CN112633393A - Automatic classification method and device for ceramic tile textures - Google Patents
Automatic classification method and device for ceramic tile textures Download PDFInfo
- Publication number
- CN112633393A CN112633393A CN202011589821.5A CN202011589821A CN112633393A CN 112633393 A CN112633393 A CN 112633393A CN 202011589821 A CN202011589821 A CN 202011589821A CN 112633393 A CN112633393 A CN 112633393A
- Authority
- CN
- China
- Prior art keywords
- tile
- image
- area image
- template library
- texture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 239000000919 ceramic Substances 0.000 title claims description 41
- 238000007781 pre-processing Methods 0.000 claims abstract description 20
- 230000004044 response Effects 0.000 claims abstract description 19
- 238000001914 filtration Methods 0.000 claims abstract description 14
- 238000012937 correction Methods 0.000 claims description 20
- 230000009466 transformation Effects 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000000926 separation method Methods 0.000 abstract description 4
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000001514 detection method Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 3
- 230000000877 morphologic effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/55—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5862—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/60—Rotation of whole images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- 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
- G06T2207/10024—Color 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
- G06T2207/20024—Filtering details
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Library & Information Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method and a device for automatically classifying tile textures. Preprocessing the collected tile images, adding the first tile area image and the corresponding textural features into a template library, wherein the steps of adding the first tile area image and the corresponding textural features comprise reserving the tile area image for four angles of 0 degree, rotating for 90 degrees, rotating for 180 degrees and rotating for 270 degrees, and extracting the corresponding textural features respectively; and for the subsequent tile area images, extracting texture features, performing relevant filtering operation of space-time constraint in the feature space of the template library, searching the position with the maximum response value, and further evaluating the similarity. And when the similarity exceeds a threshold value, outputting the corresponding version number and the corresponding rotation angle, otherwise, adding the image and the corresponding texture feature to the template library. The method does not need to configure a standard template in advance, can automatically separate the tiles of each batch, and has high calculation efficiency and high separation accuracy in the process of separating the tiles.
Description
Technical Field
The invention relates to the technical field of computer vision processing, in particular to a method and a device for automatically classifying tile textures.
Background
In the production process of the ceramic tile, the color, the patterns and the patterns of the ceramic tile are sprayed and printed on a blank through an ink jet machine, and then the blank is sequentially sintered and polished by a kiln to obtain a final product. The existing ceramic tiles in the market mainly have the types of fixed plate grains, continuous grains, infinite continuous grains and the like, and the continuous grain tiles have coherent, atmospheric and simple decoration characteristics because the continuous grain tiles can completely present the grains of natural stone, and are popular with market consumers in recent years, so that ceramic tile factories generally popularize and produce the continuous grain tiles.
The size of a single ceramic tile with continuous grains is 600 multiplied by 1200mm to 2400 multiplied by 1200mm, the size after splicing is far larger than the size which can be produced by a kiln at one time, when the continuous grains are actually produced, the ceramic tiles in the same batch often contain various different patterns, and are influenced by various factors such as material proportion and kiln temperature in the production process, and even if the ceramic tiles with the same patterns have color deviation, the color difference is inevitable. Therefore, in order to facilitate sales, a ceramic tile factory needs to divide the pattern of the ceramic tile after polishing the ceramic tile, and separate the color of the ceramic tile with different colors, and then pack the ceramic tile with the same pattern and color.
The plate-dividing and color-separating detection process of the traditional ceramic tile depends on visual distinguishing and manual marking of experienced workers, and then sorting equipment realizes sorting and packaging of the ceramic tile through identifying manual marking. However, the manual detection mode has a limit, and generally does not exceed 6 plate surface patterns, otherwise, the detection accuracy cannot be guaranteed. In addition, the worker is inevitably fatigued and distracted after working for a long time, resulting in low detection efficiency and unstable accuracy. And the existing mode can not give consideration to plate separation and color separation, and can only be finished as two independent processes, thus spending a large amount of manual labor. Therefore, the development of the full-automatic plate separating and color separating method and device has important significance for improving the production efficiency of the ceramic tiles, improving the production environment and reducing the resource waste.
