CN104655041A - Industrial part contour line multi-feature extracting method with additional constraint conditions - Google Patents
Industrial part contour line multi-feature extracting method with additional constraint conditions Download PDFInfo
- Publication number
- CN104655041A CN104655041A CN201510001034.7A CN201510001034A CN104655041A CN 104655041 A CN104655041 A CN 104655041A CN 201510001034 A CN201510001034 A CN 201510001034A CN 104655041 A CN104655041 A CN 104655041A
- Authority
- CN
- China
- Prior art keywords
- contour
- contour line
- constraint conditions
- line
- matrix
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000000605 extraction Methods 0.000 claims abstract description 38
- 239000013598 vector Substances 0.000 claims description 36
- 239000011159 matrix material Substances 0.000 claims description 33
- 239000000203 mixture Substances 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 abstract description 4
- 238000007796 conventional method Methods 0.000 abstract 1
- 238000001514 detection method Methods 0.000 description 18
- 230000011218 segmentation Effects 0.000 description 8
- 238000005259 measurement Methods 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000011179 visual inspection Methods 0.000 description 4
- 238000012952 Resampling Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 239000002023 wood Substances 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a part contour line multi-feature extracting method with additional constraint conditions. The method is characterized in that a contour line feature extracted by the conventional method is used as an initial value, and the constraint conditions that an arc is tangent with a straight line, every two adjacent arcs are tangent with each other and the like are added; furthermore, on the basis of the constraint conditions, and error equations of part contour line multi-feature basic conditions and the constraint conditions are constructed, so that integral iteration precise calculation of contour features is realized, and all the features (such as the straight lines, the arcs and the circles) forming the contour can be precisely identified section by section, and parameters of all the features are acquired. The method puts emphasis on improving the precision of part contour line extraction, and a result shows that by introduction of a constraint relation, the precision of the contour line multi-feature extraction can be effectively improved.
Description
Technical Field
The invention relates to a multi-feature extraction method of an industrial part contour line with additional constraint conditions, and belongs to the field of industrial part visual detection.
Background
With the continuous development and progress of computers and computer vision, the requirement for detecting parts by machine vision instead of human eyes is more and more urgent in the detection of industrial parts at present. The computer vision detection has many advantages, such as high detection precision, high speed, no contact and good accuracy, and in addition, the price of hardware such as computers and cameras is not high at present, and the price of human resources is continuously increased.
During the visual inspection of the part, pass throughDrawing (A)After image edge extraction, from the part to be measuredDrawing (A)The image can obtain a plane contour pixel point set belonging to the same plane contour pixel pointDrawing (A)The outline pixel point set of the element can form basic geometry such as straight line, circular arc and the likeDrawing (A)And (5) Yuan. When detecting parameters such as shape and size of a part in two-dimensional measurement, the method is generally based onDrawing (A)Size, shape and each of the elementsDrawing (A)The positional relationship of the elements, etc., are detected, and therefore, each of the elements constituting the contour of the part must be recognized before detectionDrawing (A)Features of the elements. In three-dimensional measurement, the study of visual measurement reconstruction using the contour line of a part is also carried out, so that the contour is basically reconstructedDrawing (A)The accurate segmentation and identification of the element is a key step in the visual inspection process, and the accuracy of the measurement of the dimension of the part is directly influenced. A plurality of scholars at home and abroad carry out extensive research on the characteristic angular points, such as Hsin-Teng Sheu, Wu-Chih Hu, Alexander Kolesnikov, wood steel and the like, and the characteristic angular points are extracted by combining methods of curvature, projection height, merging and splitting and the like, so that circular arcs, straight lines and the like are usedDrawing (A)The meta describes feature information of the outline. But is not fully utilized in the extraction processDrawing (A)Inherent relationships (e.g., tangent relationships) exist between elements, which necessarily affect the accuracy of corner extraction. The accuracy of contour line multi-feature extraction is improved,the method is a technical problem to be solved urgently for improving the detection precision in the field of industrial part detection.
Therefore, considering the characteristics of industrial part processing, the invention adds relevant constraint conditions on the basis of comprehensively utilizing the existing contour feature angular point detection method, realizes accurate segmentation of the contour by utilizing straight lines and circular arcs, and improves the precision of contour line multi-feature extraction.
