CN117291937B - Automatic plastering effect visual detection system based on image feature analysis - Google Patents
Automatic plastering effect visual detection system based on image feature analysis Download PDFInfo
- Publication number
- CN117291937B CN117291937B CN202311587747.7A CN202311587747A CN117291937B CN 117291937 B CN117291937 B CN 117291937B CN 202311587747 A CN202311587747 A CN 202311587747A CN 117291937 B CN117291937 B CN 117291937B
- Authority
- CN
- China
- Prior art keywords
- region
- area
- segmentation threshold
- immersed
- gray value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 28
- 230000000694 effects Effects 0.000 title claims abstract description 14
- 238000001514 detection method Methods 0.000 title claims abstract description 11
- 230000000007 visual effect Effects 0.000 title claims abstract description 9
- 230000011218 segmentation Effects 0.000 claims abstract description 201
- 230000007547 defect Effects 0.000 claims abstract description 79
- 238000000034 method Methods 0.000 claims abstract description 33
- 230000008859 change Effects 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 14
- 238000005286 illumination Methods 0.000 claims description 12
- 238000012163 sequencing technique Methods 0.000 claims description 8
- 241001632427 Radiola Species 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 238000011179 visual inspection Methods 0.000 claims 2
- 238000010276 construction Methods 0.000 claims 1
- 230000002950 deficient Effects 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 3
- 239000006185 dispersion Substances 0.000 description 4
- 238000007654 immersion Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000004570 mortar (masonry) Substances 0.000 description 2
- 238000005507 spraying Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- 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
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- 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
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of image processing, in particular to an automatic plastering effect visual detection system based on image feature analysis, which comprises the following steps: acquiring a wallboard surface image; acquiring a neighborhood region of each pixel point, and further acquiring the possibility that each pixel point is a seed point to obtain the seed point; obtaining a dark region segmentation threshold range and a bright region segmentation threshold range according to the seed points; acquiring an immersed region corresponding to each gray value in a dark region segmentation threshold range, and further acquiring the possibility that each gray value is the dark region segmentation threshold to obtain a dark region segmentation threshold; acquiring an immersed region and an unsubmerged region corresponding to each gray value in a bright region segmentation threshold range, and further acquiring the possibility that each gray value is the bright region segmentation threshold to acquire a bright region segmentation threshold; acquiring a plurality of suspected areas; the method and the device can divide the complete bubble defect area.
Description
Technical Field
The invention relates to the field of image processing, in particular to an automatic plastering effect visual detection system based on image feature analysis.
Background
In the production and processing process of FK wallboard, plastering treatment is important for the nailed wallboard, the production quality of the wallboard can be guaranteed by scientific plastering, the performances of dampproofing, weather resistance, heat insulation and the like of the wallboard are enhanced, the FK wallboard plastering equipment is automatic plastering equipment, mortar can be automatically sprayed on the nailed wallboard, gridding cloth is paved and troweling treatment is carried out on the nailed wallboard, the plastering efficiency is greatly improved, however, when bubbles are not fully removed in the mortar spraying process, the automatic plastering equipment can possibly cause bubbles to form on the surface of a wall body, so that bubble defects are common wallboard plastering defects, and detection is needed.
When the watershed algorithm is used for detecting the bubble defect areas of the wallboard, the watershed algorithm is sensitive to the selection of seed points, and the bubble defect areas in the surface image of the wallboard are difficult to distinguish by a direct means, so that the inaccurate segmentation result is easily caused by unreasonable selection of the seed points, and the gray values of the bubble defect areas are different due to the fact that the shapes of the bubble defect areas are convex under the influence of illumination influence, namely, the bubble defect areas have bright areas with gray values larger than those of the normal areas and dark areas with gray values smaller than those of the normal areas, and the complete bubble defect areas cannot be segmented only by a single threshold value of the watershed segmentation algorithm.
Disclosure of Invention
In order to solve the above problems, the present invention provides an automatic plastering effect visual detection system based on image feature analysis, the system comprising:
the wallboard surface image acquisition module is used for acquiring wallboard surface images;
a seed point acquisition module; the method comprises the steps of acquiring a neighborhood region of each pixel point in a wallboard surface image, and acquiring the possibility that each pixel point is a seed point according to the gray value variance of all pixel points in the neighborhood region of each pixel point and the difference between the gray value of each pixel point and the gray average value of all pixel points in the neighborhood region of each pixel point; acquiring a seed point according to the possibility that each pixel point is the seed point;
the segmentation threshold acquisition module is used for acquiring a dark region segmentation threshold range and a bright region segmentation threshold range according to the seed points; acquiring an immersed region corresponding to each gray value in a dark region segmentation threshold range; acquiring the possibility that each gray value in the dark region segmentation threshold range is a dark region segmentation threshold according to the gray distribution discrete degree of the immersed region corresponding to each gray value in the dark region segmentation threshold range, the area change of the immersed region, the gradient mean value of the pixel points at the edge of the immersed region and the shape similarity among the immersed regions, and acquiring the dark region segmentation threshold according to the possibility that each gray value in the dark region segmentation threshold range is a dark region segmentation threshold; acquiring an immersed region and an unsubmerged region corresponding to each gray value in a bright region segmentation threshold range; acquiring the possibility that each gray value in the bright region segmentation threshold range is a bright region segmentation threshold according to the gray distribution discrete degree of the non-immersed region corresponding to each gray value in the bright region segmentation threshold range, the area change of the immersed region, the gradient mean value of the edge pixel points of the non-immersed region and the shape similarity between the non-immersed regions, and acquiring the bright region segmentation threshold according to the possibility that each gray value in the bright region segmentation threshold range is a bright region segmentation threshold;
The bubble defect region acquisition module is used for acquiring a plurality of suspected regions according to the bright region segmentation threshold and the dark region segmentation threshold; acquiring the possibility that each suspected region is a bubble defect region according to the approximate circle degree of the suspected region, the coincidence degree of the suspected region and the gradient direction consistency of the suspected region; and acquiring bubble defect areas according to the possibility that each suspected area is the bubble defect area.
Preferably, the step of obtaining the probability that each pixel point is a seed point according to the variance of the gray value of all the pixel points in the neighborhood region of each pixel point and the difference between the gray value of each pixel point and the gray average value of all the pixel points in the neighborhood region of each pixel point, wherein the neighborhood region is used for obtaining each pixel point in the wallboard surface image comprises the following steps:
presettingConstructing by taking each pixel point in the wallboard surface image as a centerA large area is used as a neighborhood area of each pixel point;
in the method, in the process of the invention,represents the firstThe likelihood that a pixel point is a seed point;represents the firstGray value variances of all pixel points in a neighborhood region of each pixel point;represents the firstGray average value and first pixel point of all pixel points in neighborhood region of each pixel point Difference of gray values of the individual pixel points; and acquiring the possibility that each pixel point is a seed point.
