CN109580630B - Visual inspection method for defects of mechanical parts - Google Patents
Visual inspection method for defects of mechanical parts Download PDFInfo
- Publication number
- CN109580630B CN109580630B CN201811334772.3A CN201811334772A CN109580630B CN 109580630 B CN109580630 B CN 109580630B CN 201811334772 A CN201811334772 A CN 201811334772A CN 109580630 B CN109580630 B CN 109580630B
- Authority
- CN
- China
- Prior art keywords
- image
- defect
- dimensional
- mechanical part
- defects
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/002—Measuring arrangements characterised by the use of optical techniques for measuring two or more coordinates
- G01B11/005—Measuring arrangements characterised by the use of optical techniques for measuring two or more coordinates coordinate measuring machines
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
- G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
- G01N23/04—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material
- G01N23/046—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material using tomography, e.g. computed tomography [CT]
-
- 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
- G06T7/11—Region-based segmentation
-
- 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
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/10—Different kinds of radiation or particles
- G01N2223/101—Different kinds of radiation or particles electromagnetic radiation
- G01N2223/1016—X-ray
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/40—Imaging
- G01N2223/401—Imaging image processing
-
- 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
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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 visual inspection method for defects of mechanical parts, which can continuously scan the defect parts inside and on the surface of the mechanical parts, and process and analyze the acquired images to obtain the areas, the defect mass center coordinates and the defect shapes of the defects on the surface and inside, so that an operator can clearly see the defect parts; and then, restoring a three-dimensional model of the texture, the color characteristic and the internal structure of the surface of the mechanical part through three-dimensional reconstruction reverse engineering design, so as to present the positions of the defects and the defect range of the surface and the internal structure of the mechanical part, and present a three-dimensional reconstruction system of the defects as a three-dimensional PDF document.
Description
Technical Field
The invention relates to a visual inspection method, in particular to a visual inspection method for defects of mechanical parts.
Background
With the development of industrial technology, the machining mode of mechanical parts shows the development trend of high precision, high speed and automation, the traditional contact detection method has the advantages of low speed, easiness in damaging parts and capability of reducing the measurement precision, and the computer vision-based measurement technology has the advantages of high speed, good real-time performance, non-contact, low cost and the like and is widely applied to precision detection of machining defects of various parts.
The detection and inspection is one of the most basic activities in the manufacturing process, quality information of the product and the manufacturing process thereof is provided through the detection and inspection activities, and the manufacturing process of the product is controlled according to the information, namely, correction and compensation activities are carried out, so that defective products and repair product rates are reduced to the minimum degree, and the stability of the product quality forming process and the consistency of the produced products are ensured; the traditional detection and inspection mainly depends on people and is mainly finished by a manual mode, so that the time and the cost are consumed, the production period is prolonged, and the production cost is increased; meanwhile, the traditional inspection and detection activities are mainly carried out after the processing and manufacturing processes, and once defective products are detected, the loss of the defective products occurs; in addition, because the information detected manually often contains the influence of human errors, the manufacturing process is controlled according to the information, and the information can be implemented only after the process, and human errors can be introduced; therefore, real-time or on-machine process control cannot be achieved by relying on such information; new tests and assays are often based on a variety of advanced sensing technologies and are easily integrated with computer systems; with the support of appropriate software, such automated inspection or verification systems can automatically perform data acquisition, processing, feature extraction and recognition, as well as various analyses and calculations.
The invention provides a visual detection method for defects of mechanical parts, which adopts the advantages of high-precision measurement of a three-dimensional scanning technology and a tomography scanning technology to restore the defect structure of the mechanical parts and the precise size and the precise position of a damaged part so as to facilitate the development of subsequent more precise work.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a visual inspection method for the defects of mechanical parts, which enables an operator to clearly see the defective parts by continuously scanning the defective parts inside and on the surface of the mechanical parts and processing and analyzing the acquired images to obtain the areas, the defect mass center coordinates and the defect shapes of the defects on the surface and inside; and then, restoring a three-dimensional model of the texture, the color characteristic and the internal structure of the surface of the mechanical part through three-dimensional reconstruction reverse engineering design, so as to present the positions of the defects and the defect range of the surface and the internal structure of the mechanical part, and present a three-dimensional reconstruction system of the defects as a three-dimensional PDF document.