The tile version dividing method based on the deep learning model is used for pre-training before online version dividing, for example, a method and a device for classifying tiles by using computer vision and deep learning are provided by 'CN 109614994A-a tile classification and identification method and device'. Firstly, acquiring images of tiles with different patterns by using a camera, and establishing a database with a certain scale; then, training the model according to the corresponding relation between the image and the pattern label; and finally, carrying out online plate division detection on the ceramic tile by using the trained model. The problem with this approach is that the model needs to be trained in advance, and the trained model cannot be used directly for different batches of tile patterns, and is therefore difficult to deploy in the actual production process.
The existing automatic color separation method for the ceramic tiles comprises the steps of firstly collecting images by using a camera, then extracting color features of the ceramic tile images, comparing the color features with the color features of template images collected in advance, determining the color number as the current color number when the color number is smaller than a preset threshold value, and otherwise determining the color number as a new color number, such as a searching and comparing device for similar ceramic tiles provided by CN 103324759B-an intelligent ceramic tile identification device. The problem with this type of process is that it requires standard plates of each color number, which is unpredictable before the start of production and therefore presents difficulties in implementation. And the method of presetting the fixed threshold has poor adaptability for different patterns and different production batches, and industrial workers obviously lack the professional knowledge of adjusting the system work, which causes difficulty for actual deployment.
Disclosure of Invention
The invention aims to: aiming at the existing problems, the method and the device for automatically classifying the texture of the ceramic tile are provided, and a solution for automatically finishing the plate division of the ceramic tile without depending on a preset template is provided.
The technical scheme adopted by the invention is as follows:
an automatic tile texture classification method comprises the following steps:
A. preprocessing the collected tile image to obtain a corresponding tile area image;
B. the tile area image is subjected to plate division, and the method comprises the following steps:
adding the first tile area image and the corresponding texture features thereof into a template library, wherein the steps of keeping the tile area image at four angles of 0 degree, rotating 90 degrees, rotating 180 degrees and rotating 270 degrees, and respectively extracting the corresponding texture features;
extracting texture features of the subsequent tile area images, performing relevant filtering operation of space-time constraint on the extracted texture features in a feature space of the template library, and searching a position with the maximum response value; and then carrying out similarity evaluation on the detected tile area image and the template in the matched template library, if the evaluated similarity is above a given threshold value, taking the version number corresponding to the matched template as the version number corresponding to the current tile area image, and otherwise, adding the current tile area image and the corresponding texture features thereof into the template library.
Further, the preprocessing includes processes of gray value correction, foreground segmentation and geometric transformation.
Further, the process of gray value correction includes:
and carrying out gray value correction on the whole tile image according to the three channel values in the calibration area in the tile image.
Furthermore, the calibration area is an area in the tile image corresponding to a calibration object arranged near the tile in the tile image acquisition link.
Further, the template library is set with an upper limit of the number of templates; and if the similarity between the texture features of the current tile area image and the texture features of the templates in the template library is smaller than a given threshold value and the templates in the template library reach the upper limit of the number, regarding the current tile area image, taking the version number corresponding to the maximum position of the response value when the current tile area image and the template library perform the relevant filtering operation of space-time constraint as the version number corresponding to the current tile area image.
An automatic classification device for ceramic tile textures comprises an image preprocessing unit, a ceramic tile layout dividing unit and a result output unit which are sequentially connected; wherein:
the image pre-processing unit is configured to: preprocessing an input tile image and outputting a tile area image;
the tile split unit is configured to: adding the first tile area image and the corresponding texture features thereof into a template library, wherein the steps of keeping the tile area image at four angles of 0 degree, rotating 90 degrees, rotating 180 degrees and rotating 270 degrees, and respectively extracting the corresponding texture features; extracting texture features of the subsequent tile area images, performing relevant filtering operation of space-time constraint on the extracted texture features in a feature space of the template library, and searching a position with the maximum response value; carrying out similarity evaluation on the detected tile area image and a template in a matched template library, if the evaluated similarity is above a given threshold value, taking a version number corresponding to the matched template as a version number corresponding to the current tile area image, and otherwise, adding the current tile area image and the corresponding texture features thereof into the template library;
the result output unit is configured to: and outputting the corresponding version number of the ceramic tile.
Further, the image preprocessing unit preprocesses the input tile image by processes of gray value correction, foreground segmentation and geometric transformation.
Further, the process of gray value correction includes: and carrying out gray value correction on the whole tile image according to the three channel values in the calibration area in the tile image.
Furthermore, the calibration area is in the tile image acquisition link, and the calibration object arranged near the tile corresponds to the area in the tile image.