Preferably, the step of obtaining the seed point according to the probability that each pixel point is a seed point includes the following steps:
presettingSequencing all pixel points from big to small in probability, and sequentially obtainingThe pixel points are used as candidate seed points, when other candidate seed points exist in the eight neighborhood of any one candidate seed point, the candidate seed with the smallest gray value in the eight neighborhood of the candidate seed point is selectedThe seed points are used as one seed point, all seed points are obtained, and when no other alternative seed points exist in the eight adjacent positions of any alternative seed point, the alternative seed point is used as one seed point, and all seed points are obtained.
Preferably, the step of obtaining the dark region segmentation threshold range and the bright region segmentation threshold range according to the seed points includes the steps of:
sequencing all gray values between the minimum gray value in all the seed points and the gray value with the largest number of pixel points in the wallboard surface image according to the sequence from small to large, and then using the sequenced gray values as a dark area segmentation threshold range; and sequencing all gray values between the gray value with the largest number of pixels in the wallboard surface image and the maximum gray value in the wallboard surface image in order from small to large, and then taking the sequenced gray values as a bright area segmentation threshold range.
Preferably, the step of obtaining the immersed area corresponding to each gray value in the dark area segmentation threshold range includes the steps of:
and marking any gray value in the dark region segmentation threshold range as a current gray value, carrying out connected region analysis on pixel points smaller than the current gray value in the wallboard surface image, and marking the obtained connected region as an immersed region corresponding to the current gray value.
Preferably, the step of obtaining the dark region segmentation threshold according to the gray level distribution dispersion degree of the immersed region corresponding to each gray level value in the dark region segmentation threshold range, the area change of the immersed region, the gradient mean value of the edge pixels of the immersed region and the shape similarity between the immersed regions, and the probability that each gray level value in the dark region segmentation threshold range is the dark region segmentation threshold, includes the following steps:
acquiring shape similarity between immersed areas corresponding to the current gray values by using a shape context algorithm;
in the method, in the process of the invention,representing the possibility that the current gray value is a dark region segmentation threshold;representing the difference value between the area of the immersed area corresponding to the previous gray value of the current gray value in the dark area segmentation threshold range and the area of the immersed area corresponding to the current gray value; Representing the area difference value of the immersed area corresponding to the next gray value in the dark area segmentation threshold range of the current gray value and the immersed area corresponding to the current gray value;representing the area change of the immersed area corresponding to the current gray value;representing the number of pixel points in the immersed area corresponding to the current gray value;representing the first submerged area corresponding to the current gray valueGray values of the individual pixels;representing the average value of all pixel points in the immersed area corresponding to the current gray value;representing the gray level distribution discrete degree of the immersed area corresponding to the current gray level value;representing the shape similarity between the immersed areas corresponding to the current gray values;representing the number of edge pixel points of the immersed area corresponding to the current gray value;representing the first immersion area corresponding to the current gray valueGradient values of the edge pixels;representing the gradient average value of the pixel points at the edge of the immersed area corresponding to the current gray value;representing a normalization function;
and acquiring the possibility that each gray value in the dark region segmentation threshold range is the dark region segmentation threshold, and taking the gray value with the highest possibility in the dark region segmentation threshold range as the dark region segmentation threshold.
Preferably, the step of obtaining the immersed region and the non-immersed region corresponding to each gray value in the bright region segmentation threshold range includes the steps of:
Marking any gray value in the bright region segmentation threshold range as a current gray value, carrying out connected region analysis on pixel points larger than the current gray value in the wallboard surface image, and marking the obtained connected region as a non-immersed region corresponding to the current gray value; and carrying out connected domain analysis on the pixel points smaller than the current gray value in the wallboard surface image, and marking the obtained connected region as an immersed region corresponding to the current gray value.
Preferably, the step of obtaining the bright region segmentation threshold according to the gray level distribution discrete degree of the non-immersed region corresponding to each gray level value in the bright region segmentation threshold range, the area change of the immersed region, the gradient mean value of the edge pixels of the non-immersed region and the shape similarity between the non-immersed regions, and the possibility that each gray level value in the bright region segmentation threshold range is the bright region segmentation threshold, includes the steps of:
acquiring the shape similarity between the non-immersed areas corresponding to the gray values by using a shape context algorithm;
in the method, in the process of the invention,representing the possibility that the gray value is a bright area segmentation threshold value;representing the difference value between the area of the immersed area corresponding to the previous gray value of the current gray value in the bright area segmentation threshold range and the area of the immersed area corresponding to the current gray value; Representing the area difference value of the immersed area corresponding to the last gray value of the gray value in the bright area segmentation threshold range and the immersed area corresponding to the gray value;representing the area change of the immersed area corresponding to the gray value;representing the number of pixels in the non-immersed area corresponding to the gray value;represents the first non-immersed area corresponding to the gray valueGray values of the individual pixels;representing the average value of all pixel points in the non-immersed area corresponding to the gray value;representing the gray value pairThe degree of gray level distribution dispersion of the corresponding non-immersed region;representing the shape similarity between the non-immersed areas corresponding to the gray values;representing the number of edge pixel points of the non-immersed area corresponding to the gray value;represents the first non-immersed area corresponding to the gray valueGradient values of the edge pixels;representing the gradient average value of the pixel points at the edge of the non-immersed area corresponding to the gray value;representing a normalization function;
and acquiring the possibility that each gray value in the bright region segmentation threshold range is the bright region segmentation threshold, and taking the gray value with the highest possibility in the bright region segmentation threshold range as the bright region segmentation threshold.
Preferably, the obtaining a plurality of suspected areas according to the bright area segmentation threshold and the dark area segmentation threshold includes the steps of:
and carrying out connected domain analysis on pixel points smaller than a dark region segmentation threshold in the wallboard surface image to obtain each connected region, carrying out connected domain analysis on pixel points larger than a bright region segmentation threshold in the wallboard surface image to obtain each connected region, and merging the connected regions with overlapped pixels to obtain a plurality of suspected regions.