In order to achieve the purpose, the technical solution of the invention is as follows: a visual inspection method for defects of mechanical parts is characterized by comprising the following steps: the method comprises the following steps:
three-dimensional scanning and close range shooting are carried out on the surface of the mechanical part by utilizing a three-dimensional coordinate measuring instrument and a close range camera, and images are stored;
denoising noise points and reflecting regions influencing the image characteristics by adopting a median filtering method, and performing threshold automatic segmentation through an ostu algorithm to perform binarization processing on the image;
firstly, acquiring a pixel-level edge by adopting a Candy edge detection algorithm, then determining an edge sub-pixel of an edge position by adopting a two-dimensional space gray matrix by adopting a space matrix edge positioning method so as to refine the edge, and further refining the edge positioning precision to the inside of the pixel from rough to fine so as to enhance the surface smoothness and edge sharpening of the image;
then, analyzing the color of the surface of the mechanical part by adopting a clustering algorithm, analyzing the image texture characteristics of the analyzed mechanical part, matching the optimized image texture characteristics, facilitating the subsequent analysis of the continuity of the surface texture of the mechanical part, judging the continuity of the surface texture and the color characteristics of the mechanical part by adopting an identification technology, calculating the contact ratio matching of the edge of a non-smooth connection area of the surface of the mechanical part, analyzing the continuity of the surface of the mechanical part by a computer, and storing the surface image of the analyzed mechanical part;
performing three-dimensional tomography on mechanical parts by adopting a tomography thin-layer acquisition method, performing image enhancement on an obtained tomography image through gray level conversion, performing binarization processing on the image after image enhancement through a binarization algorithm, and calculating and storing the defect area, the defect mass center and the defect shape in the image;
restoring a three-dimensional model of textures, color characteristics and an internal structure of the surface of the mechanical part through three-dimensional reconstruction reverse engineering design, so as to present the positions and the defect ranges of the surface and the internal defects of the mechanical part;
and restoring the defect structure of the mechanical part and the position and the shape of the damaged part, and presenting a three-dimensional reconstruction system of the defect as a three-dimensional PDF document.
As a further improvement of the present invention, the three-dimensional scanning and close-range image capturing and image storage of the surface of the mechanical part by using the three-dimensional coordinate measuring apparatus and the close-range camera is specifically that when the surface of the mechanical part is scanned by using the three-dimensional laser scanner, the millimeter-scale and submillimeter-scale three-dimensional laser scanning systems are comprehensively used, the millimeter-scale three-dimensional scanner is used for overall control, the submillimeter-scale three-dimensional scanner is used for local data acquisition, and then close-range photogrammetry measurement is combined to realize the complete reservation of the surface information of the mechanical part.
As further limitation of three-dimensional scanning and close-range shooting, firstly, millimeter-scale and submillimeter-scale three-dimensional laser scanning instruments are used for obtaining dense point data of the whole mechanical part, and deformation and defect shapes of local structures on the surface of the mechanical part are obtained through multi-time phase point data comparison; and then close-range photogrammetry is adopted, an image of the mechanical part is obtained in a projection or texture expansion mode, and the slight change of texture color is detected by means of an image recognition technology.
As a further improvement of the present invention, the specific method adopted in the image recognition technology is as follows: firstly, correcting and preprocessing a close-range image to obtain a calibration image of the surface of the mechanical part, then extracting a spectral curve corresponding to a certain pixel point in the image as a spectral feature of the point, then carrying out PCA (principal component analysis) dimensionality reduction processing on the image to reduce the image to 3 dimensions, extracting the feature by using a feature extractor based on CNN (computer-based network), taking the extracted feature as a spatial feature, finally carrying out linear feature fusion on the spectral feature and the spatial feature to form a spectrum-space feature set, and classifying by using an SVM (support vector machine) as a classifier to obtain a result of color continuity between the surfaces of the mechanical part.
As a further improvement of the method, the optimization processing of the noise and reflection regions influencing the image characteristics mainly comprises point cloud splicing, noise reduction and redundancy removal, data segmentation, filtering and point cloud optimization processing.