Further, the template library is provided with an upper limit of the number of templates; and if the similarity between the texture features of the current tile area image and the texture features of the templates in the template library is smaller than a given threshold value and the templates in the template library reach the upper limit of the number, taking the version number corresponding to the maximum position of the response value when the space-time constrained related filtering operation is performed in the feature space of the current tile area image and the template library as the version number corresponding to the current tile area image.
In summary, due to the adoption of the technical scheme, the invention has the beneficial effects that:
1. the scheme for automatically classifying the ceramic tile textures does not need to configure a standard template in advance, and can automatically classify the ceramic tiles of each batch.
2. According to the scheme for automatically classifying the ceramic tile textures, the template library is configured, and images and characteristics of 4 angles of the ceramic tile are stored respectively, so that matching operation of the templates is facilitated. Meanwhile, the fuzzy matching process is set without calculating the similarity one by one, and the matching efficiency is greatly improved.
3. The scheme for automatically classifying the texture of the ceramic tile can adapt to the number of patterns of the ceramic tile to classify the ceramic tile, and can ensure the accuracy of classification.
Drawings
The invention will now be described, by way of example, with reference to the accompanying drawings, in which:
fig. 1 is a flowchart of an embodiment of a tile texture automatic classification method.
Fig. 2 is a flowchart of another embodiment of the tile texture automatic classification method.
Detailed Description
All of the features disclosed in this specification, or all of the steps in any method or process so disclosed, may be combined in any combination, except combinations of features and/or steps that are mutually exclusive.
Any feature disclosed in this specification (including any accompanying claims, abstract) may be replaced by alternative features serving equivalent or similar purposes, unless expressly stated otherwise. That is, unless expressly stated otherwise, each feature is only an example of a generic series of equivalent or similar features.
Example one
An automatic tile texture classification method, as shown in fig. 1, includes:
A. and preprocessing the collected tile image to obtain a corresponding tile area image.
For the preprocessing step of the tile image, the flow comprises the processes of gray value correction, foreground segmentation and geometric transformation. Firstly, carrying out gray value correction on the whole tile image according to three channels in a calibration area in the tile image so as to eliminate the change of the image gray value caused by the brightness of a light source and the response of a camera. The position of the tile region is then determined by foreground segmentation and the interference region is removed using morphological processing. And finally, positioning four corner points of the tile area, correcting the azimuth through geometric transformation, eliminating deflection angles, and extracting an image of the tile area.
The calibration area is a reference area designed in the tile image, and in some embodiments, in the tile image capturing step, the calibration object set near the tile corresponds to an area in the tile image. And calibrating three channel values in the area as reference values, and performing color correction on the tile images to correct the image gray scale of each tile image to the truest state so as to eliminate the influence of noise brought by the external environment on the version result.
B. And (5) carrying out plate division on the tile area image.
And after the tile area image is obtained from the tile image, classifying the tile area image according to the texture characteristics of the tile area image. Firstly, a template library is constructed, at the beginning, no template exists in the template library, for a first tile area image, the default is used as a first template to be added into the template library, corresponding texture features are generated, and a version number corresponding to the first tile area image is set (for example, a serial number in a feature space of the template library is used as the version number). Because the directions of the tiles are different and the texture features are also different, when the texture features are added to the template library, the texture features of four angles of 0 degree, 90 degree rotation, 180 degree rotation and 270 degree rotation are reserved, namely, the tile area image is reserved for 0 degree, 90 degree rotation, 180 degree rotation and 270 degree rotation respectively, and the corresponding texture features are extracted respectively.
And for the subsequently detected tile area images, extracting texture features of the tile area images, performing relevant filtering operation of space-time constraint on the extracted texture features in a feature space of a template library, and searching a position with the maximum response value. And correspondingly obtaining the suspected version number and the rotation position of the current tile area image according to the position with the maximum response value. And then, carrying out similarity evaluation on the detected tile area image and a template in a matched template library, generally the similarity evaluation between texture features, if the evaluated similarity is smaller than a given threshold (or the similarity distance is larger than a corresponding given threshold), indicating that the current tile area image is not in the template library, adding the current tile area image and the corresponding texture features thereof into the template library, if the evaluated similarity is above the given threshold, indicating that the same plate type exists in the template library in the current tile area image, and taking the plate number corresponding to the matched template as the plate number corresponding to the current tile area image. Preferably, when outputting the version number of the detected tile (i.e., the current tile), the rotation angle of the tile with respect to the form is also output to facilitate sorting arrangement.