Preferably, the possibility that each suspected region is a bubble defect region is obtained according to the degree of approximate circle of the suspected region, the coincidence degree of the suspected region and the gradient direction consistency of the suspected region; acquiring the bubble defect areas according to the possibility that each suspected area is the bubble defect area, wherein the method comprises the following steps:
acquiring the mode of the gradient directions of all pixel points in the wallboard surface image, marking the mode as a first gradient direction, and taking the opposite vector of the first gradient direction as the illumination direction angle of the wallboard surface;
any suspected area is marked as a current suspected area;
in the method, in the process of the invention,representing the possibility that the current suspected region is a bubble defect region;representing the number of overlapped pixel points of all the connected areas in the current suspected area; Representing the number of edge pixel points of all the connected areas in the current suspected area;representing the coincidence degree of the current suspected region;representing the number of overlapped pixel points of each communication area in the current suspected area;representing the first connected region of the current suspected regionThe distances between the overlapped pixel points and the centroid of the current suspected region;representing the average value of the distances between all overlapped pixel points of each connected region in the current suspected region and the mass center of the current suspected region;representing the degree to which the current suspected region is approximately circular;representing the gradient direction consistency of the current suspected region;representing the first connected region of the current suspected regionGradient direction angles of the overlapped pixel points;an illumination direction angle representing an image of a wallboard surface;representing absolute value symbols; acquiring the possibility that each suspected region is a bubble defect region;
presetting a likelihood thresholdWhen the probability that any suspected region is a bubble defect region is greater than or equal to the probability thresholdAnd when the suspected area is a bubble defect area, obtaining all bubble defect areas.
The invention has the following beneficial effects: according to the invention, through the gray level distribution characteristics of the dark areas in the bubble defect areas of the wallboard surface images, the possibility that each pixel point is a seed point is obtained, and then the seed point is obtained according to the possibility that each pixel point is a seed point; the seed point is positioned at the position of the dark area in the bubble defect area, so that the accuracy of subsequent segmentation is improved; then obtaining a dark area segmentation threshold range and a bright area segmentation threshold range according to the seed points; acquiring an immersed region corresponding to each gray value in a dark region segmentation threshold range; acquiring the possibility that each gray value in the dark region segmentation threshold range is a dark region segmentation threshold according to the gray distribution discrete degree of the immersed region corresponding to each gray value in the dark region segmentation threshold range, the area change of the immersed region, the gradient average value of the pixel points at the edge of the immersed region and the shape similarity between the immersed regions, and obtaining the dark region segmentation threshold; acquiring an immersed region and an unsubmerged region corresponding to each gray value in a bright region segmentation threshold range; according to the gray level distribution discrete degree of the non-immersed area corresponding to each gray level value in the bright area segmentation threshold range, the area change of the immersed area, the gradient average value of the edge pixel points of the non-immersed area and the shape similarity among the non-immersed areas, the possibility that each gray level value in the bright area segmentation threshold range is a dark area segmentation threshold is obtained, the bright area segmentation threshold is obtained, and the problem that a single threshold cannot segment a complete bubble defect area is solved; finally, a plurality of suspected areas are obtained according to the bright area segmentation threshold and the dark area segmentation threshold; acquiring the possibility that each suspected region is a bubble defect region according to the approximate circle degree of the suspected region, the coincidence degree of the suspected region and the gradient direction consistency of the suspected region; and acquiring the bubble defect areas according to the possibility that each suspected area is the bubble defect area, so that the acquired bubble defect areas are more complete and accurate.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a system block diagram of an automatic plastering effect visual detection system based on image feature analysis according to an embodiment of the invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of the specific implementation, structure, characteristics and effects of the automatic plastering effect vision detection system based on image characteristic analysis according to the invention with reference to the attached drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the automatic plastering effect visual detection system based on image feature analysis provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, an automatic plastering effect visual detection system based on image feature analysis according to an embodiment of the invention is shown, and comprises the following modules:
wallboard surface image acquisition module 101 acquires wallboard surface images.
It should be noted that, the purpose of the present invention is to detect the bubble defect area in the wallboard surface image, so that the wall surface image needs to be acquired first, therefore, in the embodiment of the present invention, the camera is erected on the plastering device to acquire the plastered wallboard image, in order to facilitate the subsequent analysis, the plastered wallboard image is subjected to the graying treatment to obtain the gray image, and in addition, since the watershed segmentation algorithm is sensitive to noise, the segmentation result is more accurate in order to reduce the segmentation error, in the embodiment of the present invention, the gaussian filter is used to perform the denoising treatment on the gray image, and the denoised gray image is recorded as the wallboard surface image. It should be noted that, the gaussian filter is a known technique for denoising an image, and in the embodiment of the present invention, too much description is not given.
Thus, wallboard surface images were acquired.
The seed point acquisition module 102 acquires seed points in the wallboard surface image.
It should be noted that, the watershed segmentation algorithm regards the wallboard surface image as a geodesic topological feature, the gray value of each pixel point in the wallboard surface image represents the altitude of each pixel point, and the watershed is formed by simulating water immersion or precipitation to complete segmentation of different areas in the wallboard surface image, but since the watershed algorithm is sensitive to the selection of seed points, and the bubble defect areas in the wallboard surface image are difficult to distinguish by direct means, the segmentation result is inaccurate easily caused by unreasonable selection of the seed points, so that the position of the seed points needs to be determined first, in order to obtain a complete segmentation result, we need to set at least one seed point in each bubble defect area, and the seed points should be set in the dark areas in the bubble defect areas because the gray value of the dark areas in the bubble defect areas is low.
The automatic plastering device is uniform in spraying the wallboard, so that the surface illumination intensity of the automatic plastering device is uniform, the pixel gray value distribution of a normal area in a wallboard surface image is centralized, the pixel gray value distribution of a bubble defect area is discrete due to the fact that the internal heights of the bubble defect area are nonuniform, the illumination intensity difference of different positions is large, the seed points are arranged in a dark area in the bubble defect area according to the characteristics of a watershed algorithm, and the gray value of the pixel points in the dark area in the bubble defect area is low compared with that of surrounding areas due to the fact that the backlight of the dark area in the bubble defect area is low, and therefore in the embodiment of the invention, the positions of the seed points can be obtained according to the following rules.
In the embodiment of the invention, each pixel point in the wallboard surface image is taken as the center to constructA large area is used as a neighborhood area of each pixel point; in the embodiment of the invention, the presetIn other embodiments, the practitioner may set according to the particular implementationIs a value of (2).