Disclosure of Invention
The invention aims to solve the defects of the prior art, improve the contour line multi-feature extraction precision frequently used in the visual inspection of industrial parts and provide a multi-feature extraction scheme for additional constraint conditions of the contour line of the industrial parts. The method comprises the steps of adding constraint conditions such as tangency of arcs on the outline and tangency of adjacent arcs on the outline on the basis of extracting initial multiple features of the part outline, carrying out iterative accurate extraction on the features on the outline, and carrying out accurate segmentation identification on each feature forming the outline, such as the straight line, the arcs and the circles, so as to obtain parameters of each feature.
In order to solve the technical problems, the invention adopts the following technical scheme:
the method for extracting the multiple characteristics of the contour line of the industrial part with additional constraint conditions comprises the following steps:
(1) extracting multiple characteristics of the contour line of the industrial part from the actual part image to be used as an initial value of the method;
(2) establishing a multi-feature basic condition and constraint condition error equation on the part contour line;
(3) and (4) integrally and accurately solving the contour line multi-feature parameters with the constraint conditions.
In the step (1), because the visual inspection system is generally set to be a semi-closed environment and has good illumination conditions, the contour line multi-feature extraction method with additional constraint conditions is obtainedThe contrast ratio of the image target and the background color is large, so that a traditional edge extraction algorithm can be used for obtaining an initial contour, then an eight-neighborhood tracking is used for obtaining a contour pixel point set which is arranged in sequence from a starting point to an end point, before contour characteristic corner points are extracted, in order to reduce the calculation amount, a neighborhood value comparison method is used for removing contour points which cannot be corner points, and a rotation-invariant corner point judgment method and an angle-based judgment method introduced in the existing literature are used for removing redundant corner points caused by linear resampling. But still have many redundant angular points on the circular arc to remove, the removal method is: first, the corner points are divided into two types according to the characteristics of the neighborhood of the corner points, wherein the first type is an abrupt corner point (Type), curve segments on both sides of the corner point belong to different characteristics, the second is smooth corner pointType), it may be a tangent point of a straight line and a circular arc or a redundant angular point on a circular arc, and curve segments on both sides of it may belong to the same feature and may need to be removed; then using adaptationThe curvature method distinguishes the corner point type.
To be basic to the profileDrawing (A)The elements are identified by segmentation and determinedDrawing (A)Type of element, i.e. judgmentDrawing (A)Whether the element is a straight line or a circular arc. The invention judges the distance between two contour angular points based on the point projection height methodDrawing (A)Meta-types, thereby enabling to further classify corner points into the following four types: (1)type, i.e., line to line; (2)type, i.e., arc to line; (3)type, i.e., line to arc; (4)type, i.e., arc to arc (arc).
Consider the foregoingType andthe invention finally divides the angular points into the following types:，，，，，andseven sub-classes. Such asRepresenting smooth corner points connecting the arcs and the straight lines,a projection for connecting an arc and a straight lineAnd changing the corner point. For a smooth feature corner point, the two end curve segments may belong to the same feature and be merged.
Drawing (A)The identification of the meta-parameters is determined according to the corner type of two end points of the contour segment, but omission is inevitable in the detection of the characteristic corners, for exampleThe corner points are followed byCorner points, known from logical relationsAndmust exist betweenCorner points, so that it is now necessary to insert between themThe corner point method is as follows: by straight-line connectionCorner point andcorner point, then fromCorner point andfinding out the point with the farthest distance from the connecting line of the two corner points from the contour points between the corner points, i.e. the point is insertedA corner point.
For the case of continuous arc features, it needs to be judged whether the arc features can be combined or divided: firstly, estimating the circle center and the radius of the circular arc; then the ratio of the sum of the distance from each point to the center of the circle and the difference value of the radius to the length of the arc is utilizedThe arc discrimination function of (2) merging or splitting arcs according to the following criteria: after the two arcs mergeWhen the value is smaller, the two end arcs are combined, and when one is insertedAfter the corner point, calculatedWhen the value is smaller, the arc is divided; and finally obtaining an accurate arc segmentation result.
When the feature corner points on the contour are identified, the part contour can be represented by straight lines and circular arcs.Type andthe curve segments between the model corner points are expressed by straight lines,type andthe curve segments between the model corners are expressed by circular arcs.
In step (2), the contour line multi-feature extraction method with additional constraint condition is proved to be only used for straight line in the prior researchThe free contour information can be well expressed by the lines and the circular arcs, so the invention describes how to carry out the addition of the constraint conditions by the contour comprising the lines and the circular arcsDrawing (A)The method comprises the following steps of performing combined extraction on meta-features, wherein the specific process comprises the establishment of a basic conditional error equation and a constraint conditional error equation.
1) Basic conditions.