In some embodiments, the above process comprises:
and (3) constructing a template library: and respectively reserving four angles of 0 degree, 90 degrees, 180 degrees and 270 degrees for the first tile area image to obtain 4 images, respectively extracting the texture features of each image, adding the texture features corresponding to each image into a template library, and obtaining templates with the sequence numbers of 1, 2, 3 and 4 respectively.
And extracting the texture features of the newly input tile area image, performing space-time constraint related filtering operation on the extracted texture features and the feature space of the template library, and searching the position with the maximum response value so as to preliminarily determine the version number and the rotation angle of the tile area image. And then calculating the similarity between the texture features of the newly input tile area image and the texture features at the position with the maximum response value. When the similarity is smaller than a given threshold, adding the newly input tile area image under 4 angles (0 degree is reserved, 90 degrees is rotated, 180 degrees is rotated, and 270 degrees is rotated) and the corresponding texture features (4 feature vectors) into the template library to serve as a new template, for example, when the second tile area image serves as the new template, templates with the sequence numbers of 5, 6, 7, and 8 are obtained. And when the similarity is above a given threshold, taking the version number corresponding to the matched template as the version number corresponding to the newly input tile area image.
The template-corresponding plate numbers may be the number of the template, or the same plate number may be set for each template of the same tile, for example, the above-mentioned template numbers 1, 2, 3, and 4 correspond to the first plate number, and the template numbers 5, 6, 7, and 8 correspond to the second plate number.
For the split result of the tile, the number of versions may be defined. For example, the tile type may be limited, such as solid page, lined tile, or infinite lined tile. In the corresponding plate dividing result, the fixed plate surface can limit and output a plate type; a continuous grain tile with n (n is a positive integer) patterns (layouts) can be limited to outputting up to n layouts. In the actual plate dividing process, because tiles with the same plate type are produced, deviation is inevitable, and machine plate dividing can be regarded as a new plate type for the tiles in the condition. In this regard, especially for the continuous grain tiles, as shown in fig. 2, on the premise of limiting the upper limit of the number of output versions (which can be realized by limiting the upper limit of the number of templates in the template library), if the similarity between the texture features of the newly input (current) tile area image and the texture features of each template in the template library is smaller than a given threshold, and the template in the template library has reached the upper limit of the number, that is, the newly input tile area image cannot be regarded as a new plate, the version number corresponding to the maximum position of the response value when performing the spatio-temporal constraint correlation filtering operation in the feature space of the template library is used as the version number corresponding to the newly input tile area image for the newly input tile area image.
After the newly input tile area image is subjected to plate division, the method further comprises the step of sorting the corresponding tiles to the corresponding areas based on the plate division result. Preferably, before the ceramic tiles are sorted, the method further comprises the step of marking the corresponding ceramic tiles by the version number/version type based on the version dividing result.
Example two
An automatic classification device for ceramic tile textures comprises an image preprocessing unit, a ceramic tile layout dividing unit and a result output unit which are sequentially connected. Wherein:
the image preprocessing unit is configured to: and preprocessing the input tile image and outputting a tile area image.
The preprocessing includes the processes of gray value correction, foreground segmentation and geometric transformation. The grey value correction is to perform grey value correction on the whole tile image according to three channels in a calibration area in the tile image so as to eliminate image grey value changes caused by light source brightness and camera response. And performing foreground segmentation on the image to determine the position of the tile area, and removing the interference area by using morphological processing. The geometric transformation process is the process of correcting the image azimuth, and the correction of the tile area azimuth is completed by positioning four corner points of the tile area and then carrying out geometric transformation, so that the deflection angle is eliminated, and the tile area image is extracted.
The calibration area is a reference area designed in the tile image, and in some embodiments, in the tile image capturing step, the calibration object set near the tile corresponds to the area in the tile image. And calibrating three channel values in the area as reference values, and performing color correction on the tile images to correct the image gray scale of each tile image to the truest state so as to eliminate the influence of noise brought by the external environment on the version result.
A tile tiling unit configured to: and classifying the input tile area image according to the texture characteristics of the input tile area image.
