Search

CN-121607346-B - Plastic package sorting method, system, electronic device and storage medium

CN121607346BCN 121607346 BCN121607346 BCN 121607346BCN-121607346-B

Abstract

The application provides a plastic package sorting method, a plastic package sorting system, electronic equipment and a storage medium, wherein the method comprises the steps of carrying out spectral feature analysis on hyperspectral data and determining the material type of plastic; the method comprises the steps of carrying out morphological analysis on image model data, carrying out multi-mode matching verification on a shape analysis result and space distribution of hyperspectral data, determining geometric attributes, generating a correlation model based on a specified sorting strategy database, material types and geometric attributes, extracting pixel-level contour sets of objects according to the image model data, mapping the pixel-level contour sets to corresponding air injection equipment parameters to obtain a target air hole matrix, and determining air valve control instructions of the air injection equipment according to the air injection equipment parameters, the correlation model, the target air hole matrix and conveying speed. Through the double combination of hyperspectral data and image model data, the problem that single vision technology cannot distinguish the defects of same-color heterogeneous plastics (such as transparent PET and PVC) is solved.

Inventors

  • LIU XIAOJUN
  • LI XIZHUO

Assignees

  • 广州九爪智能科技有限公司

Dates

Publication Date
20260512
Application Date
20260202

Claims (9)