As a further improvement of the present invention, the method for determining continuity of surface texture between surfaces of mechanical parts specifically adopts the following method:
firstly, counting the number of corner points: designing an arc length threshold value A as a comparison standard, and obtaining the value A by weighting the average arc length between the model value points; when the adjacent distance of the continuous points is continuously less than or equal to A, the area is regarded as having 1 corner point, and the type value points in the corner point area are marked;
then, removing redundant points in the texture point cloud: designing an included angle threshold B, and calculating adjacent three points Pti-1、Pt、Pti+1The included angle a between two vectors consisting of 1 ≦ i ≦ N-1iWherein N is a positive integer greater than 1, if aiIf the value is greater than the threshold value B, deleting the intermediate type value point Pt until all the type value points are judged;
finally, the starting and ending points Pt are connected0PtN-1Calculating the distance h from each point to the connection lineiAnd calculating a data point Pt with a corner markiAnd two adjacent data points Pti-1、Pti+1The resulting angle a of the two vectors formediDistance h from the middle data pointiRatio a ofi/hiHaving a minimum value min (a)i/hi) The middle data points are 1 known corner point, the corner point marks of the data points of adjacent regions are deleted, the complex texture section is split, the corner point counter is reduced by 1, the step is continuously carried out on the new texture section, and the counter is reduced by 1 when each corner point is judged until the counter is reduced to 0, and the judgment is stopped.
The invention aims to highlight the local characteristics in the tomographic image, clarify the original unclear internal defects, enlarge the difference between different object characteristics in the image, enhance the differentiation between different parts in the image, blur the characteristics of irrelevant parts, form contrast, enrich the quality and information content of partial images to be researched in the image, more quickly and effectively identify the reinforced parts, meet the requirements of image analysis and research, and enhance the interpretation and identification effects of internal defects.
The gray-scale transform image processing techniques are performed in the spatial domain, which can be represented by the following equation:
g (x, y) = T [ f (x, y) ], where f (x, y) is the input image, g (x, y) is the image after processing, T is an operator defined in the domain of the point (x, y) with respect to f, the operator can be applied to a single image or to a set of images, a neighborhood with respect to the point (x, y) in one image in the spatial domain, the neighborhood moving from one pixel to another in the image to generate an output image, the processing steps of which are as follows:
1) moving the neighborhood origin from one pixel to another, applying an operator T to the pixels in the neighborhood and producing an output at that location, the value of the output image g at these coordinates for any given location (x, y) being equal to the result of applying the operator T to the neighborhood of f having (x, y) as the origin;
2) the origin of the neighborhood is moved further to the next position and the process is repeated cyclically, starting from the upper left-hand position of the input image and then proceeding pixel by pixel in a horizontal scan, one line at a time, to produce the value of the next output image g.
The gray scale change in the present invention specifically uses inversion transformation, and can obtain an inverted image of an image with a gray scale range of [0, L-1], which is given by the following formula: s = L-1-r; reversing the grey scale of an image in this way can result in a negative of an equivalent photograph, this type being particularly useful for enhancing white or grey detail embedded in dark areas of an image, particularly when black areas predominate in size.
After image enhancement, in order to highlight the internal image defects of the mechanical parts obtained by tomography, the gray value of a pixel point on the image is set to be 0 or 255, that is, the whole image has an obvious black-and-white effect, so that the internal defects are more obviously highlighted.
On the basis of image enhancement, carrying out binarization processing on an image; the binarization processing in the invention specifically selects a maximum inter-class variance method.
In the binarization process, 256 brightness levels are processed by selecting a proper threshold value, so that an image reflecting local or overall characteristics of the image can be obtained; in image processing, the image is binarized, which plays a very important role in subsequent image analysis.
Firstly, the image binarization process can effectively prepare for further processing of the image, so that the image has only two states of 0 and 255, the data volume is reduced, and the target contour to be acquired can be highlighted; secondly, the premise of processing and analyzing the binary image is that the binary image can be obtained only by carrying out the binary processing on the gray level image, so that the subsequent analysis of the image is facilitated, the set property of the image is only related to the position of a point with a pixel value of 0 or 255, the multi-level value of the pixel is not related any more, the processing becomes simple, and the processing and compression amount of data is small.
The binarization processing of the image is to set the gray level of a point on the image to be 0 or 255, all pixels with the gray level larger than or equal to a threshold value are defined as a part to be highlighted, the gray level of the part is 255 for representation, the pixel points of the rest part are determined to be outside the region, the gray level is set to be 0, and the background or the region which is not interested is represented; in other words, the whole image exhibits obvious black and white effect, and in order to obtain an ideal binary image, a closed and connected boundary is generally adopted to define a non-overlapping region.
After image binarization processing, obtaining black holes in the mechanical parts, namely defects in the mechanical parts; the defect area, defect mass center and defect shape data of the internal defect of the mechanical part can be obtained by calculating the internal defect black area and then carrying out statistical analysis on the internal defect pixel area of the mechanical part.