The tile layout unit is provided with a template library, the template library takes the input first tile area image and the corresponding texture characteristics as templates, and the templates in the template library are set with layout numbers. Specifically, when the tile area image is added to the template library as a template, the tile area image is respectively retained by 0 degree, rotated by 90 degrees, rotated by 180 degrees, and rotated by 270 degrees, and corresponding texture features are respectively extracted. For example, the first tile area image is respectively reserved with four angles of 0 degree, 90 degrees, 180 degrees and 270 degrees to obtain 4 images, the texture features of each image are respectively extracted, the texture features corresponding to each image are added into a template library, and templates with the sequence numbers of 1, 2, 3 and 4 are obtained. The template number may be set as a plate number, or one plate number may be set for templates corresponding to the same tile, for example, the template corresponding to the plate number one of the above-mentioned numbers 1, 2, 3, and 4.
And extracting the texture features of the subsequently input tile region images, performing space-time constrained correlation filtering operation on the extracted texture features in a feature space of a template library, and searching a position (such as a sequence number of a template) with the maximum response value. And correspondingly obtaining the suspected version number and the rotation position (angle) of the current tile area image according to the position with the maximum response value. Performing similarity evaluation on the input tile area image and a template in a matched template library, if the evaluated similarity is smaller than a given threshold value, indicating that the current tile area image is not in the template library, and adding the current tile area image and corresponding texture features thereof into the template library (similarly, adding an image of a newly input tile area image under 4 angles (0 degree is reserved, 90 degrees is rotated, 180 degrees is rotated, and 270 degrees is rotated) and corresponding texture features (4 feature vectors) into the template library); and if the evaluated similarity is above a given threshold, indicating that the same plate type exists in the template library of the current tile area image, and taking the plate number corresponding to the matched template as the plate number corresponding to the current tile area image.
Generally speaking, tiles have a more general classification, such as fixed layout, lined tiles, and infinite lined tiles. The results of the plate can be defined for different types (which can be achieved by defining the number of templates in the template library). In the corresponding plate dividing result, the fixed plate surface can limit and output a plate type; a continuous grain tile with n (n is a positive integer) patterns (layouts) can be limited to outputting up to n layouts.
In the actual plate dividing process, because tiles with the same plate type are produced, deviation is inevitable, and machine plate dividing can be regarded as a new plate type for the tiles in the condition. On the premise of limiting the upper limit of the number of output versions, if the similarity between the texture features of the newly input tile region image and the texture features of each template in the template library is smaller than a given threshold value, and the version number corresponding to the template in the template library has reached the upper limit of the number of output versions, that is, the newly input tile region image cannot be regarded as a new plate, then, regarding the input tile region image, the version number corresponding to the position where the response value is maximum when the input tile region image and the template library are subjected to the relevant filtering operation of space-time constraint in the feature space is taken as the version number corresponding to the newly input tile region image.
A result output unit configured to: the corresponding tile version number is output (or the rotation angle of the tile relative to the form is also output to facilitate the arrangement during sorting).
The invention is not limited to the foregoing embodiments. The invention extends to any novel feature or any novel combination of features disclosed in this specification and any novel method or process steps or any novel combination of features disclosed.
Claims (10)
1. A tile texture automatic classification method is characterized by comprising the following steps:
A. preprocessing the collected tile image to obtain a corresponding tile area image;
B. the tile area image is subjected to plate division, and the method comprises the following steps:
adding the first tile area image and the corresponding texture features thereof into a template library, wherein the steps of keeping the tile area image at four angles of 0 degree, rotating 90 degrees, rotating 180 degrees and rotating 270 degrees, and respectively extracting the corresponding texture features;
extracting texture features of the subsequent tile area images, performing relevant filtering operation of space-time constraint on the extracted texture features in a feature space of the template library, and searching a position with the maximum response value; and then carrying out similarity evaluation on the detected tile area image and the template in the matched template library, if the evaluated similarity is above a given threshold value, taking the version number corresponding to the matched template as the version number corresponding to the current tile area image, and otherwise, adding the current tile area image and the corresponding texture features thereof into the template library.
2. The automatic tile texture classification method according to claim 1, characterized in that the preprocessing comprises the processes of grey value correction, foreground segmentation and geometric transformation.
3. The automatic tile texture classification method according to claim 2, wherein the gray value correction process comprises:
and carrying out gray value correction on the whole tile image according to the three channel values in the calibration area in the tile image.
4. The method for automatically classifying tile texture according to claim 3, wherein the calibration area is an area in the tile image corresponding to a calibration object disposed near the tile in the tile image capturing step.