  1. 1. A method of sorting plastic packages, comprising: acquiring parameters of air injection equipment, hyperspectral data, image model data and conveying speed, wherein the hyperspectral data are spectral data of plastic packages acquired by utilizing a hyperspectral imager when conveyed on a belt at the conveying speed, and the image model data are image data of the plastic packages acquired by utilizing the hyperspectral imager when conveyed on the belt at the conveying speed; performing spectral feature analysis on the hyperspectral data, extracting material feature absorption peaks, calculating spectral similarity, and determining the material type of plastics; performing morphological analysis on the image model data, performing multi-mode matching verification on a shape analysis result and the spatial distribution of the hyperspectral data, and determining geometric attributes; Generating a correlation model based on a specified sorting strategy database, the material type and the geometric attribute, wherein the specified sorting strategy database is used for storing a mapping relation between the material type and a basic jet intensity value of jet equipment; Extracting a pixel level contour set of an object according to the image model data and mapping the pixel level contour set to parameters corresponding to air injection equipment to obtain a target air hole matrix; Determining an air valve control instruction of the air injection equipment according to the air injection equipment parameters, the correlation model, the target air hole matrix and the conveying speed; wherein the morphological analysis is performed on the image model data, the shape analysis result and the spatial distribution of the hyperspectral data are subjected to multi-modal matching verification, and the determining of the geometric attribute comprises: Carrying out connected domain analysis on the image model data to obtain a pixel connected region of a target object; Constructing a minimum circumscribed rectangle for the pixel connected region to obtain the number of long-axis pixels and the number of short-axis pixels; calculating the ratio of the number of the long-axis pixels to the number of the short-axis pixels to obtain an aspect ratio characteristic value; Calculating the ratio of the area of the pixel communication area to the minimum circumscribed rectangular area to obtain a filling rate characteristic value; comparing the aspect ratio characteristic value with the filling rate characteristic value input form classification rule base; Generating geometrical attributes representing thin films when the aspect ratio characteristic value falls into a first preset interval and the filling rate characteristic value is smaller than a preset threshold value; Generating geometrical attributes representing bottle flakes when the aspect ratio characteristic value falls into a second preset interval and the filling rate characteristic value is larger than a preset threshold value; Establishing a pixel-level space mapping relation between hyperspectral data and image model data through coordinate conversion; Determining an effective spectral response area of the same object in the hyperspectral data based on the pixel-level spatial mapping relation; Calculating a pixel communication area in the image model data to obtain geometric centroid coordinates; calculating an effective spectral response area in the hyperspectral data to obtain a spectrum centroid coordinate; calculating the geometric centroid coordinates and the spectrum centroid coordinates to obtain Euclidean distances; judging whether the Euclidean distance is smaller than a preset space deviation threshold value or not; And under the condition that the Euclidean distance is not smaller than a preset space deviation threshold value, judging that the currently identified object is an optical ghost or noise point, and stopping generating the geometric attribute of the plastic.
  2. 2. The method of claim 1, wherein generating the association model based on the specified sort policy database, the material type, and the geometric attribute comprises: searching the material type based on a specified sorting strategy database to obtain a basic jet intensity value; Determining jet correction weights according to the geometric features; And carrying out association storage on the basic jet intensity value and the jet correction weight to generate the association model.
  3. 3. The method according to claim 2, wherein the air injection device parameters include one-dimensional coordinate axis coordinates of an arrangement direction of the air injection devices, the extracting a pixel level contour set of the object according to the image model data and mapping the pixel level contour set to the air injection device parameters, and obtaining the target air hole matrix includes: Positioning vertex coordinates of an object in the image model data by using a specified convolution algorithm to obtain a region of interest; performing graying, binarization and edge detection algorithms on the region of interest to obtain a pixel level contour set of the object; mapping the pixel level outline set to a one-dimensional coordinate axis corresponding to the arrangement direction of the air injection equipment, identifying a coordinate interval covered by the outline, and generating a target air hole matrix consisting of air valve index numbers.
  4. 4. The method of claim 3, wherein the jet device parameters include a mounting distance of a gas valve, and wherein determining the gas valve control command for the jet device based on the jet device parameters, the correlation model, the target vent matrix, and the transport speed comprises: Traversing each air valve index number in the target air hole matrix, and obtaining target opening duration of each air valve according to the basic air injection intensity value and the air injection correction weight in the correlation model; Calculating the opening trigger time stamp of each air valve according to the conveying speed and the installation distance of the air valve, and generating the opening trigger time stamp of each air valve; and determining a gas valve control instruction of the gas injection equipment according to the opening trigger time stamp, the gas valve index number and the target opening duration of each gas valve.
  5. 5. The method of claim 4, wherein the performing spectral feature analysis on the hyperspectral data, extracting material feature absorption peaks and calculating spectral similarity, determining the material type of the plastic comprises: Performing smoothing and differential operation on the hyperspectral data to generate a derivative spectrum curve and extracting a characteristic peak value corresponding to a preset wavelength from the derivative spectrum curve; calculating similarity coefficients of the derivative spectral curve and each standard curve in a standard material spectral library by utilizing a spectral angle drawing algorithm or an Euclidean distance algorithm to obtain the similarity coefficients; Determining a current dynamic classification threshold based on the conveying speed and a preset speed threshold anti-correlation mapping function; And comparing the similarity coefficient with the dynamic classification threshold, and if the similarity coefficient is larger than or equal to the dynamic classification threshold, taking the standard material type corresponding to the similarity coefficient as the material type of the plastic.
  6. 6. The method as recited in claim 1, further comprising: performing principal component analysis processing on the hyperspectral data to obtain principal component characteristics; inputting the principal component characteristics into a pre-trained random forest classification model to obtain a classification result representing whether the materials are stacked or not.
  7. 7. A plastic package sorting system, comprising: The first acquisition module is used for acquiring parameters of the air injection equipment, hyperspectral data, image model data and conveying speed, wherein the hyperspectral data are spectral data acquired by utilizing a hyperspectral imager when the plastic package is conveyed on a belt at the conveying speed, and the image model data are image data acquired by utilizing the hyperspectral imager when the plastic package is conveyed on the belt at the conveying speed; The first determining module is used for carrying out spectral feature analysis on the hyperspectral data, extracting material feature absorption peaks, calculating spectral similarity and determining the material type of plastics; The second determining module is used for carrying out morphological analysis on the image model data and carrying out multi-mode matching verification on a shape analysis result and the spatial distribution of the hyperspectral data to determine geometric attributes; the first generation module is used for generating a correlation model based on a specified sorting strategy database, the material type and the geometric attribute, wherein the specified sorting strategy database is used for storing a mapping relation between the material type and a basic jet intensity value of jet equipment; The first obtaining module is used for extracting a pixel level outline set of the object according to the image model data and mapping the pixel level outline set to parameters corresponding to the air injection equipment to obtain a target air hole matrix; the third determining module is used for determining an air valve control instruction of the air injection equipment according to the air injection equipment parameters, the association model, the target air hole matrix and the conveying speed; wherein the morphological analysis is performed on the image model data, the shape analysis result and the spatial distribution of the hyperspectral data are subjected to multi-modal matching verification, and the determining of the geometric attribute comprises: Carrying out connected domain analysis on the image model data to obtain a pixel connected region of a target object; Constructing a minimum circumscribed rectangle for the pixel connected region to obtain the number of long-axis pixels and the number of short-axis pixels; calculating the ratio of the number of the long-axis pixels to the number of the short-axis pixels to obtain an aspect ratio characteristic value; Calculating the ratio of the area of the pixel communication area to the minimum circumscribed rectangular area to obtain a filling rate characteristic value; comparing the aspect ratio characteristic value with the filling rate characteristic value input form classification rule base; Generating geometrical attributes representing thin films when the aspect ratio characteristic value falls into a first preset interval and the filling rate characteristic value is smaller than a preset threshold value; Generating geometrical attributes representing bottle flakes when the aspect ratio characteristic value falls into a second preset interval and the filling rate characteristic value is larger than a preset threshold value; Establishing a pixel-level space mapping relation between hyperspectral data and image model data through coordinate conversion; Determining an effective spectral response area of the same object in the hyperspectral data based on the pixel-level spatial mapping relation; Calculating a pixel communication area in the image model data to obtain geometric centroid coordinates; calculating an effective spectral response area in the hyperspectral data to obtain a spectrum centroid coordinate; calculating the geometric centroid coordinates and the spectrum centroid coordinates to obtain Euclidean distances; judging whether the Euclidean distance is smaller than a preset space deviation threshold value or not; And under the condition that the Euclidean distance is not smaller than a preset space deviation threshold value, judging that the currently identified object is an optical ghost or noise point, and stopping generating the geometric attribute of the plastic.
  8. 8. An electronic device, comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.
  9. 9. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein computer instructions which, when executed by a processor, implement the method of any of claims 1-6.

