CN-121982429-A - Intelligent color selection method and system for plastic sheet based on image processing

CN121982429ACN 121982429 ACN121982429 ACN 121982429ACN-121982429-A

Abstract

The invention relates to the technical field of image processing and discloses an intelligent color selection method and system for a plastic sheet based on image processing, wherein the method comprises the steps of mapping an original color image to a preset brightness-color opposite color space so as to achieve complete decoupling of brightness information and color information in the plastic sheet and obtain an initial color image; the method comprises the steps of filtering an initial color image to obtain a standardized image, carrying out binarization on the standardized image to obtain a binary image, fusing edge continuity constraint conditions of geometric features of key edges to obtain an area identification image of a plastic sheet, calculating high-order color moment features of a corresponding pixel set, carrying out similarity measurement on the high-order color moment features to obtain a complete classification result set, analyzing the complete classification result set, and combining a real-time quality control strategy of the plastic sheet to obtain an optimal sorting control instruction sequence to realize intelligent color selection control.

Inventors

WANG BAOWEN
ZHU BO
Mu Leichao
WANG GANG
XU HAIBO
HAN KAIDI
MU CHENGLEI
ZHANG PU
XIA XUESHI

Assignees

陕西东方保发环保科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260401

Claims (10)

1. An intelligent color selection method for plastic sheets based on image processing is characterized by comprising the following steps: S1, mapping an original color image to a preset brightness-color opposite color space so as to achieve complete decoupling of brightness information and color information in a plastic sheet and obtain an original color image of the plastic sheet; s2, carrying out multi-scale morphological filtering processing on brightness components and color opposite components in the initial color image, and reserving key edge geometric characteristics of the plastic sheet to obtain a standardized image of the plastic sheet; s3, carrying out binarization processing on the standardized image based on the color numerical distribution characteristics of the standardized image to obtain a binary image of the plastic sheet; S4, fusing edge continuity constraint conditions of the key edge geometric features to analyze component connectivity in the binary image so as to obtain an area identification image of the plastic sheet; S5, identifying a corresponding pixel set of the region identification image on the binary image, calculating high-order color moment characteristics of the corresponding pixel set, and carrying out the similarity measurement of the mahalanobis distance on the high-order color moment characteristics to obtain a complete classification result set of the plastic sheet; S6, carrying out decision tree logic analysis on the complete classification result set, and combining with a real-time quality control strategy of the plastic sheet to obtain an optimal sorting control instruction sequence of the plastic sheet so as to realize intelligent color sorting control of the plastic sheet.
2. The intelligent color selection method for plastic sheets based on image processing according to claim 1, wherein the mapping the original color image to a preset brightness-color contrast color space to achieve complete decoupling of brightness information and color information in the plastic sheets, to obtain an initial color image of the plastic sheets, comprises: Acquiring an original color image of a plastic sheet; Channel separation is carried out on the original color image to obtain a red component image, a green component image and a blue component image of the original color image; Eliminating color shift of the red component image, the green component image and the blue component image caused by uneven illumination to obtain balanced red component, balanced green component and balanced blue component of the original color image; Mapping the balanced red component, the balanced green component and the balanced blue component to a preset brightness-color opposite color space to obtain a brightness component, a first color opposite component and a second color opposite component of the original color image; And carrying out dynamic range normalization on the brightness component, the first color opposite component and the second color opposite component to obtain an initial color image of the plastic sheet.
3. The intelligent color selection method for plastic sheets based on image processing according to claim 1, wherein the steps of performing multi-scale morphological filtering processing on the brightness component and the color opposite component in the initial color image, and retaining the key edge geometric features of the plastic sheets to obtain the standardized image of the plastic sheets include: Determining a local texture complexity index of the brightness component according to the association relation between the brightness component and the color opposite component in the initial color image; based on the local texture complexity index, geometrical characteristics of different granularity levels in the plastic sheet are adaptively arranged to obtain a multi-scale structural element sequence of the plastic sheet; Applying the brightness component to the multi-scale structure element sequence for morphological opening and closing analysis to obtain a brightness smooth component of the multi-scale structure element sequence; Mapping the color opposite component to the multi-scale structure element sequence for morphological gradient evaluation to obtain a color smooth component of the multi-scale structure element sequence; And fusing the brightness smooth component and the color smooth component, and superposing the key edge geometric features of the plastic sheet to obtain the standardized image of the plastic sheet.
4. A method of intelligent color selection of plastic sheeting based on image processing as set forth in claim 3 wherein said fusing said brightness smoothing component with said color smoothing component and superimposing key edge geometries of said plastic sheeting to obtain a standardized image of said plastic sheeting comprises: Synchronously extracting brightness intensity characteristics and color direction characteristics from the brightness smooth components and the color smooth components; Calculating an edge weighting coefficient of the plastic sheet according to the brightness intensity characteristic and the color direction characteristic, wherein the calculation formula of the edge weighting coefficient is as follows: ; In the formula, For the edge weighting coefficients to be used, For the brightness intensity characteristics to be described, For the luminance direction characteristics of the luminance smoothing component, As a feature of the color direction of the light, Is the circumference ratio; based on the edge weighting coefficient, carrying out self-adaptive spectrum reconstruction on the brightness smooth component and the color smooth component to obtain a primary fusion image of the plastic sheet; And carrying out edge connectivity optimization on the primary fusion image based on the key edge geometrical characteristics to obtain a standardized image of the plastic sheet.