The calculation of the defect area, the defect mass center and the defect shape in the image is specifically to obtain the internal defect of the tomography image in the mechanical part, circularly traverse the boundary of each communicated region, then calculate the area surrounded by the defect boundary, mark the defect mass center and solve the defect mass center coordinate; and then circularly processing each closed defect boundary, acquiring boundary coordinates, calculating the point number for describing the closed region boundary, and acquiring the shape of the defect.
As a further improvement of the invention, the fault thin layer collection method specifically adopts the following method: thin-layer scanning and storing are carried out on mechanical parts by adopting a tomography device, segmentation results are input, a proper segmentation method is selected according to image characteristics, and an isosurface is constructed by utilizing the segmentation results; then, the two-dimensional slice images are analyzed and processed, and a series of two-dimensional slice images obtained through scanning are subjected to tomographic interpolation to construct a three-dimensional data volume.
As a further improvement of the invention, the CT device comprises a scanning part, a computer system and an image display and storage system; the scanning part consists of an X-ray tube, a detector and a scanning frame; the computer system stores and operates the information data collected by scanning and constructs a three-dimensional model; the image display storage system displays the image processed and reconstructed by the computer on a display screen.
The three-dimensional reconstruction reverse engineering design deduces and obtains the internal structure of the mechanical part by reversely analyzing the structure of the mechanical part so as to evaluate the continuity of the internal structure of the mechanical part.
The three-dimensional reconstruction reverse engineering design is a product design technology reproduction process, namely, a project standard product is subjected to reverse analysis and research, so that design factors such as processing flow, organization structure, functional characteristics, technical specification and the like of the product are deduced and obtained, and similar products are manufactured.
The invention can make the operator see the defect part clearly by continuously scanning the defect parts inside and on the surface of the mechanical part and then processing and analyzing the acquired image to obtain the area, the defect mass center coordinate and the defect shape of the defect on the surface and inside; and then fitting the surface image of the mechanical part obtained by three-dimensional scanning and the internal image of the mechanical part obtained by fault scanning, and then restoring a three-dimensional model of the texture, the color characteristic and the internal structure of the surface of the mechanical part by three-dimensional reconstruction reverse engineering design, so that the positions and the defect ranges of the surface and the internal defects of the mechanical part are presented, and a three-dimensional reconstruction system of the defects is presented as a three-dimensional PDF document.
The invention has the beneficial effects that:
1. when the internal defect of the mechanical part is detected, the surface and the interior of the mechanical part are scanned by adopting a three-dimensional scanning technology and a tomography technology, and then a three-dimensional model of the structure of the surface and the interior of the mechanical part is restored by 3D reconstruction reverse engineering design, so that the internal defect position and the defect range are presented.
2. The method adopts millimeter-scale and submillimeter-scale three-dimensional laser scanning instruments to obtain dense point cloud data of the whole mechanical part, and obtains the deformation and defect shape of the local structure of the mechanical part through multi-temporal phase point cloud data comparison; and then, close-range photogrammetry is adopted, an orthophoto image of the mechanical part is obtained in a projection or texture expansion mode, and the slight change of texture color is detected by means of an image recognition technology, so that the matching of a non-smooth connection area on the surface of the subsequent mechanical part is greatly facilitated, and the accuracy of the computer for identifying the surface defects of the mechanical part is improved.
3. According to the method, a Candy edge detection algorithm is adopted to obtain a pixel-level edge, then a space moment edge positioning method is adopted to determine edge sub-pixels of an edge position by utilizing a two-dimensional space gray matrix so as to thin the edge, and then the edge positioning precision is thinned to the inside of the pixel from rough to fine, so that the surface smoothness and the edge sharpening of the image can be effectively enhanced.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention; in the drawings:
fig. 1 is a schematic view of a method for visually inspecting a defect of a mechanical component according to the present invention.
Fig. 2 is a schematic flow chart of a visual inspection method for defects of mechanical parts according to the present invention.
Detailed Description
The present invention and its embodiments are described in further detail below with reference to the accompanying drawings.