(1) The linear equation expression is:
（1）
the error equation is:
（2）。
(2) the equation for the circle is expressed as:
（3）
the error equation is:
（4）
and expressing the two types of error equations by using a matrix to obtain:
（5）
wherein:；a vector representing the composition of all the straight-line parameters,a vector representing the composition of all the circular parameters,a coefficient matrix composed of all the linear observation error equation coefficients,for the corresponding constant term vector, the constant term vector,then the corresponding residual vector is obtained;a coefficient matrix composed of all the circular observation error equation coefficients,for the corresponding constant term vector, the constant term vector,then the corresponding residual vector.
2) A constraint condition.
(1) For the general form of the straight-line equation,representing the normal vector of the line, so the line parametersAndsatisfies the following conditions:
（6）
the linearization can result in:
（7）
for this class of constraints, i.e., all equations (7), the matrix form is written as:
（8）
whereinFor the matrix of coefficients thereof,is the corresponding constant vector.
(2) Constraint condition of tangency of circular arc and straight line:
（9）
order to （10）
Then its linearized form is:
（11）
constraints of this type can be written in matrix form:
（12）
whereinAndare respectively asAnda corresponding matrix of coefficients is then formed,are the corresponding constant vectors.
(3) Constraint condition of tangency of the circular arc and the circular arc:
let two arc parameters be respectively，The tangent condition is (both internally tangent and externally tangent):
（13）
wherein: （14）
the linearized form of this constraint is:
（15）
writing these constraints in matrix form (all constraints in (equation 15)) then is
（16）
WhereinIn the form of a matrix of coefficients,then it is a constant vector.
In order to realize the purpose of the invention, the contour line multi-feature extraction method with additional constraint conditions comprises the following specific steps of (3) firstly establishing a parameter estimation integral adjustment model with the constraint conditions, thereby realizing the integral accurate solution of contour line multi-feature parameters:
order toAccording to the above error equation and constraint conditions, the parameter estimation adjustment model adopted by the invention can be expressed as follows (using a unit weight matrix):
（17）
introducing auxiliary Lagrange coefficientsThe langerhan function is constructed as follows:
（18）
using Lagrange extremum criterion, forThe partial derivatives of the parameters are calculated and set to zero, and the following normal equation can be obtained:
（19）
calculating all contour points according to the specific conditions by the following equations (5), (8), (12) and (16)Andthe matrix is then substituted into (equation 19), and the contours can be iteratively solvedDrawing (A)The exact parameter values of the meta-features.
Drawings
FIG. 1 shows a schematic view of aDesign structure schematic of part multi-feature extraction hardware system with constraint conditions attached to embodimentDrawing (A)。
FIG. 2Part multi-feature extraction hardware system real object with constraint conditions attached to embodimentDrawing (A)。
FIG. 3The method is an operation flow in contour line multi-feature detection with constraint conditions.
FIG. 4And precisely extracting the embodiment process of the contour line multi-feature with the constraint condition.
FIG. 5Arc splitting and merging schematic in contour multi-feature initial value extractionDrawing (A)。
FIG. 6To simulateDrawing (A)Image profileDrawing (A)And extracting a result of the meta-feature corner points.
FIG. 7As a practical partDrawing (A)Image profileDrawing (A)Meta-multi-feature extraction result and detail amplificationDrawing (A)。
FIG. 8Comparing the extracted corner positions under the conditions of no constraint condition and constraint condition.
FIG. 9For parts with circular arcs abuttingDrawing (A)Image profileDrawing (A)Meta-multi-feature extraction result detailsDrawing (A)。
FIG. 1 shows a schematic view of a0 is the extraction result without restriction and the extraction result with restriction.
In the drawings: the device comprises a telecentric lens 1, a CCD camera 2, an LED light source 3, a workpiece platform 4, a stand column 5, a computer 6, a working platform 7, a workpiece 8 and a height adjusting knob 9.
Detailed Description
For the sake of reference, the embodiment of the invention provides a complete constraint-attached part contour line multi-feature extraction practical operation process from shooting. The prior art of removing the shooting mode and extracting the initial value of the multi-feature angular point of the part contourIn addition, the invention provides the originality obtained on the photographed partDrawing (A)The method is a new technical scheme for improving the multi-feature detection precision of parts by processing on the basis of the extracted initial contour angular points, and can be automatically realized by adopting a computer means. The following combinationsDrawingsAnd the embodiment illustrates the technical scheme of the invention.