Description

Plastic package sorting method, system, electronic device and storage medium Technical Field The application relates to the technical field of plastic package sorting methods, in particular to a plastic package sorting method, a plastic package sorting system, electronic equipment and a storage medium. Background Currently, the sorting process of plastic packages relies mainly on two key techniques, gravimetric sorting and visual sorting. The gravimetric sorting technique achieves effective separation by accurately measuring the weight differences of different plastic materials, which is particularly efficient when dealing with large batches of the same type of plastic. The vision sorting technology utilizes a high-resolution camera and an advanced image processing algorithm to sort by identifying the surface characteristics of plastics such as color, shape, texture and the like, so that more complex mixed materials can be processed. The two technologies have the advantages of being often combined in practical application so as to improve the sorting accuracy and efficiency and meet the requirements of environmental protection and resource regeneration. When the existing plastic packaging is sorted, the plastic material types are difficult to accurately distinguish due to weight sorting and visual sorting technologies, so that the sorting effect is poor. Disclosure of Invention The embodiment of the application provides a plastic package sorting method, a plastic package sorting system, electronic equipment and a storage medium, which are used for solving the problems of the related technology, and the technical scheme is as follows: In a first aspect, an embodiment of the present application provides a plastic package sorting method, including: Acquiring parameters of air injection equipment, hyperspectral data, image model data and conveying speed, wherein the hyperspectral data are spectral data acquired by utilizing a hyperspectral imager when plastic packages are conveyed on a belt at the conveying speed, and the image model data are image data acquired by utilizing the hyperspectral imager when the plastic packages are conveyed on the belt at the conveying speed; Carrying out spectral feature analysis on the hyperspectral data, extracting material feature absorption peaks, calculating spectral similarity, and determining the material type of plastics; Carrying out morphological analysis on the image model data, carrying out multi-mode matching verification on the shape analysis result and the spatial distribution of the hyperspectral data, and determining geometric attributes; Generating a correlation model based on a specified sorting strategy database, material types and geometric attributes, wherein the specified sorting strategy database is used for storing a mapping relation between the material types and basic jet intensity values of jet equipment; extracting a pixel level contour set of the object according to the image model data and mapping the pixel level contour set to parameters corresponding to the air injection equipment to obtain a target air hole matrix; And determining an air valve control instruction of the air injection equipment according to the air injection equipment parameters, the correlation model, the target air hole matrix and the conveying speed. In one embodiment of the application, generating the association model based on the specified sort strategy database, material type, and geometric attributes includes: retrieving the material type based on a specified sorting strategy database to obtain a basic jet intensity value; determining jet correction weights according to the geometric features; And storing the basic jet intensity value and the jet correction weight in a correlated way to generate a correlation model. In one embodiment of the present application, the air injection device parameters include one-dimensional coordinate axis coordinates of an arrangement direction of the air injection device, extracting a pixel level contour set of the object according to the image model data and mapping the pixel level contour set to the air injection device parameters, and obtaining the target air hole matrix includes: Positioning vertex coordinates of an object in the image model data by using a specified convolution algorithm to obtain an interested region; Gray scale, binarization and edge detection algorithm are carried out on the region of interest to obtain a pixel level outline set of the object; Mapping the pixel-level outline set onto a one-dimensional coordinate axis corresponding to the arrangement direction of the air injection equipment, identifying a coordinate interval covered by the outline, and generating a target air hole matrix consisting of air valve index numbers. In one embodiment of the present application, the air injection device parameters include an installation distance of an air valve, and determining an air valve control instruction of the air inj