A method for visually inspecting a mechanical component for defects, the method comprising:
three-dimensional scanning and close range shooting are carried out on the surface of the mechanical part by utilizing a three-dimensional coordinate measuring instrument and a close range camera, and images are stored;
denoising noise points and reflecting regions influencing the image characteristics by adopting a median filtering method, and performing threshold automatic segmentation through an ostu algorithm to perform binarization processing on the image;
firstly, acquiring a pixel-level edge by adopting a Candy edge detection algorithm, then determining an edge sub-pixel of an edge position by adopting a two-dimensional space gray matrix by adopting a space matrix edge positioning method so as to refine the edge, and further refining the edge positioning precision to the inside of the pixel from rough to fine so as to enhance the surface smoothness and edge sharpening of the image;
then, analyzing the color of the surface of the mechanical part by adopting a clustering algorithm, analyzing the image texture characteristics of the analyzed mechanical part, matching the optimized image texture characteristics, facilitating the subsequent analysis of the continuity of the surface texture of the mechanical part, judging the continuity of the surface texture and the color characteristics of the mechanical part by adopting an identification technology, calculating the contact ratio matching of the edge of a non-smooth connection area of the surface of the mechanical part, analyzing the continuity of the surface of the mechanical part by a computer, and storing the surface image of the analyzed mechanical part;
performing three-dimensional tomography on mechanical parts by adopting a tomography thin-layer acquisition method, performing image enhancement on an obtained tomography image through gray level conversion, performing binarization processing on the image after image enhancement through a binarization algorithm, and calculating and storing the defect area, the defect mass center and the defect shape in the image;
restoring a three-dimensional model of textures, color characteristics and an internal structure of the surface of the mechanical part through three-dimensional reconstruction reverse engineering design, so as to present the positions and the defect ranges of the surface and the internal defects of the mechanical part;
and restoring the defect structure of the mechanical part and the position and the shape of the damaged part, and presenting a three-dimensional reconstruction system of the defect as a three-dimensional PDF document.
The method for three-dimensional scanning and close range shooting of the surface of the mechanical part and storing the image by using the three-dimensional coordinate measuring instrument and the close range camera specifically comprises the following steps: when the three-dimensional laser scanner is adopted to scan the surface of the mechanical part, the millimeter-scale and submillimeter-scale three-dimensional laser scanning system is comprehensively utilized, the millimeter-scale three-dimensional scanner is used for carrying out overall control, the submillimeter-scale three-dimensional scanner is used for carrying out local data acquisition, and then close-range photogrammetry is combined to realize the complete reservation of the surface information of the mechanical part.
The steps are as follows: firstly, obtaining dense point data of the whole mechanical part by using a millimeter-scale and submillimeter-scale three-dimensional laser scanning instrument, and obtaining the deformation and defect shape of the local structure of the surface of the mechanical part by comparing the data of the dense point data with multi-time phase data; and then close-range photogrammetry is adopted, an image of the mechanical part is obtained in a projection or texture expansion mode, and the slight change of texture color is detected by means of an image recognition technology.
The specific method adopted in the image recognition technology comprises the following steps: firstly, correcting and preprocessing a close-range image to obtain a calibration image of the surface of the mechanical part, then extracting a spectral curve corresponding to a certain pixel point in the image as a spectral feature of the point, then carrying out PCA (principal component analysis) dimensionality reduction processing on the image to reduce the image to 3 dimensions, extracting the feature by using a feature extractor based on CNN (computer-based network), taking the extracted feature as a spatial feature, finally carrying out linear feature fusion on the spectral feature and the spatial feature to form a spectrum-space feature set, and classifying by using an SVM (support vector machine) as a classifier to obtain a result of color continuity between the surfaces of the mechanical part.
The optimization processing of the noise and reflection regions influencing the image characteristics mainly comprises point cloud splicing, noise reduction and redundancy removal, data segmentation, filtering and point cloud optimization processing; the binarization processing specifically selects a maximum inter-class variance method.
The method specifically comprises the following steps of calculating and storing the defect area, the defect mass center and the defect shape inside the image: acquiring internal defects of a tomography image in the mechanical part, circularly traversing the boundary of each communicated region, then calculating the area enclosed by the defect boundary, marking the defect centroid and solving the coordinates of the defect centroid; and then circularly processing each closed defect boundary, acquiring boundary coordinates, calculating the point number describing the boundary of the closed region, acquiring the shape of the defect, and finally storing the defect area, the defect mass center and the defect shape data in the acquired image.
The fault thin-layer acquisition method comprises the steps of CT equipment data acquisition and three-dimensional reconstruction, wherein the CT equipment data acquisition is to perform thin-layer scanning storage on mechanical parts and then restore a three-dimensional structure model through three-dimensional reconstruction reverse engineering design.
The fault thin layer collection method specifically adopts the following method: thin-layer scanning and storing mechanical parts by adopting a tomography device, inputting a segmentation result, segmenting an image according to image characteristics, and constructing an isosurface by utilizing the segmentation result; then, the two-dimensional slice images are analyzed and processed, and a series of two-dimensional slice images obtained through scanning are subjected to tomographic interpolation to construct an internal three-dimensional data structure.