Acquisition of the firstThe likelihood that a pixel point is a seed point:
in the method, in the process of the invention,represents the firstThe likelihood that a pixel point is a seed point;represents the firstGray value variances of all pixel points in a neighborhood region of each pixel point;represents the firstGray average value and first pixel point of all pixel points in neighborhood region of each pixel pointDifference of gray values of the individual pixel points; when the first isThe gray value variance of all the pixels in the neighborhood region of each pixel is larger, the description is thatThe gray scale distribution of the pixel points in the neighborhood region of each pixel point is more discrete, at the moment, the first pixel pointThe pixel points are more likely to be positioned in the bubble defect area and are more likely to be seed points;when the first isGray average value and first pixel point of all pixel points in neighborhood region of each pixel pointThe gray value difference of each pixel point is larger, the description is thatThe pixel points are more likely to be located in the dark area in the bubble defect, thusThe individual pixel points are more likely to be seed points; The larger the value of (2), the description of the (1)The greater the likelihood that a pixel point is a seed point. And similarly, the possibility that each pixel point is a seed point is obtained.
Sequencing all pixel points from big to small in probability of being seed points, and sequentially obtainingThe pixel points are used as the alternative seed points, and in the embodiment of the invention, the presetIn other embodiments, the practitioner may set according to the particular implementationIs a value of (2). When other alternative seed points exist in the eight neighborhood of any alternative seed point, the alternative seed point with the minimum gray value in the eight neighborhood of the alternative seed point is used as a seed point, and when other alternative seed points do not exist in the eight neighborhood of any alternative seed point, the alternative seed point is used as a seed point, all seed points are obtained, and all seed points are obtained.
To this end, all seed points in the wallboard surface image were obtained.
The segmentation threshold acquisition module 103 acquires a dark region segmentation threshold and a bright region segmentation threshold according to seed points in the wallboard surface image.
It should be noted that, because the shape of the bubble defect area is convex, the gray value of the bubble defect area is different, that is, there is a bright area larger than the gray value of the normal area in the bubble defect area, and there is a dark area smaller than the gray value of the normal area in the bubble defect area, the complete bubble defect area cannot be segmented only by a single threshold value of the watershed segmentation algorithm, so the invention aims to obtain a dark area segmentation threshold value and a bright area segmentation threshold value.
In the embodiment of the invention, a dark region segmentation threshold range is acquired: sequencing all gray values between the minimum gray value in all the seed points and the gray value with the largest number of pixel points in the wallboard surface image according to the sequence from small to large, and then using the sequenced gray values as a dark area segmentation threshold range; and sequencing all gray values between the gray value with the largest number of pixels in the wallboard surface image and the maximum gray value in the wallboard surface image in order from small to large, and then taking the sequenced gray values as a bright area segmentation threshold range.
It is to be noted that the dark region segmentation threshold is assumed to beSince the dark area is small in the wallboard surface image, the gray value gradually increasesThe area change of the immersed area in the wallboard surface image is smaller, and the gray level distribution is concentrated due to the larger area of the normal area, when the gray level value exceedsWhen the ratio of the difference between the areas of the immersed area corresponding to the next gray value and the difference between the areas of the immersed area corresponding to the previous gray value to the difference between the areas of the immersed area corresponding to the previous gray value and the corresponding immersed area reaches the maximum value, the current gray value is more likely to be a dark area segmentation threshold value; the gray scale distribution in the known dark area is more discrete, so when the gray scale value distribution discrete degree of the immersed area corresponding to any gray scale value is larger, the current gray scale value is more likely to be a dark area segmentation threshold value; the dark area has larger gray scale difference from the normal area and the bright area, so that similar points on the edge of the dark area have larger gradient values, and therefore, when the gradient value of a pixel point on the edge of the immersed area corresponding to any gray scale value is larger, the immersed area corresponding to the current gray scale value is more likely to be a complete dark area, and the current gray scale value is more likely to be a dark area segmentation threshold; when the shapes of the immersed areas corresponding to any gray value are more similar, the immersed area corresponding to the current gray value is more likely to be a complete dark area; the probability that each gradation value in the dark region division threshold range is the dark region division threshold is thus obtained from the above-described features.
In the embodiment of the invention, any gray value in a dark region segmentation threshold range is marked as a current gray value, connected region analysis is carried out on pixels smaller than the current gray value in the wallboard surface image, and the obtained connected region is marked as an immersed region corresponding to the current gray value; the shape similarity between the immersed areas corresponding to the current gray values is obtained by using a shape context algorithm, and it should be noted that the shape context algorithm is a known technology, and in the embodiment of the present invention, too much description is not given.
Acquiring the possibility that the current gray value is a dark region segmentation threshold value:
in the method, in the process of the invention,representing the possibility that the current gray value is a dark region segmentation threshold;representing the difference value between the area of the immersed area corresponding to the previous gray value of the current gray value in the dark area segmentation threshold range and the area of the immersed area corresponding to the current gray value;representing the area difference value of the immersed area corresponding to the next gray value in the dark area segmentation threshold range of the current gray value and the immersed area corresponding to the current gray value;representing the area change of the immersed area corresponding to the current gray value, the area occupied by the dark area in the wallboard surface image is smaller, therefore, when When the value of the current gray value reaches the maximum value, the later gray value of the current gray value in the dark region segmentation threshold value range exceeds the dark region segmentation threshold value, and the immersed region corresponding to the current gray value only comprises the dark region, so that the probability that the current gray value is the dark region segmentation threshold value is higher;representing the number of pixel points in the immersed area corresponding to the current gray value;representing the first submerged area corresponding to the current gray valueGray values of the individual pixels;representing the average value of all pixel points in the immersed area corresponding to the current gray value;representing the gray level distribution discrete degree of the immersed area corresponding to the current gray level value, when the value is larger, the pixel points of the dark area contained in the immersed area corresponding to the current gray level value are indicated, and the probability that the current gray level value is the dark area segmentation threshold value is larger;representing the shape similarity between the immersed areas corresponding to the current gray value, when the shape similarity is larger, the shape between the immersed areas corresponding to the current gray value is more similar, and the probability that the current gray value is a dark area segmentation threshold value is higher;representing a normalization function;representing the number of edge pixel points of the immersed area corresponding to the current gray value; Representing the first immersion area corresponding to the current gray valueGradient values of the edge pixels;representing the gradient mean value of the pixel points at the edge of the immersed area corresponding to the current gray value, and when the gradient mean value is larger, indicating that the edge of the immersed area corresponding to the current gray value is a wallThe greater the likelihood that the current gray level will be the dark region segmentation threshold, i.e. the complete dark region in the wallboard surface image can be segmented using the current gray level,the larger the value of (c), the greater the likelihood that the current gray value is the dark region segmentation threshold.
And acquiring the possibility that each gray value in the dark region segmentation threshold range is the dark region segmentation threshold, and taking the gray value with the highest possibility in the dark region segmentation threshold range as the dark region segmentation threshold.
