment and storage medium | |
| CN110136193B (en) | Rectangular box three-dimensional size measuring method based on depth image and storage medium | |
| CN114092396A (en) | Method and device for detecting corner collision flaw of packaging box | |
| CN109544513A (en) | A kind of steel pipe end surface defect extraction knowledge method for distinguishing |
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 |