5. The intelligent color selection method for plastic sheets based on image processing according to claim 1, wherein the binarizing the standardized image based on the color numerical distribution characteristics of the standardized image to obtain the binary image of the plastic sheet comprises: Analyzing the color numerical distribution characteristics of the standardized image to obtain a main color aggregation area of the standardized image; in the main color aggregation area, taking a high threshold area as a main characteristic and a low threshold area as a detail characteristic to construct a self-adaptive threshold curved surface of the plastic sheet; comparing the color characteristic values of the pixel points in the standardized image with the self-adaptive threshold curved surface point by point to obtain a preliminary binary segmentation result of the standardized image; And carrying out space consistency constraint on the preliminary binary segmentation result to eliminate isolated noise points in the preliminary binary segmentation result and fill local holes so as to obtain a binary image of the plastic sheet.
6. The image processing-based intelligent color selection method for plastic sheets according to claim 1, wherein the fusing the edge continuity constraint condition of the key edge geometric feature to analyze the component connectivity in the binary image to obtain the area identification image of the plastic sheet comprises: determining structural connectivity decision criteria of the plastic sheet according to edge continuity constraint conditions in the key edge geometric features; Based on the structural connectivity judgment criterion, performing connectivity analysis on adjacent components in the binary image to obtain an initial connected region of the binary image; Verifying the integrity of the edge closure degree in the initial connected region to obtain a connected region of the binary image; and taking the connected region as an independent identifier to identify the plastic sheet, and obtaining a region identification image of the plastic sheet.
7. An intelligent color selection method for plastic sheets based on image processing according to claim 1, wherein said identifying the corresponding pixel set of the area identification image on the binary image comprises: Identifying boundary contours in the region identification image to obtain a minimum circumscribed rectangle of the region identification image; A dense sampling strategy is adopted for the slender area of the minimum circumscribed rectangle, and a sparse sampling strategy is adopted for the near-circular area; integrating the dense sampling strategy and the sparse sampling strategy into a spatial sampling strategy of the region identification image; according to the spatial sampling strategy, combining the distribution density of sampling points on the binary image with the shape complexity between the connected areas to construct an adaptive sampling grid of the binary image; screening out representative pixel points of the region identification image in the binary image in the self-adaptive sampling grid; and performing spatial cluster analysis on the representative pixel points to obtain a corresponding pixel set of the region identification image.
8. The image processing-based intelligent color selection method for plastic sheets according to claim 7, wherein said calculating the higher-order color moment characteristics of the corresponding pixel sets comprises: respectively extracting the bias state characteristics, the kurtosis characteristics and the projection characteristics of the corresponding pixel set in the preset brightness-color opposite color space; constructing a multi-scale color moment descriptor of the corresponding pixel set according to the bias state characteristic, the kurtosis characteristic and the projection characteristic; Performing feature orthogonalization processing on the multi-scale color moment descriptors to obtain standardized moment descriptors of the corresponding pixel sets; dynamically coupling the standardized moment descriptor with the shape feature of the connected region to obtain a high-order color moment feature of the corresponding pixel set, wherein the calculation formula of the high-order color moment feature is as follows: ; In the formula, For the high-order color moment feature, For the normalized moment descriptor in question, Is a preset coupling coefficient of the shape characteristics, As the feature vector of the shape feature, Reference values are referenced for vectors of the shape features.
9. The intelligent color selection method for plastic sheets based on image processing according to claim 8, wherein the performing a similarity measurement of mahalanobis distance on the high-order color moment features to obtain a complete classification result set of the plastic sheets comprises: Determining the reference color moment characteristics of the plastic sheet according to the reference characteristic template library of the plastic sheet; performing mahalanobis distance measurement on the high-order color moment features and the reference color moment features to obtain an initial classification label of the plastic sheet; Verifying the consistency of the spatial distribution and the color continuity in the initial classification label to obtain an optimized classification label of the plastic sheet; and carrying out category identification distribution on the connected areas of the plastic sheets according to the optimized classification labels to obtain a complete classification result set of the plastic sheets.
10. An intelligent color selection system for plastic sheets based on image processing, for implementing the intelligent color selection method for plastic sheets based on image processing as claimed in claim 1, the system comprising: The initial image module is used for mapping the original color image to a preset brightness-color opposite color space so as to achieve complete decoupling of brightness information and color information in the plastic sheet and obtain an initial color image of the plastic sheet; the standard image module is used for carrying out multi-scale morphological filtering processing on the brightness component and the color opposite component in the initial color image, and retaining the key edge geometric characteristics of the plastic sheet to obtain a standardized image of the plastic sheet; the binary image module is used for carrying out binarization processing on the standardized image based on the color numerical distribution characteristics of the standardized image so as to obtain a binary image of the plastic sheet; the region image module is used for fusing the edge continuity constraint conditions of the key edge geometric features so as to analyze the connectivity of components in the binary image and obtain a region identification image of the plastic sheet; the classification result module is used for identifying a corresponding pixel set of the region identification image on the binary image, calculating high-order color moment characteristics of the corresponding pixel set, and carrying out the similarity measurement of the mahalanobis distance on the high-order color moment characteristics to obtain a complete classification result set of the plastic sheet; And the intelligent sorting module is used for carrying out decision tree logic analysis on the complete sorting result set, and combining with the real-time quality control strategy of the plastic sheet to obtain the optimal sorting control instruction sequence of the plastic sheet so as to realize intelligent color sorting control of the plastic sheet.