The tomography equipment comprises a scanning part, a computer system and an image display and storage system; the scanning part consists of an X-ray tube, a detector and a scanning frame; the computer system stores and operates the information data collected by scanning; the image display storage system displays the image processed and reconstructed by the computer on a display screen.
The invention can make the operator see the defect part clearly by continuously scanning the defect parts inside and on the surface of the mechanical part and then processing and analyzing the acquired image to obtain the area, the defect mass center coordinate and the defect shape of the defect on the surface and inside; and then fitting the surface image of the mechanical part obtained by three-dimensional scanning and the internal image of the mechanical part obtained by fault scanning, and then restoring a three-dimensional model of the texture, the color characteristic and the internal structure of the surface of the mechanical part by three-dimensional reconstruction reverse engineering design, so that the positions and the defect ranges of the surface and the internal defects of the mechanical part are presented, and a three-dimensional reconstruction system of the defects is presented as a three-dimensional PDF document.
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 should be equivalent or changed within the scope of the present invention.
Claims (8)
1. A visual inspection method for defects of mechanical parts is characterized by comprising the following steps: the method comprises the following steps:
three-dimensional scanning and close range shooting are carried out on the surface of the mechanical part by utilizing a three-dimensional laser scanner and a close range camera, and images are stored;
denoising noise points and reflecting regions influencing the image characteristics by adopting a median filtering method, and performing threshold automatic segmentation through an ostu algorithm to perform binarization processing on the image;
firstly, acquiring a pixel-level edge of the image subjected to binarization processing by adopting a Candy edge detection algorithm, then determining an edge sub-pixel of an edge position by utilizing a two-dimensional space gray matrix by adopting a space matrix edge positioning method so as to refine the edge, and further refining the edge positioning precision to the inside of the pixel from coarse to fine so as to enhance the surface smoothness and edge sharpening of the image;
then analyzing the color of the surface of the mechanical part by adopting a clustering algorithm, optimizing the image texture characteristics of the analyzed mechanical part, matching the optimized image texture characteristics, facilitating the subsequent analysis of the continuity of the surface texture of the mechanical part, judging the continuity of the surface texture and the color characteristics of the mechanical part by adopting an identification technology, calculating the contact ratio matching of the edge of a non-smooth connection area of the surface of the mechanical part, analyzing the continuity of the surface of the mechanical part by a computer, and storing the surface image of the analyzed mechanical part; the specific method adopted in the identification technology is as follows: firstly, correcting and preprocessing a close-range image to obtain a calibration image of the surface of a mechanical part, then extracting a spectral curve corresponding to a certain pixel point in the image as a spectral feature of the point, then carrying out PCA (principal component analysis) dimensionality reduction processing on the image to reduce the image to 3 dimensions, extracting the feature by using a feature extractor based on CNN (computer network), taking the extracted feature as a spatial feature, finally carrying out linear feature fusion on the spectral feature and the spatial feature to form a spectrum-space feature set, and classifying by using an SVM (support vector machine) as a classifier to obtain a result of color continuity between the surfaces of the mechanical part;
performing three-dimensional tomography scanning on mechanical parts by adopting a tomography thin-layer acquisition method, performing image enhancement on an obtained tomography image through gray level conversion processing, performing binarization processing on the image after image enhancement through a binarization algorithm, and calculating and storing the defect area, the defect mass center and the defect shape in the mechanical parts; the calculation and storage of the defect area, the defect mass center and the defect shape in the mechanical part are specifically as follows: acquiring internal defects of a tomography image in the mechanical part, circularly traversing the boundary of each communicated region, then calculating the area enclosed by the defect boundary, marking the defect centroid and solving the coordinates of the defect centroid; then circularly processing each closed defect boundary, acquiring boundary coordinates, calculating the point number describing the boundary of the closed region, acquiring the shape of the defect, and finally storing the defect area, the defect mass center and the defect shape data in the obtained image;
restoring a three-dimensional model of textures, color characteristics and an internal structure of the surface of the mechanical part through three-dimensional reconstruction reverse engineering design, so as to present the positions and the defect ranges of the surface and the internal defects of the mechanical part;
and restoring the defect structure of the mechanical part and the position and the shape of the damaged part, and presenting a three-dimensional reconstruction system of the defect as a three-dimensional PDF document.
2. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 1, wherein: the three-dimensional scanning and close range shooting and image storage of the surface of the mechanical part by using the three-dimensional laser scanner and the close range camera are specifically as follows: when adopting three-dimensional laser scanner scanning machine spare part surface, the three-dimensional laser scanner of comprehensive utilization millimeter level and submillimeter level carries out overall control with the three-dimensional laser scanner of millimeter level, carries out local data acquisition with the three-dimensional laser scanner of submillimeter level, then combines close-range photogrammetry, realizes that the whole of machine spare part surface information is left and is got.
3. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 2, wherein: firstly, obtaining dense point data of the whole mechanical part by using a millimeter-scale and submillimeter-scale three-dimensional laser scanning instrument, and obtaining the deformation and defect shape of the local structure of the surface of the mechanical part by comparing the data of the dense point data with multi-time phase data; and then close-range photogrammetry is adopted, an image of the mechanical part is obtained in a projection or texture expansion mode, and the slight change of texture color is detected by means of an image recognition technology.
4. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 1, wherein: the processing of noise points and light reflecting areas influencing the image characteristics comprises point cloud splicing, noise reduction and redundancy removal, data segmentation, filtering and point cloud optimization.
5. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 4, wherein: and automatically segmenting a threshold value through an ostu algorithm, and specifically selecting a maximum inter-class variance method for the binarization processing of the image.
6. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 1, wherein: the fault thin-layer acquisition method comprises the steps of acquiring data and three-dimensional reconstruction by CT equipment, wherein the data acquired by the CT equipment is obtained by carrying out thin-layer scanning and storage on mechanical parts, and then restoring an internal structure three-dimensional model by three-dimensional reconstruction reverse engineering design.
7. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 6, wherein: the fault thin layer collection method specifically adopts the following method: thin-layer scanning and storing are carried out on mechanical parts by adopting a tomography device, segmentation results are input, a proper segmentation method is selected according to image characteristics, and an isosurface is constructed by utilizing the segmentation results; then, the two-dimensional slice images are analyzed and processed, and a series of two-dimensional slice images obtained through scanning are subjected to tomographic interpolation to construct a three-dimensional data volume.
8. The visual inspection method of the defects of the mechanical parts and components as claimed in claim 7, wherein: the tomography equipment comprises a scanning part, a computer system and an image display and storage system; the scanning part consists of an X-ray tube, a detector and a scanning frame; the computer system stores and operates the information data collected by scanning; the image display storage system displays the image processed and reconstructed by the computer on a display screen.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811334772.3A CN109580630B (en) | 2018-11-10 | 2018-11-10 | Visual inspection method for defects of mechanical parts |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811334772.3A CN109580630B (en) | 2018-11-10 | 2018-11-10 | Visual inspection method for defects of mechanical parts |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109580630A CN109580630A (en) | 2019-04-05 |
| CN109580630B true CN109580630B (en) | 2022-02-18 |
Family
ID=65922026
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811334772.3A Active CN109580630B (en) | 2018-11-10 | 2018-11-10 | Visual inspection method for defects of mechanical parts |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109580630B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110052569A (en) * | 2019-05-22 | 2019-07-26 | 安庆市吉安汽车零件锻轧有限公司 | A kind of intelligence gear shaft forging rolling technique |
| CN110197479B (en) * | 2019-05-23 | 2021-03-23 | 中国科学院力学研究所 | Surface modification quality detection method |
| CN110400286A (en) * | 2019-06-05 | 2019-11-01 | 山东科技大学 | The detection localization method of metal needle in a kind of X ray CT image |
| CN110738180B (en) * | 2019-10-21 | 2022-05-17 | 易思维（杭州）科技有限公司 | Method for evaluating signal accuracy and system precision in detection process |
| CN110806736B (en) * | 2019-11-19 | 2021-10-15 | 北京工业大学 | Method for detecting quality information of forge pieces of die forging forming intelligent manufacturing production line |
| CN112484664A (en) * | 2020-11-26 | 2021-03-12 | 江苏国和智能科技有限公司 | Defect identification device and method based on laser three-dimensional scanning |