It is to be noted that, assuming that the bright region division threshold isSince the area of the normal region is large and the gray distribution is concentrated, when the gray value is gradually increasedWhen the area of the immersed area in the wallboard surface image changes greatly; because the bright area occupies smaller area in the wallboard surface image, when the gray value exceedsWhen the area change of the immersed area in the wallboard surface image is smaller, for any gray value, when the ratio of the area difference between the immersed area corresponding to the previous gray value and the immersed area corresponding to the previous gray value to the area difference between the immersed area corresponding to the next gray value and the immersed area corresponding to the next gray value reaches the maximum value, the gray value is more likely to be a bright area segmentation threshold; the gray scale distribution in the known bright region is more discrete, so that when the gray scale distribution of the non-immersed region corresponding to any gray scale value is more discrete, the gray scale value is more likely to be a bright region segmentation threshold value; the bright area has larger gray scale difference from the normal area and the dark area, so the pixel points on the edge of the bright area have larger gradient values, and therefore, when any gray scale value corresponds to the edge of the non-immersed area When the gradient value of the pixel point of (2) is larger, the non-immersed area corresponding to the gray value is more likely to be a complete bright area, and the gray value is more likely to be a bright area segmentation threshold; when the shapes of the non-immersed areas corresponding to any gray value are more similar, the immersed areas corresponding to the gray value are more likely to be complete bright areas; the likelihood that each gray value in the bright region segmentation threshold range is the bright region segmentation threshold is thus obtained from the above-described features.
In the embodiment of the invention, any gray value in a bright region segmentation threshold range is marked as a current gray value, a pixel point which is larger than the current gray value in a wallboard surface image is subjected to connected region analysis, and the obtained connected region is marked as a non-immersed region corresponding to the current gray value; carrying out connected domain analysis on pixel points smaller than the current gray value in the wallboard surface image, and marking the obtained connected region as an immersed region corresponding to the current gray value;
acquiring the shape similarity between the non-immersed areas corresponding to the gray values by using a shape context algorithm;
the possibility that the gray value is a bright area segmentation threshold value is obtained:
in the method, in the process of the invention,representing the possibility that the gray value is a bright area segmentation threshold value; Representing the difference value between the area of the immersed area corresponding to the previous gray value of the current gray value in the bright area segmentation threshold range and the area of the immersed area corresponding to the current gray value;representing the area difference value of the immersed area corresponding to the last gray value of the gray value in the bright area segmentation threshold range and the immersed area corresponding to the gray value;representing the area change of the immersed area corresponding to the gray value, the area ratio of the bright area in the wallboard surface image is known to be small, so that whenWhen the ratio of the gray value is maximum, the gray value is a dark area and a normal area in the immersed area corresponding to the previous gray value in the bright area segmentation threshold range,the current gray value is larger, and a part of bright area is added in the immersed area corresponding to the last gray value in the bright area segmentation threshold range,the value of (2) is smaller, thus whenWhen the value of (2) is larger, the probability that the gray value is a bright area segmentation threshold value is higher;representing the number of pixels in the non-immersed area corresponding to the gray value;represents the first non-immersed area corresponding to the gray valueGray values of the individual pixels;representing the average value of all pixel points in the non-immersed area corresponding to the gray value; Representing the gray distribution dispersion degree of the non-immersed area corresponding to the gray value, when the gray distribution dispersion degree is larger, the image of the bright area contained in the non-immersed area corresponding to the gray value is describedThe greater the probability that the gray value is the bright area segmentation threshold value is;representing the shape similarity between the non-immersed areas corresponding to the gray value, when the shape similarity is larger, the shape between the non-immersed areas corresponding to the gray value is more similar, and the probability that the gray value is a bright area segmentation threshold value is higher;
representing the number of edge pixel points of the non-immersed area corresponding to the gray value;represents the first non-immersed area corresponding to the gray valueGradient values of the edge pixels;representing the gradient mean value of the pixel points at the edge of the non-immersed area corresponding to the gray value, when the gradient mean value is larger, the probability that the gray value is a bright area segmentation threshold value is larger,the larger the value of (c) is, the greater the possibility that the gray value is the bright region division threshold value is.
And acquiring the possibility that each gray value in the bright region segmentation threshold range is the bright region segmentation threshold, and taking the gray value with the highest possibility as the bright region segmentation threshold.
So far, according to the seed points in the wallboard surface image, the dark area segmentation threshold value and the bright area segmentation threshold value are obtained.
The bubble defect area obtaining module 104 obtains a plurality of suspected areas according to the dark area segmentation threshold and the bright area segmentation threshold, obtains the possibility that each suspected area is a bubble defect area, and obtains the bubble defect area according to the possibility that each suspected area is a bubble defect area.
It should be noted that, according to the obtained dark region segmentation threshold and the bright region segmentation threshold, the dark region and the bright region which may be the bubble defect regions are segmented, and it is known that overlapping pixels exist on the edges of the dark region and the bright region which belong to the same bubble defect region, so that the segmented dark region and the bright region are combined according to the characteristics to obtain each suspected region, at this time, the bubble defect regions need to be screened out from the suspected regions, and because the dark region and the pixel on one edge of the bright region of the same bubble defect region are completely overlapped, when the ratio of the number of overlapping pixels of all the connected regions in the suspected region to the number of edge pixels of all the connected regions in the suspected region is larger, that is, when the coincidence of the suspected region is larger, the current suspected region is more likely to be the bubble defect region; since the bubble defect area is approximately circular, when the distances from each overlapping pixel point of all the connected areas in the suspected area to the centroid of the suspected area are similar, the current suspected area is more likely to be the bubble defect area; finally, when the gradient direction of each overlapped pixel point of the communication area in the suspected area is more similar to the difference of the illumination direction angle of the wallboard surface image In this case, it is explained that the more likely the gradient directions of the overlapping pixel points of the connected regions in the suspected region are uniform, the more likely the suspected region is a bubble defect region.
In the embodiment of the invention, a suspected region is acquired: and carrying out connected domain analysis on pixel points smaller than a dark region segmentation threshold in the wallboard surface image to obtain each connected region, carrying out connected domain analysis on pixel points larger than a bright region segmentation threshold in the wallboard surface image to obtain each connected region, and merging the connected regions with overlapped pixels to obtain a plurality of suspected regions.
Obtaining the illumination direction angle of the wallboard surface image: and (3) acquiring the mode of the gradient directions of all pixel points in the wallboard surface image, marking the mode as a first gradient direction, and taking the opposite vector of the first gradient direction as the illumination direction angle of the wallboard surface.