5. The automatic tile texture classification method according to any one of claims 1 to 4, wherein an upper limit of the number of templates is set in the template library;
and if the similarity between the texture features of the current tile area image and the texture features of the templates in the template library is smaller than a given threshold value and the templates in the template library reach the upper limit of the number, regarding the current tile area image, taking the version number corresponding to the maximum position of the response value when the current tile area image and the template library perform the relevant filtering operation of space-time constraint as the version number corresponding to the current tile area image.
6. An automatic classification device for ceramic tile textures is characterized by comprising an image preprocessing unit, a ceramic tile layout dividing unit and a result output unit which are sequentially connected; wherein:
the image pre-processing unit is configured to: preprocessing an input tile image and outputting a tile area image;
the tile split unit is configured to: adding the first tile area image and the corresponding texture features thereof into a template library, wherein the steps of keeping the tile area image at four angles of 0 degree, rotating 90 degrees, rotating 180 degrees and rotating 270 degrees, and respectively extracting the corresponding texture features; extracting texture features of the subsequent tile area images, performing relevant filtering operation of space-time constraint on the extracted texture features in a feature space of the template library, and searching a position with the maximum response value; carrying out similarity evaluation on the detected tile area image and a template in a matched template library, if the evaluated similarity is above a given threshold value, taking a version number corresponding to the matched template as a version number corresponding to the current tile area image, and otherwise, adding the current tile area image and the corresponding texture features thereof into the template library;
the result output unit is configured to: and outputting the corresponding version number of the ceramic tile.
7. The automatic tile texture classification device according to claim 6, wherein the preprocessing of the input tile image by the image preprocessing unit includes processes of gray value correction, foreground segmentation, and geometric transformation.
8. The automatic tile texture sorting apparatus according to claim 7, wherein the gray value correction process comprises: and carrying out gray value correction on the whole tile image according to the three channel values in the calibration area in the tile image.
9. The automatic tile texture sorting device according to claim 8, wherein the calibration area is in a tile image capturing stage, and the calibration object disposed near the tile corresponds to an area in the tile image.
10. An automatic tile texture sorting device according to any one of claims 6 to 9, wherein the template library is provided with an upper limit on the number of templates; and if the similarity between the texture features of the current tile area image and the texture features of the templates in the template library is smaller than a given threshold value and the templates in the template library reach the upper limit of the number, taking the version number corresponding to the maximum position of the response value when the space-time constrained related filtering operation is performed in the feature space of the current tile area image and the template library as the version number corresponding to the current tile area image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011589821.5A CN112633393B (en) | 2020-12-29 | 2020-12-29 | Automatic classification method and device for ceramic tile textures |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011589821.5A CN112633393B (en) | 2020-12-29 | 2020-12-29 | Automatic classification method and device for ceramic tile textures |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112633393A true CN112633393A (en) | 2021-04-09 |
| CN112633393B CN112633393B (en) | 2022-11-15 |
Family
ID=75285924
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011589821.5A Active CN112633393B (en) | 2020-12-29 | 2020-12-29 | Automatic classification method and device for ceramic tile textures |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112633393B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114820841A (en) * | 2022-03-31 | 2022-07-29 | 广东东唯新材料有限公司 | Image processing method and device for random continuous lines of tile and rock plate product layout |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003031956A1 (en) * | 2001-10-05 | 2003-04-17 | Millennium Venture Holdings Ltd | System and method for classifying workpieces according to tonal variations |
| CN104772817A (en) * | 2015-04-16 | 2015-07-15 | 佛山市简一陶瓷有限公司 | Distributing system for producing full-body marble tile |
| CN105718947A (en) * | 2016-01-21 | 2016-06-29 | 吉林大学 | Lung cancer image fine classification method based on fusion of LBP and wavelet moment features |