| CN112304960B (en) * | 2020-12-30 | 2021-08-10 | 中国人民解放军国防科技大学 | High-resolution image object surface defect detection method based on deep learning |
| CN112967280A (en) * | 2021-04-07 | 2021-06-15 | 苏州工业职业技术学院 | Three-dimensional visual defect detection method based on CT |
| CN114034704A (en) * | 2021-10-18 | 2022-02-11 | 武汉科技大学 | Waste roller three-dimensional detection device and method based on multi-source data fusion |
| CN114092506B (en) * | 2022-01-20 | 2022-04-15 | 北京微链道爱科技有限公司 | Scanning device with 3D (three-dimensional) characteristics and scanning method |
| CN116595214A (en) * | 2023-07-14 | 2023-08-15 | 中国航发北京航空材料研究院 | Component non-contact automatic shape correction system based on image processing |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103239256A (en) * | 2013-05-29 | 2013-08-14 | 中国人民解放军第三军医大学第一附属医院 | Virtual X-ray imaging method and virtual X-ray imaging system for human body bone joint |
| CN104266587A (en) * | 2014-09-22 | 2015-01-07 | 电子科技大学 | Three-dimensional measurement system and method for obtaining actual 3D texture point cloud data |
| CN105912874A (en) * | 2016-04-29 | 2016-08-31 | 青岛大学附属医院 | Liver three-dimensional database system constructed on the basis of DICOM (Digital Imaging and Communications in Medicine) medical image |
| CN106204733A (en) * | 2016-07-22 | 2016-12-07 | 青岛大学附属医院 | Liver and the three-dimensional constructing system of kidney CT image associating |
-
2018
- 2018-11-10 CN CN201811334772.3A patent/CN109580630B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103239256A (en) * | 2013-05-29 | 2013-08-14 | 中国人民解放军第三军医大学第一附属医院 | Virtual X-ray imaging method and virtual X-ray imaging system for human body bone joint |
| CN104266587A (en) * | 2014-09-22 | 2015-01-07 | 电子科技大学 | Three-dimensional measurement system and method for obtaining actual 3D texture point cloud data |
| CN105912874A (en) * | 2016-04-29 | 2016-08-31 | 青岛大学附属医院 | Liver three-dimensional database system constructed on the basis of DICOM (Digital Imaging and Communications in Medicine) medical image |
| CN106204733A (en) * | 2016-07-22 | 2016-12-07 | 青岛大学附属医院 | Liver and the three-dimensional constructing system of kidney CT image associating |
Non-Patent Citations (1)
| Title |
|---|
| 基于三维CT图像的涡轮叶片缺陷检测技术研究;金炎芳;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20070615(第06期);摘要和图3-3、3-12以及图5-4 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109580630A (en) | 2019-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109580630B (en) | Visual inspection method for defects of mechanical parts | |
| CN110866924B (en) | Line structured light center line extraction method and storage medium | |
| CN109655019B (en) | Cargo volume measurement method based on deep learning and three-dimensional reconstruction | |
| CN106875381B (en) | Mobile phone shell defect detection method based on deep learning | |
| CN111462120B (en) | Defect detection method, device, medium and equipment based on semantic segmentation model | |
| RU2680765C1 (en) | Automated determination and cutting of non-singular contour of a picture on an image | |
| CN107228860B (en) | Gear defect detection method based on image rotation period characteristics | |
| CN113688807B (en) | Self-adaptive defect detection method, device, recognition system and storage medium | |
| CN112651968B (en) | Wood board deformation and pit detection method based on depth information | |
| CN111833306A (en) | Defect detection method and model training method for defect detection | |
| CN109540925B (en) | Complex ceramic tile surface defect detection method based on difference method and local variance measurement operator | |
| CN110596120A (en) | Glass boundary defect detection method, device, terminal and storage medium | |
| CN109559324A (en) | A kind of objective contour detection method in linear array images | |
| CN113240626A (en) | Neural network-based method for detecting and classifying concave-convex flaws of glass cover plate | |
| CN111354047B (en) | Computer vision-based camera module positioning method and system | |
| CN106327451A (en) | Image restorative method of ancient animal fossils | |
| CN109410139A (en) | Inside a kind of historical relic and surface disease digital assay appraisal procedure | |
| CN111986170A (en) | Defect detection algorithm based on Mask R-CNN (deep neural network) | |
| CN114550021A (en) | Surface defect detection method and device based on feature fusion | |
| CN112489042A (en) | Metal product printing defect and surface damage detection method based on super-resolution reconstruction | |
| CN113609984A (en) | Pointer instrument reading identification method and device and electronic equipment | |
| CN114581805A (en) | Coating roller surface defect detection method adopting 3D line laser profile technology | |
| Böhm et al. | Model refinement by integrated processing of laser scanning and photogrammetry | |
| CN116539619A (en) | Product defect detection method, system, device and storage medium | |
| CN115393290A (en) | Edge defect detection method, device and equipment |
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 |