See alsoFIG. 1 of the drawingsThe hardware system for extracting and detecting the multiple features of the part contour line of the embodiment comprises the following parts:
(1) a camera with a telecentric lens. The system is provided with a USB or 1394 data line, data are transmitted to a computer in real time, and experiments prove that the distortion of a telecentric lens is negligible;
(2) and (4) a computer. The invention provides a part multi-feature extraction method with constraint conditions, which can be loaded on a computer by adopting a software technology;
(3) a backlight source. The backlight source of the embodiment is an LED light source, is externally connected with an adjustable power line and can adjust the brightness of the LED. The background light is an illumination mode that a camera and a light source are placed on different sides of a measured object, most of images acquired by the mode are projections of opaque objects, and the shielded parts are black, otherwise, the images are white, so that the images are clear in black and white and have protruding edges, the edge detection is particularly facilitated, and the method is suitable for the fields of shape recognition, position and size measurement and the like of the opaque objects;
(4) a part platform. The device is used for placing the detected industrial parts;
(5) and (4) a total support. The device is used for supporting and fixing the camera, the lens, the backlight source and the part platform. See alsoFIG. 2 of the drawingsThe support can be adjusted during specific implementation so that the lens of the camera is aligned with the industrial part on the part platform, and the backlight source can uniformly provide light for clear acquisition of the contour image of the industrial part.
The specific steps of the operation flow in the detection are as followsFIG. 3 of the drawings: placing the planar industrial part to be detected on the table top of the part platform, and adjusting the planar industrial part to be detected through an adjustable power lineAnd (4) adjusting the brightness of the backlight source (if the brightness is proper, the adjustment is not needed), and photographing the industrial parts by using a camera after the adjustment is completed. Shooting industrial parts to obtain part originalsDrawing (A)The image is transmitted to the computer through the connection between the camera and the computer, the computer is atDrawing (A)And on the basis of the image, executing contour multi-feature detection with constraint conditions on the basis of extracting the initial part contour feature angular points in the existing literature, and then outputting a detection result.
The technical scheme of the part contour multi-feature detection with additional constraint conditions executed by the computer in the embodiment of the invention is as follows.
Step 1, sending out from an actual part image, acquiring an initial value of a feature angular point on a contour by using a contour line feature angular point extraction method, and taking the obtained initial contour line feature as the initial value of the method.
Step 1.1, extracting and identifying the segmentation angular points of the contour line segments of the parts.
Because the vision detection system is generally set to be a semi-closed environment and has good lighting conditions, the image quality of the object to be detected obtained is generally good, the contrast between the target and the background is high, and the initial contour of the part can be obtained by utilizing an image edge extraction algorithm and then performing edge refinement; acquiring a contour pixel point set which is sequentially arranged from a starting point to an end point by using an eight-neighborhood tracking method after an initial contour is obtained; before extracting the characteristic angular points of the contour, in order to reduce the calculation amount, firstly, the contour points are classified in advance by adopting a neighborhood value comparison method, the contour points which cannot be taken as the angular points are removed, and the remaining contour points are removed by adopting an angular point judgment method with unchanged rotation and an angle-based judgment method to remove redundant angular points caused by linear resampling.
And a plurality of redundant angular points still exist on the circular arc of the angular point extraction result obtained through the steps and need to be further removed. The removing method comprises the following steps: first, the corner points are divided into two types according to the characteristics of the neighborhood of the corner points, wherein the first type is an abrupt corner point (Type), curve segments on both sides of the corner point belong to different characteristics, the second is smooth corner pointType), it may be a tangent point of a straight line and a circular arc or a redundant angular point on a circular arc, and curve segments on both sides of it may belong to the same feature and may need to be removed; then using adaptationThe curvature method distinguishes the corner point type.
To be basic to the profileDrawing (A)The elements are identified by segmentation and determinedDrawing (A)Type of element, i.e. judgmentDrawing (A)Whether the element is a straight line or a circular arc. The invention judges the distance between two contour angular points based on the point projection height methodDrawing (A)Meta-types, thereby enabling to further classify corner points into the following four types: (1)type, i.e., line to line; (2)type, i.e., arc to line; (3)type, i.e., line to arc; (4)type, i.e., arc to arc (arc).
Consider the foregoingType andthe invention finally divides the corner points intoThe following classes:，，，，，andseven sub-classes. Such asRepresenting smooth corner points connecting the arcs and the straight lines,representing the abrupt corner connecting the arc and the line.
Step 1.2, basic conditions common to industrial partsDrawing (A)And identifying the meta-parameters.