| CN107730500A (en) * | 2017-11-01 | 2018-02-23 | 广东工业大学 | A kind of ceramic tile texture detection, system, device and readable storage medium storing program for executing |
| CN108038853A (en) * | 2017-12-18 | 2018-05-15 | 浙江工业大学 | A kind of ceramic tile surface defect identification method based on convolutional neural networks and Active Learning |
| CN109051106A (en) * | 2018-06-28 | 2018-12-21 | 广东科达洁能股份有限公司 | A kind of ceramic tile potted flower color packing method and system |
| CN110163244A (en) * | 2019-04-04 | 2019-08-23 | 睿视智觉（厦门）科技有限公司 | A kind of ceramic tile texture classifying method and device |
-
2020
- 2020-12-29 CN CN202011589821.5A patent/CN112633393B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003031956A1 (en) * | 2001-10-05 | 2003-04-17 | Millennium Venture Holdings Ltd | System and method for classifying workpieces according to tonal variations |
| CN104772817A (en) * | 2015-04-16 | 2015-07-15 | 佛山市简一陶瓷有限公司 | Distributing system for producing full-body marble tile |
| CN105718947A (en) * | 2016-01-21 | 2016-06-29 | 吉林大学 | Lung cancer image fine classification method based on fusion of LBP and wavelet moment features |
| CN107730500A (en) * | 2017-11-01 | 2018-02-23 | 广东工业大学 | A kind of ceramic tile texture detection, system, device and readable storage medium storing program for executing |
| CN108038853A (en) * | 2017-12-18 | 2018-05-15 | 浙江工业大学 | A kind of ceramic tile surface defect identification method based on convolutional neural networks and Active Learning |
| CN109051106A (en) * | 2018-06-28 | 2018-12-21 | 广东科达洁能股份有限公司 | A kind of ceramic tile potted flower color packing method and system |
| CN110163244A (en) * | 2019-04-04 | 2019-08-23 | 睿视智觉（厦门）科技有限公司 | A kind of ceramic tile texture classifying method and device |
Non-Patent Citations (4)
| Title |
|---|
| JIAO ZHANG 等: "An improved BP algorithm for classification of architectural ceramic material", 《2010 INTERNATIONAL CONFERENCE ON COMPUTER DESIGN AND APPLICATIONS》 * |
| 卞玉丽: "基于机器视觉的瓷砖的在线分类识别的研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
| 彭琼: "基于纹理特征的瓷砖图像分类", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
| 陈瑶等: "基于机器视觉的瓷砖自动分类系统设计", 《武汉工业学院学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114820841A (en) * | 2022-03-31 | 2022-07-29 | 广东东唯新材料有限公司 | Image processing method and device for random continuous lines of tile and rock plate product layout |
| CN114820841B (en) * | 2022-03-31 | 2023-03-24 | 广东东唯新材料有限公司 | Image processing method and device for random continuous lines of tile and rock plate product layout |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112633393B (en) | 2022-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104866862B (en) | A kind of method of belt steel surface area-type defect recognition classification | |
| CN111474184A (en) | AOI character defect detection method and device based on industrial machine vision | |
| CN102542275B (en) | Automatic identification method for identification photo background and system thereof | |
| CN106251353A (en) | Weak texture workpiece and the recognition detection method and system of three-dimensional pose thereof | |
| CN112330628A (en) | Metal workpiece surface defect image detection method | |
| CN104463199A (en) | Rock fragment size classification method based on multiple features and segmentation recorrection | |
| CN102426649A (en) | Simple steel seal digital automatic identification method with high accuracy rate | |
| CN110443791B (en) | Workpiece detection method and device based on deep learning network | |
| CN106846313A (en) | Surface Flaw Detection method and apparatus | |
| CN115078365A (en) | Soft package printing quality defect detection method | |
| CN114910480A (en) | Wafer surface defect detection method based on machine vision | |
| CN107871137A (en) | A kind of material matching process based on image recognition | |
| CN112633393B (en) | Automatic classification method and device for ceramic tile textures | |
| CN112304957A (en) | Machine vision-based intelligent detection method and system for appearance defects | |
| CN111667475A (en) | Machine vision-based Chinese date grading detection method | |
| CN113252103A (en) | Method for calculating volume and mass of material pile based on MATLAB image recognition technology | |
| CN110728286B (en) | Abrasive belt grinding material removal rate identification method based on spark image | |
| CN115830359A (en) | Workpiece identification and counting method based on target detection and template matching in complex scene | |
| CN112798108B (en) | Ceramic tile self-adaptive color separation method and device | |
| Kuo et al. | Automatic pattern recognition and color separation of embroidery fabrics | |
| CN107516309B (en) | Visual inspection method for defects of printing panel | |
| CN113470125B (en) | Method and system for carrying out color matching on patterns | |
| CN114264607B (en) | Machine vision-based tile color difference online detection system and method | |
| CN113705487B (en) | Precision workpiece identification and technological parameter association system and identification method | |
| CN113591923A (en) | Engine rocker arm part classification method based on image feature extraction and template matching |
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 |