In the embodiment of the invention, any suspected area is marked as a current suspected area, and the possibility that the current suspected area is a bubble defect area is obtained:
in the method, in the process of the invention,representing the possibility that the current suspected region is a bubble defect region;representing the number of overlapped pixel points of all the connected areas in the current suspected area;representing the number of edge pixel points of all the connected areas in the current suspected area; Representing the coincidence degree of the current suspected region, and indicating that the current suspected region is more likely to be a bubble defect region when the value is larger;representing the number of overlapped pixel points of each communication area in the current suspected area;representing the first connected region of the current suspected regionThe distances between the overlapped pixel points and the centroid of the current suspected region;representing the average value of the distances between all overlapped pixel points of each connected region in the current suspected region and the mass center of the current suspected region;the smaller the value of the degree representing the approximate circle of the current suspected region, the more the shape of the current suspected region approaches to the circle, the greater the possibility that the current suspected region is a bubble defect region;
representing the first connected region of the current suspected regionGradient direction angles of the overlapped pixel points;an illumination direction angle representing an image of a wallboard surface; the closer the difference between the gradient direction angle of each overlapping pixel point of each connected region in the current suspected region and the illumination direction angle of the wallboard surface imageWhen the gradient directions of the overlapped pixel points of the communication areas in the current suspected area are more likely to be consistent, namely, the pixel points overlapped by the communication areas in the current suspected area are taken as the pixel points on the overlapped edges of the bright area and the dark area of the bubble defect area; And when the gradient direction consistency representing the current suspected region is larger, the probability that the current suspected region is a bubble defect region is larger.
Presetting a likelihood thresholdWhen the probability that any suspected region is a bubble defect region is greater than or equal to the probability thresholdWhen the suspected region is considered to be a bubble defect region, in the embodiment of the invention, a probability threshold is presetIn itIn other embodiments, the practitioner may set the likelihood threshold according to the particular implementationIs a value of (2).
And acquiring a plurality of suspected areas according to the dark area segmentation threshold and the bright area segmentation threshold, acquiring the possibility that each suspected area is a bubble defect area, and acquiring the bubble defect area according to the possibility that each suspected area is the bubble defect area.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (3)
1. An automatic plastering effect visual detection system based on image feature analysis, which is characterized by comprising:
the wallboard surface image acquisition module is used for acquiring wallboard surface images;
A seed point acquisition module; the method comprises the steps of acquiring a neighborhood region of each pixel point in a wallboard surface image, and acquiring the possibility that each pixel point is a seed point according to the gray value variance of all pixel points in the neighborhood region of each pixel point and the difference between the gray value of each pixel point and the gray average value of all pixel points in the neighborhood region of each pixel point; acquiring a seed point according to the possibility that each pixel point is the seed point;
the segmentation threshold acquisition module is used for acquiring a dark region segmentation threshold range and a bright region segmentation threshold range according to the seed points; acquiring an immersed region corresponding to each gray value in a dark region segmentation threshold range; acquiring the possibility that each gray value in the dark region segmentation threshold range is a dark region segmentation threshold according to the gray distribution discrete degree of the immersed region corresponding to each gray value in the dark region segmentation threshold range, the area change of the immersed region, the gradient mean value of the pixel points at the edge of the immersed region and the shape similarity among the immersed regions, and acquiring the dark region segmentation threshold according to the possibility that each gray value in the dark region segmentation threshold range is a dark region segmentation threshold; acquiring an immersed region and an unsubmerged region corresponding to each gray value in a bright region segmentation threshold range; acquiring the possibility that each gray value in the bright region segmentation threshold range is a bright region segmentation threshold according to the gray distribution discrete degree of the non-immersed region corresponding to each gray value in the bright region segmentation threshold range, the area change of the immersed region, the gradient mean value of the edge pixel points of the non-immersed region and the shape similarity between the non-immersed regions, and acquiring the bright region segmentation threshold according to the possibility that each gray value in the bright region segmentation threshold range is a bright region segmentation threshold;
The bubble defect region acquisition module is used for acquiring a plurality of suspected regions according to the bright region segmentation threshold and the dark region segmentation threshold; acquiring the possibility that each suspected region is a bubble defect region according to the approximate circle degree of the suspected region, the coincidence degree of the suspected region and the gradient direction consistency of the suspected region; acquiring bubble defect areas according to the possibility that each suspected area is a bubble defect area;
the step for obtaining the neighborhood region of each pixel point in the wallboard surface image, and obtaining the possibility that each pixel point is a seed point according to the gray value variance of all the pixel points in the neighborhood region of each pixel point and the difference between the gray value of each pixel point and the gray average value of all the pixel points in the neighborhood region of each pixel point, comprises the following steps:
presettingConstruction of +.>A large area is used as a neighborhood area of each pixel point;
in the method, in the process of the invention,represents->The likelihood that a pixel point is a seed point; />Represents->Gray value variances of all pixel points in a neighborhood region of each pixel point; />Represents->Gray average value and +.th of all pixels in neighborhood region of each pixel >Difference of gray values of the individual pixel points; acquiring the possibility that each pixel point is a seed point;
the step of obtaining the seed points according to the possibility that each pixel point is the seed point comprises the following steps:
presettingOrdering all pixel points from big to small, and sequentially obtaining +.>The method comprises the steps of taking each pixel point as an alternative seed point, taking the alternative seed point with the smallest gray value in the eight neighborhood of any alternative seed point as one seed point when other alternative seed points exist in the eight neighborhood of any alternative seed point, obtaining all seed points, taking the alternative seed point as one seed point when other alternative seed points do not exist in the eight neighborhood of any alternative seed point, and obtaining all seed pointsSeed points;
the step of obtaining the immersed area corresponding to each gray value in the dark area segmentation threshold range comprises the following steps:
marking any gray value in a dark area segmentation threshold range as a current gray value, carrying out connected domain analysis on pixel points smaller than the current gray value in the wallboard surface image, and marking the obtained connected region as an immersed region corresponding to the current gray value;
according to the gray level distribution discrete degree of the immersed area corresponding to each gray level value in the dark area segmentation threshold range, the area change of the immersed area, the gradient mean value of the edge pixel points of the immersed area and the shape similarity among the immersed areas, the possibility that each gray level value in the dark area segmentation threshold range is the dark area segmentation threshold is obtained, and the possibility that each gray level value in the dark area segmentation threshold range is the dark area segmentation threshold is obtained according to the possibility that each gray level value in the dark area segmentation threshold range is the dark area segmentation threshold, the steps include:
Acquiring shape similarity between immersed areas corresponding to the current gray values by using a shape context algorithm;
in the method, in the process of the invention,representing the possibility that the current gray value is a dark region segmentation threshold; />Representing the difference value between the area of the immersed area corresponding to the previous gray value of the current gray value in the dark area segmentation threshold range and the area of the immersed area corresponding to the current gray value; />Representing the area difference value of the immersed area corresponding to the next gray value in the dark area segmentation threshold range of the current gray value and the immersed area corresponding to the current gray value; />Representing the area change of the immersed area corresponding to the current gray value; />Representing the number of pixel points in the immersed area corresponding to the current gray value; />Represents the +.f in the submerged area corresponding to the current gray value>Gray values of the individual pixels; />Representing the average value of all pixel points in the immersed area corresponding to the current gray value;representing the gray level distribution discrete degree of the immersed area corresponding to the current gray level value; />Representing the shape similarity between the immersed areas corresponding to the current gray values; />Representing the number of edge pixel points of the immersed area corresponding to the current gray value;represents the +.f. of the immersed area corresponding to the current gray value >Gradient values of the edge pixels; />Representing the current gray levelGradient average value of the edge pixel points of the immersed area corresponding to the value; />Representing a normalization function;
acquiring the possibility that each gray value in the dark region segmentation threshold range is the dark region segmentation threshold, and taking the gray value with the highest possibility in the dark region segmentation threshold range as the dark region segmentation threshold;
the step of acquiring the immersed area and the non-immersed area corresponding to each gray value in the bright area segmentation threshold range comprises the following steps:
marking any gray value in the bright region segmentation threshold range as a current gray value, carrying out connected region analysis on pixel points larger than the current gray value in the wallboard surface image, and marking the obtained connected region as a non-immersed region corresponding to the current gray value; carrying out connected domain analysis on pixel points smaller than the current gray value in the wallboard surface image, and marking the obtained connected region as an immersed region corresponding to the current gray value;
according to the gray level distribution discrete degree of the non-immersed area corresponding to each gray level value in the bright area segmentation threshold range, the area change of the immersed area, the gradient mean value of the edge pixel points of the non-immersed area and the shape similarity among the non-immersed areas, the possibility that each gray level value in the bright area segmentation threshold range is the bright area segmentation threshold is obtained, and according to the possibility that each gray level value in the bright area segmentation threshold range is the bright area segmentation threshold, the bright area segmentation threshold is obtained, the steps are as follows:
Acquiring the shape similarity between the non-immersed areas corresponding to the gray values by using a shape context algorithm;
in the method, in the process of the invention,representing the possibility that the gray value is a bright area segmentation threshold value;/>representing the difference value between the area of the immersed area corresponding to the previous gray value of the current gray value in the bright area segmentation threshold range and the area of the immersed area corresponding to the current gray value; />Representing the area difference value of the immersed area corresponding to the last gray value of the gray value in the bright area segmentation threshold range and the immersed area corresponding to the gray value; />Representing the area change of the immersed area corresponding to the gray value;representing the number of pixels in the non-immersed area corresponding to the gray value; />Represents the +.f. in the non-immersed area corresponding to the current gray value>Gray values of the individual pixels; />Representing the average value of all pixel points in the non-immersed area corresponding to the gray value; />Representing the gray level distribution discrete degree of the non-immersed area corresponding to the gray level value; />Representing the shape similarity between the non-immersed areas corresponding to the gray values; />Representing the number of edge pixel points of the non-immersed area corresponding to the gray value; />Represents the +.f. of the non-immersed area corresponding to the gray value of this time >Gradient values of the edge pixels; />Representing the gradient average value of the pixel points at the edge of the non-immersed area corresponding to the gray value; />Representing a normalization function;
acquiring the possibility that each gray value in the bright area segmentation threshold range is the bright area segmentation threshold, and taking the gray value with the highest possibility in the bright area segmentation threshold range as the bright area segmentation threshold;
acquiring the possibility that each suspected region is a bubble defect region according to the approximate circle degree of the suspected region, the coincidence degree of the suspected region and the gradient direction consistency of the suspected region; acquiring the bubble defect areas according to the possibility that each suspected area is the bubble defect area, wherein the method comprises the following steps:
acquiring the mode of the gradient directions of all pixel points in the wallboard surface image, marking the mode as a first gradient direction, and taking the opposite vector of the first gradient direction as the illumination direction angle of the wallboard surface;
any suspected area is marked as a current suspected area;
in the method, in the process of the invention,representing the current doubtThe likelihood that the alike area is a bubble defect area; />Representing the number of overlapped pixel points of all the connected areas in the current suspected area; />Representing the number of edge pixel points of all the connected areas in the current suspected area; / >Representing the coincidence degree of the current suspected region; />Representing the number of overlapped pixel points of each communication area in the current suspected area; />Represents the +.about.of each connected region in the current suspected region>The distances between the overlapped pixel points and the centroid of the current suspected region; />Representing the average value of the distances between all overlapped pixel points of each connected region in the current suspected region and the mass center of the current suspected region; />Representing the degree to which the current suspected region is approximately circular; />Representing the gradient direction consistency of the current suspected region; />Represents the +.about.of each connected region in the current suspected region>Gradient direction angles of the overlapped pixel points; />An illumination direction angle representing an image of a wallboard surface; />Representing absolute value symbols; acquiring the possibility that each suspected region is a bubble defect region;
presetting a likelihood thresholdWhen the possibility that any of the suspected regions is a bubble defective region is equal to or greater than the possibility threshold +.>And when the suspected area is a bubble defect area, obtaining all bubble defect areas.
2. The visual inspection system for the effect of automatic plastering based on image feature analysis according to claim 1, wherein the steps of obtaining the dark region segmentation threshold range and the bright region segmentation threshold range according to the seed points include:
Sequencing all gray values between the minimum gray value in all the seed points and the gray value with the largest number of pixel points in the wallboard surface image according to the sequence from small to large, and then using the sequenced gray values as a dark area segmentation threshold range; and sequencing all gray values between the gray value with the largest number of pixels in the wallboard surface image and the maximum gray value in the wallboard surface image in order from small to large, and then taking the sequenced gray values as a bright area segmentation threshold range.