Description

Intelligent color selection method and system for plastic sheet based on image processing Technical Field The invention relates to the technical field of image processing, in particular to an intelligent color selection method and system for plastic sheets based on image processing. Background The prior art has obvious defects in an image preprocessing link of intelligent color selection of a plastic sheet, does not map an original color image to a special color space to realize complete decoupling of brightness and color information, only adopts a conventional color space to process the image, so that color cast caused by uneven illumination cannot be effectively eliminated, brightness and color information are mutually interfered, does not carry out multi-scale morphological filtering on brightness components and color opposite components, only adopts single-scale filtering or simple smoothing processing, is difficult to consider image noise reduction and retention of key edge geometric characteristics, easily causes image detail loss or noise residue, has poor quality of a generated standardized image, and cannot provide a precise image foundation for subsequent color selection. The defects of the plastic sheet classification and separation control links in the prior art are prominent. The method comprises the steps of determining the continuity constraint condition analysis assembly connectivity of the geometrical characteristics of the non-fusion key edges, identifying the plastic sheet areas by means of a simple threshold value or an area growth method, enabling the area identification to be inaccurate, enabling the situation that adjacent plastic sheets are adhered or single plastic sheets are split to easily occur, calculating the high-order color moment characteristics, carrying out similarity measurement by means of the mahalanobis distance, classifying the plastic sheets with similar colors only based on the simple color characteristics or gray level characteristics, and enabling the classification result to be difficult to accurately distinguish, wherein the accuracy of the classification result is low, carrying out decision tree logic analysis on the classification result without combining a real-time quality control strategy, generating a classification instruction only according to fixed classification labels, and enabling the classification instruction to be lack of flexibility and adaptability, so that intelligent color selection requirements under different quality requirements can not be met. Disclosure of Invention The invention provides an intelligent color selection method and system for plastic sheets based on image processing, which are used for solving the problems in the background technology. In order to achieve the above purpose, the invention provides an intelligent color selection method for plastic sheets based on image processing, which comprises the following steps: S1, mapping an original color image to a preset brightness-color opposite color space so as to achieve complete decoupling of brightness information and color information in a plastic sheet and obtain an original color image of the plastic sheet; s2, carrying out multi-scale morphological filtering processing on brightness components and color opposite components in the initial color image, and reserving key edge geometric characteristics of the plastic sheet to obtain a standardized image of the plastic sheet; s3, carrying out binarization processing on the standardized image based on the color numerical distribution characteristics of the standardized image to obtain a binary image of the plastic sheet; S4, fusing edge continuity constraint conditions of the key edge geometric features to analyze component connectivity in the binary image so as to obtain an area identification image of the plastic sheet; S5, identifying a corresponding pixel set of the region identification image on the binary image, calculating high-order color moment characteristics of the corresponding pixel set, and carrying out the similarity measurement of the mahalanobis distance on the high-order color moment characteristics to obtain a complete classification result set of the plastic sheet; S6, carrying out decision tree logic analysis on the complete classification result set, and combining with a real-time quality control strategy of the plastic sheet to obtain an optimal sorting control instruction sequence of the plastic sheet so as to realize intelligent color sorting control of the plastic sheet. In a preferred embodiment, the mapping the original color image to a preset brightness-color contrast color space to achieve complete decoupling of brightness information and color information in the plastic sheet, to obtain an initial color image of the plastic sheet, includes: Acquiring an original color image of a plastic sheet; Channel separation is carried out on the original color image to obtain a red component image, a green com