When the characteristic angular points on the contour are identified, the contour of the part can be represented by straight lines and circular arcs. If the characteristic angular point is an abrupt angular point, the curve segments at the two ends of the characteristic angular point are not combined, and if the characteristic angular point is smooth, the curve segments at the two ends of the characteristic angular point are possibly the same characteristic and are possibly combined.Type andthe curve segments between the model corner points are expressed by straight lines,type andthe curve segments between the model corners are expressed by circular arcs, and if the two conditions are not met, new corners need to be inserted.
And step 1.2.1, judging the linear characteristics.
Drawing (A)The identification of the meta-parameters is determined according to the corner type of two end points of the contour segment, but omission is inevitable in the detection of the characteristic corners, for exampleThe corner points are followed byCorner points, known from logical relationsAndmust exist betweenCorner points, which are inserted between themThe corner point method comprises the following steps: by straight-line connectionCorner point andcorner point, then fromCorner point andfinding out the point with the farthest distance from the connecting line of the two corner points from the contour points between the corner points, i.e. the point is insertedA corner point. Corner points on the contourAnd corner pointThe line segments between the two points are straight line featuresOr) And fitting by using a linear equation.
And 1.2.2, segmenting and fusing circular arcs.
Considering the possibility of merging and dividing the continuous arc features at the same time, it is necessary to judge whether the continuous arcs can be merged or need to be divided: firstly, estimating the circle center and the radius of the circular arc; then the ratio of the sum of the distance from each point to the center of the circle and the difference value of the radius to the length of the arc is utilizedThe arc discrimination function of (2) merging or splitting arcs according to the following criteria: after the two arcs mergeWhen the value is smaller, the arcs at the two ends are combined; when inserting oneAfter the corner point, calculatedWhen the value is smaller, the arc is divided.
The method for splitting and combining the circular arcs comprises the following specific steps:
such asFIG. 5 shows a schematic representation of the structure(a) Shown by usingRepresenting a continuous circular segment, in whichAndis composed of，，，Or，Represents the ith segmentAngular point, then:
(1) using pairs of circular equationsIn the middle of the contour pointFitting to obtainThe center and radius of the arc are calculatedValue, order，；
(2) Will be provided withAdding and utilizingBetween contour points to re-fit the centre and radius of the arc, i.e. to calculateAnd the center and radius of the circle, and calculating itA value;
(3) if it is notThen mergeIs a segment of circular arc and makes，Then returning to the step 2; otherwise make；
(4) In thatInterject intoTo split the segment of the arc and then calculateAnd the center and radius of the circle, and calculating itA value;
(5) if it is notDescription of the inventionIf the identification is unique, directly jumping to the step (7);
(6) if it is notThen, then，On the contrary orderReturning to the step (4);
(7) in thatIs inserted with a new corner pointAnd is removedAll corner points in between. Then, the judgment is continuedAndin the case of (1): order toThen go back to the first step to judge whether it is newAnd fusing the splitting condition by the inter-arc till the end.
As can be seen from the above method, the steps (2) and (3) are used for merging the circular arcs, and the steps (4) and (6) are used for splitting the circular arcs, so that the two purposes of merging and splitting can be achieved simultaneously. The plane industrial part contour can directly use the method to realize the arc splitting or merging.
And 2, integrally solving the contour line multi-feature parameters with the constraint conditions, wherein the calculation process is as follows.
1) Basic conditions.
(1) The linear equation expression is:
（1）
the error equation is:
（2）。
(2) the equation for the circle is expressed as:
（3）
the error equation is:
（4）
and expressing the two types of error equations by using a matrix to obtain:
（5）
wherein:；a vector representing the composition of all the straight-line parameters,a vector representing the composition of all the circular parameters,a coefficient matrix composed of all the linear observation error equation coefficients,for the corresponding constant term vector, the constant term vector,then the corresponding residual vector is obtained;a coefficient matrix composed of all the circular observation error equation coefficients,for the corresponding constant term vector, the constant term vector,then the corresponding residual vector.
2) A constraint condition.
(1) For the general form of the straight-line equation,representing the normal vector of the line, so the line parametersAndsatisfies the following conditions:
（6）
the linearization can result in:
（7）
for this class of constraints, i.e., all equations (7), the matrix form is written as:
（8）
whereinFor the matrix of coefficients thereof,is the corresponding constant vector.
(2) Constraint condition of tangency of circular arc and straight line:
（9）
order to （10）
Then its linearized form is:
（11）
constraints of this type can be written in matrix form:
（12）
whereinAndare respectively asAnda corresponding matrix of coefficients is then formed,are the corresponding constant vectors.
