CPC B07C 5/3422 (2013.01) [B02C 23/10 (2013.01); B07C 5/02 (2013.01); B07C 2501/0054 (2013.01); H04N 7/18 (2013.01)] | 11 Claims |
1. A sorting-based garbage classification device, comprising:
a conveying assembly (1); and a sorting assembly (2), an identifying assembly (3) and a processing assembly (4) sequentially installed above the conveying assembly (1) from left to right in that order;
wherein the sorting assembly (2) comprises a fixing frame (5) and lifting cylinders (6), a lower end of the fixing frame (5) is connected to the conveying assembly (1), and lifting cylinders (6) are installed on the fixing frame (5); the lifting cylinders (6) are distributed in a plurality of groups, expandable and contractable ends of the lifting cylinders (6) are provided with rotating cylinders (7) respectively; rotating ends of the rotating cylinders (7) are provided with S-shaped baffles (8) respectively, the rotating cylinders (7) are located below the fixing frame (5), and a cross section of the fixing frame (5) is U-shaped;
wherein the identifying assembly (3) comprises an identifying equipment (9) and a fixed seat (10), a lower part of the identifying equipment (9) is fixedly connected to the fixed seat (10), and the fixed seat (10) is connected to the conveying assembly (1); and
wherein the processing assembly (4) comprises a crushing box (11) and electric heating plates (12), an inner side of the crushing box is connected to the electric heating plates (12), and the crushing box is connected to the conveying assembly (1);
wherein the device is based on a sorting-based garbage classification method, comprising:
step 1, feature information acquisition, comprising:
collecting domestic garbage components firstly, and then obtaining quantitative feature results of the domestic garbage components, to thereby obtain initial quantitative data of the domestic garbage components;
step 2, information preprocessing, comprising:
preprocessing the initial quantitative data to obtain original hyperspectral data, constructing an information extraction algorithm model, and inputting the original hyperspectral data into the information extraction algorithm model to carry out feature compression on the original hyperspectral data and express the original hyperspectral data by principal components;
step 3, algorithm model optimization, comprising:
optimizing the information extraction algorithm model with parameter indexes, and optimizing the number of the principal components of a feature compression part, and parameters of kernel functions including a linear kernel function, a radial basis kernel function and a polynomial kernel function of support vector classification model to obtain optimum parameter conditions, thereby generating a classification model used for sorting of domestic garbage components according to quantitative feature indexes; and
step 4, garbage components sorting, comprising:
grouping domestic garbage components in an ascending or descending order according to the classification model for subsequent disposal and utilization.
|