3. The visual inspection system for automatically inspecting plastering effect based on image feature analysis according to claim 1, wherein the steps of obtaining a plurality of suspected areas according to the bright area segmentation threshold and the dark area segmentation threshold comprise:
and carrying out connected domain analysis on pixel points smaller than a dark region segmentation threshold in the wallboard surface image to obtain each connected region, carrying out connected domain analysis on pixel points larger than a bright region segmentation threshold in the wallboard surface image to obtain each connected region, and merging the connected regions with overlapped pixels to obtain a plurality of suspected regions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311587747.7A CN117291937B (en) | 2023-11-27 | 2023-11-27 | Automatic plastering effect visual detection system based on image feature analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311587747.7A CN117291937B (en) | 2023-11-27 | 2023-11-27 | Automatic plastering effect visual detection system based on image feature analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117291937A CN117291937A (en) | 2023-12-26 |
| CN117291937B true CN117291937B (en) | 2024-03-05 |
Family
ID=89244837
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311587747.7A Active CN117291937B (en) | 2023-11-27 | 2023-11-27 | Automatic plastering effect visual detection system based on image feature analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117291937B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117557568B (en) * | 2024-01-12 | 2024-05-03 | 吉林省迈达医疗器械股份有限公司 | Focal region segmentation method in thermal therapy process based on infrared image |
| CN117911408B (en) * | 2024-03-19 | 2024-05-28 | 盈客通天下科技(大连)有限公司 | Road pavement construction quality detection method and system |
| CN118154578B (en) * | 2024-04-11 | 2024-09-24 | 四川省第六建筑有限公司 | ALC panel splicing integrity detection method |
| CN118071776B (en) * | 2024-04-22 | 2024-06-21 | 大连云智信科技发展有限公司 | Image processing system for detecting intestinal health condition of poultry |
| CN118115502B (en) * | 2024-04-30 | 2024-08-27 | 中建安装集团有限公司 | Sewage discharge monitoring method and system based on image features |
| CN118608531B (en) * | 2024-08-08 | 2024-10-11 | 瑞安市瑞鑫电器有限公司 | Motor carbon brush frame wear unbalance degree detection method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101996405A (en) * | 2010-08-30 | 2011-03-30 | 中国科学院计算技术研究所 | Method and device for rapidly detecting and classifying defects of glass image |
| CN106683076A (en) * | 2016-11-24 | 2017-05-17 | 南京航空航天大学 | Texture feature clustering-based locomotive wheelset tread damage detection method |
| CN115861320A (en) * | 2023-02-28 | 2023-03-28 | 天津中德应用技术大学 | Intelligent detection method for automobile part machining information |
| WO2023077404A1 (en) * | 2021-11-05 | 2023-05-11 | 宁德时代新能源科技股份有限公司 | Defect detection method, apparatus and system |
| CN116402810A (en) * | 2023-06-05 | 2023-07-07 | 山东天力润滑油有限公司 | Image processing-based lubricating oil anti-abrasive particle quality detection method |
| CN116843678A (en) * | 2023-08-28 | 2023-10-03 | 青岛冠宝林活性炭有限公司 | Hard carbon electrode production quality detection method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9269155B2 (en) * | 2012-04-05 | 2016-02-23 | Mediatek Singapore Pte. Ltd. | Region growing method for depth map/color image |
-
2023
- 2023-11-27 CN CN202311587747.7A patent/CN117291937B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101996405A (en) * | 2010-08-30 | 2011-03-30 | 中国科学院计算技术研究所 | Method and device for rapidly detecting and classifying defects of glass image |
| CN106683076A (en) * | 2016-11-24 | 2017-05-17 | 南京航空航天大学 | Texture feature clustering-based locomotive wheelset tread damage detection method |
| WO2023077404A1 (en) * | 2021-11-05 | 2023-05-11 | 宁德时代新能源科技股份有限公司 | Defect detection method, apparatus and system |
| CN115861320A (en) * | 2023-02-28 | 2023-03-28 | 天津中德应用技术大学 | Intelligent detection method for automobile part machining information |
| CN116402810A (en) * | 2023-06-05 | 2023-07-07 | 山东天力润滑油有限公司 | Image processing-based lubricating oil anti-abrasive particle quality detection method |
| CN116843678A (en) * | 2023-08-28 | 2023-10-03 | 青岛冠宝林活性炭有限公司 | Hard carbon electrode production quality detection method |
Non-Patent Citations (3)
| Title |
|---|
| An Automatic Approach for Retinal Vessel Segmentation by Multi-Scale Morphology and Seed Point Tracking;Wang, WH等;JOURNAL OF MEDICAL IMAGING AND HEALTH INFORMATICS;全文 * |
| 形态学分水岭和Fisher线性判别的图像分割算法研究;黄娜;中国优秀硕士学位论文全文数据库信息科技辑(第07期);全文 * |
| 李建奇 ; 阳春华 ; 曹斌芳 ; 朱红求 ; 刘金平 ; .面向参数测量的改进分水岭浮选泡沫图像分割方法.仪器仪表学报.2013,(06),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117291937A (en) | 2023-12-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117291937B (en) | Automatic plastering effect visual detection system based on image feature analysis | |
| CN107679535B (en) | Automatic reading identification system and method for pointer type water meter based on template matching | |
| CN104992434B (en) | The method for carrying out typhoon location on present satellites image according to historical satellite image | |
| CN115020267B (en) | Semiconductor surface defect detection method | |
| CN106228161B (en) | A kind of pointer-type dial plate automatic reading method | |
| CN106651880B (en) | Offshore moving target detection method based on multi-feature fusion thermal infrared remote sensing image | |
| CN117408995B (en) | Power adapter appearance quality detection method based on multi-feature fusion | |
| CN116758057A (en) | Communication equipment defect detection method based on artificial intelligence | |
| CN115641327B (en) | Building engineering quality supervision and early warning system based on big data | |
| CN110211128B (en) | Loess plateau terrace extraction method based on remote sensing image and DEM | |
| CN117274240B (en) | Bearing platform foundation concrete surface crack identification method | |
| CN110807763A (en) | Method and system for detecting ceramic tile surface bulge | |
| CN115983141B (en) | Deep learning-based method, medium and system for inverting wave height of ocean waves | |
| CN115049736A (en) | Method for accurately positioning surface defects of cable | |
| CN117557584B (en) | Water body extraction method and device, electronic equipment and storage medium | |
| CN116883408A (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
| CN113899349B (en) | Sea wave parameter detection method, equipment and storage medium | |
| CN118225803A (en) | Visual detection method for appearance of blade surface of bulldozer | |
| CN117455870A (en) | Connecting wire and connector quality visual detection method | |
| CN115984360B (en) | Method and system for calculating length of dry beach based on image processing | |
| CN116309447B (en) | Dam slope crack detection method based on deep learning | |
| CN115578695B (en) | Water gauge water level machine vision detection method and device with free shooting visual angle | |
| CN111307267A (en) | Conductor galloping monitoring method based on concentric circle detection | |
| CN115546115A (en) | Ice-coated ice-shaped picture segmentation and identification method based on geographic morphology | |
| CN114677428A (en) | Power transmission line icing thickness detection method based on unmanned aerial vehicle image processing |
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 |