(3) Constraint condition of tangency of circular arc and circular arc
Let two arc parameters be respectively，The tangent condition is (both internally tangent and externally tangent):
（13）
wherein: （14）
the linearized form of this constraint is:
（15）
wherein,（16）
writing these constraints in matrix form (all constraints in (equation 15)) then is
（17）
WhereinIn the form of a matrix of coefficients,then it is a constant vector.
3) Parameter estimation integral adjustment model with constraint condition
Order toAccording to the above error equation and constraint conditions, the parameter estimation adjustment model adopted by the invention can be expressed as follows (using a unit weight matrix):
（18）
introducing auxiliary Lagrange coefficientsThe langerhan function is constructed as follows:
（19）
using Lagrange extremum criterion, forThe partial derivatives of the parameters are calculated and set to zero, and the following normal equation can be obtained:
（20）
calculating all contour points according to the specific conditions by the following equations (5), (8), (12) and (17)Andthe matrix is then substituted into (equation 20), and the contours can be obtained by iterative solutionDrawing (A)The exact parameter values of the meta-features.
In order to verify the effectiveness of the contour line multi-feature extraction method with additional constraint conditions, simulation is respectively adoptedDrawing (A)Image and actual partDrawing (A)Quantitative and qualitative tests were performed.
Considering that the actual value of the actual part cannot be obtained in the actual test, the design is carried out according to the known part sizeDrawing (A)Draw a simulationDrawing (A)The image is quantitatively analyzed and verified by the method provided by the invention. For simulationDrawing (A)Like, the extraction result of the method of the present inventionAs shown in fig. 6(a) And 6(b) comprises 5 straight lines, 2 sections of circular arcs and 3 circles, wherein for the extraction results of the two sections of circular arcs, the comparison of the extraction values of the constraint conditions and the constraint conditions which are not attached with the simulation design values is shown in the specificationTABLE 1（x0、y0R in pixels, sita in °). FromTABLE 1The introduction of the constraint condition can be found in the comparison, so that the extraction result is more accurate, and the method can quantitatively realize the accurate extraction of the multiple features of the part contour line.
For actual parts obtained in examplesDrawing (A)Image of a part, the result is asFIG. 7 of the drawingsAnd 9. These results show that: the method provided by the invention can accurately extract the contourDrawing (A)The angular point of the division between the elements, even for the parts with continuous arcs adjacent to each other, can also accurately extract the arc tangent point as the division point (such asFIG. 9 of the drawings) (ii) a In addition fromFIG. 8 of the drawingsComparison of details of FIGS. 10In the drawingsThe outline of the additional constraint can be seenDrawing (A)The extraction position of the meta-feature segmentation corner point is more accurate and reasonable.
When the quality of the part is detected, the extracted result is compared with the design value of the contour parameter of the part, if the difference value is within the allowable range of the condition, the quality of the part is qualified, otherwise, the quality of the part is unqualified. The judgment result is used as the result of the detection process and can be output through the computer display equipment to prompt that unqualified parts are processed.
Claims (3)
1. The multi-feature extraction method of the contour line of the industrial part with additional constraint conditions is characterized by comprising the following steps in sequence:
(1) extracting multiple characteristics of the contour line of the industrial part from the image of the actual industrial part as the initial value of the method;
(2) establishing a multi-feature basic condition and constraint condition error equation on the part contour line;
(3) and (4) integrally and accurately solving the contour line multi-feature parameters with the constraint conditions.
2. The method for extracting multiple features from the contour line of an industrial part with additional constraint conditions as claimed in claim 1, wherein in step (2), since it has been proved that the free contour information can be well expressed only by using straight lines and circular arcs, the invention describes how to perform the joint extraction of the multiple primitive features with additional constraint conditions by using the contour including straight lines and circular arcs, and the specific process comprises the establishment of a basic condition error equation and a constraint condition error equation:
1) basic conditions
(1) The linear equation expression is:
（1）
the error equation is:
（2）
(2) the equation for the circle is expressed as:
（3）
the error equation is:
（4）
and expressing the two types of error equations by using a matrix to obtain:
（5）
wherein:；a vector representing the composition of all the straight-line parameters,a vector representing the composition of all the circular parameters,a coefficient matrix composed of all the linear observation error equation coefficients,for the corresponding constant term vector, the constant term vector,then the corresponding residual vector is obtained;a coefficient matrix composed of all the circular observation error equation coefficients,for the corresponding constant term vector, the constant term vector,then the corresponding residual vector is obtained;
2) constraint conditions
(1) For the general form of the straight-line equation,representing the normal vector of the line, so the line parametersAndsatisfies the following conditions:
（6）
the linearization can result in:
（7）
for this class of constraints, i.e., all equations (7), the matrix form is written as:
（8）
whereinFor the matrix of coefficients thereof,is a corresponding constant vector;
(2) constraint condition of tangency of circular arc and straight line:
（9）
order to （10）
Then its linearized form is:
（11）
constraints of this type can be written in matrix form:
（12）
whereinAndare respectively asAnda corresponding matrix of coefficients is then formed,are the corresponding constant vectors;
(3) constraint condition of tangency of the circular arc and the circular arc:
let two arc parameters be respectively，The tangent condition is (both internally tangent and externally tangent):
（13）
wherein: （14）
the linearized form of this constraint is:
（15）
writing these constraints in matrix form (all constraints in (equation 15)) then is
（16）
WhereinIn the form of a matrix of coefficients,then it is a constant vector.
3. The method for extracting multiple features of the contour line of an industrial part with additional constraint conditions as claimed in claim 1, wherein in step (3), a parameter estimation integral adjustment model with additional constraint conditions is firstly established, so as to realize the integral accurate solution of the multiple feature parameters of the contour line, and the specific process is as follows:
order toThe parameter estimation global adjustment model used in the present invention can be expressed as follows (using the unit weight matrix) according to the error equation of the basic condition and the constraint condition established in claim 2:
（17）
introducing auxiliary Lagrange coefficientsThe langerhan function is constructed as follows:
（18）
using Lagrange extremum criterion, forThe partial derivatives of the parameters are calculated and set to zero, and the following normal equation can be obtained:
（19）
calculating all contour points according to the specific conditions by the following equations (5), (8), (12) and (16)Andand (3) substituting the matrix into a parameter estimation integral adjustment model (formula 19) with constraint conditions, and then iteratively solving to obtain accurate parameter values of the characteristics of each primitive of the contour.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510001034.7A CN104655041B (en) | 2015-01-05 | 2015-01-05 | A kind of industrial part contour line multi-feature extraction method of additional constraint condition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510001034.7A CN104655041B (en) | 2015-01-05 | 2015-01-05 | A kind of industrial part contour line multi-feature extraction method of additional constraint condition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104655041A true CN104655041A (en) | 2015-05-27 |
| CN104655041B CN104655041B (en) | 2018-11-23 |
Family
ID=53246453
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510001034.7A Expired - Fee Related CN104655041B (en) | 2015-01-05 | 2015-01-05 | A kind of industrial part contour line multi-feature extraction method of additional constraint condition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104655041B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105108358A (en) * | 2015-09-15 | 2015-12-02 | 南京理工大学 | Real-time detection and control device and method for slit width image of carbon dioxide laser cut die plate |
| CN107292893A (en) * | 2016-03-31 | 2017-10-24 | 富士通株式会社 | Image processing apparatus, image processing method and image processing equipment |
| CN108734054A (en) * | 2017-04-17 | 2018-11-02 | 湖南生物机电职业技术学院 | Unobstructed citrusfruit image-recognizing method |
| CN109509182A (en) * | 2018-10-29 | 2019-03-22 | 首都航天机械有限公司 | A kind of typical products geometric dimension measurement method and system based on image procossing |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101802548A (en) * | 2007-07-24 | 2010-08-11 | 海克斯康测量技术有限公司 | Method of compensating measurement errors of a measuring machine deriving from the deformations of the machine bed caused by the load exerted by the mobile unit of the machine on the machine bed, and measuring machine operating according to said method |
| CN102129684A (en) * | 2011-03-17 | 2011-07-20 | 南京航空航天大学 | Method for matching images of different sources based on fit contour |
| CN102425989A (en) * | 2011-08-22 | 2012-04-25 | 天津大学 | Image detection-based two-dimensional characteristic size measurement method |
| CN103473766A (en) * | 2013-09-03 | 2013-12-25 | 浙江工商大学 | Two-dimensional shape feature point matching method based on visual perception |
| CN103914690A (en) * | 2014-04-11 | 2014-07-09 | 大连理工大学 | Shape matching method based on projective invariant |
| CN104048645A (en) * | 2014-06-25 | 2014-09-17 | 山东理工大学 | Integral orientation method of ground scanning point cloud through linear fitting |
| WO2014169904A1 (en) * | 2013-04-19 | 2014-10-23 | Schoen + Sandt Machinery Gmbh | Checking method |
-
2015
- 2015-01-05 CN CN201510001034.7A patent/CN104655041B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101802548A (en) * | 2007-07-24 | 2010-08-11 | 海克斯康测量技术有限公司 | Method of compensating measurement errors of a measuring machine deriving from the deformations of the machine bed caused by the load exerted by the mobile unit of the machine on the machine bed, and measuring machine operating according to said method |
| CN102129684A (en) * | 2011-03-17 | 2011-07-20 | 南京航空航天大学 | Method for matching images of different sources based on fit contour |
| CN102425989A (en) * | 2011-08-22 | 2012-04-25 | 天津大学 | Image detection-based two-dimensional characteristic size measurement method |
| WO2014169904A1 (en) * | 2013-04-19 | 2014-10-23 | Schoen + Sandt Machinery Gmbh | Checking method |
| CN103473766A (en) * | 2013-09-03 | 2013-12-25 | 浙江工商大学 | Two-dimensional shape feature point matching method based on visual perception |
| CN103914690A (en) * | 2014-04-11 | 2014-07-09 | 大连理工大学 | Shape matching method based on projective invariant |
| CN104048645A (en) * | 2014-06-25 | 2014-09-17 | 山东理工大学 | Integral orientation method of ground scanning point cloud through linear fitting |
Non-Patent Citations (1)
| Title |
|---|
| 郑顺义等: "基于模型和广义点摄影测量的圆柱体自动三维重建与检测", 《测绘学报》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105108358A (en) * | 2015-09-15 | 2015-12-02 | 南京理工大学 | Real-time detection and control device and method for slit width image of carbon dioxide laser cut die plate |
| CN107292893A (en) * | 2016-03-31 | 2017-10-24 | 富士通株式会社 | Image processing apparatus, image processing method and image processing equipment |
| CN108734054A (en) * | 2017-04-17 | 2018-11-02 | 湖南生物机电职业技术学院 | Unobstructed citrusfruit image-recognizing method |
| CN109509182A (en) * | 2018-10-29 | 2019-03-22 | 首都航天机械有限公司 | A kind of typical products geometric dimension measurement method and system based on image procossing |
| CN109509182B (en) * | 2018-10-29 | 2021-03-26 | 首都航天机械有限公司 | Typical product geometric dimension measuring method and system based on image processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104655041B (en) | 2018-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Concrete crack assessment using digital image processing and 3D scene reconstruction | |
| CN108369650B (en) | Method for identifying possible characteristic points of calibration pattern | |
| US9443297B2 (en) | System and method for selective determination of point clouds | |
| JP7133283B2 (en) | Systems and methods for efficiently scoring probes in images with a vision system | |
| CN111583114B (en) | Automatic measuring device and measuring method for pipeline threads | |
| CN104634242A (en) | Point adding system and method of probe | |
| JP7505866B2 (en) | Inspection support method, inspection support system, and inspection support program | |
| CN104655041B (en) | A kind of industrial part contour line multi-feature extraction method of additional constraint condition | |
| CN108257125B (en) | Depth image quality non-reference evaluation method based on natural scene statistics | |
| CN112446870A (en) | Pipeline damage detection method, device, equipment and storage medium | |
| JP2016119090A (en) | System and method for determining clutter in acquired image | |
| CN113393439A (en) | Forging defect detection method based on deep learning | |
| Shrestha et al. | A real world dataset for multi-view 3d reconstruction | |
| Stent et al. | Precise deterministic change detection for smooth surfaces | |
| CN113538399A (en) | Method for obtaining accurate contour of workpiece, machine tool and storage medium | |
| JP7230722B2 (en) | Image processing device and image processing method | |
| WO2015140484A1 (en) | A method, apparatus, system, and computer readable medium for enhancing differences between two images of a structure | |
| US20220198707A1 (en) | Method and apparatus with object pose estimation | |
| CN109214398B (en) | Method and system for measuring rod position from continuous images | |
| JP2013182416A (en) | Feature amount extraction device, feature amount extraction method, and feature amount extraction program | |
| CN111681270A (en) | Method, device and storage medium for realizing registration between image frames | |
| JPWO2020039470A1 (en) | Image processing system | |
| CN110772258A (en) | Multi-view ranging method for measuring human body size | |
| Dou et al. | Internal thread defect detection system based on multi-vision | |
| CN117689652A (en) | Quality evaluation method for multi-camera image fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181123 |