Premium
Color difference classification of dyed fabrics via a kernel extreme learning machine based on an improved grasshopper optimization algorithm
Author(s) -
Li Jianqiang,
Shi Weimin,
Yang Donghe
Publication year - 2021
Publication title -
color research and application
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.393
H-Index - 62
eISSN - 1520-6378
pISSN - 0361-2317
DOI - 10.1002/col.22581
Subject(s) - color difference , extreme learning machine , artificial intelligence , algorithm , kernel (algebra) , computer science , grasshopper , machine learning , mathematics , enhanced data rates for gsm evolution , ecology , combinatorics , artificial neural network , biology
Conventional artificial color difference detection is error prone and inefficient. Herein, a novel color difference classification model is proposed for dyed fabrics via a kernel extreme learning machine based on an improved grasshopper optimization algorithm. First, in order to prevent the grasshopper optimization algorithm from succumbing to local optimality, it is proposed to optimize the initial population of the algorithm using differential evolution, resulting in a better solution direction at the outset. Then, this novel grasshopper algorithm is used to adjust the kernel bandwidth and penalty parameters of the learning maching model, thereby forming a color difference classification model for dyed fabrics based on the differential evolution grasshopper optimization algorithm kernel extreme learning machine. Finally, the key indicator values representing color difference are extracted from the printed and dyed product images collected under the standard light source. The color difference calculated by substituting these values into the color difference formula generates a color difference dataset, which is used to train and test the color difference classification model. Experimental results show that the proposed differential evolution grasshopper optimization algorithm kernel extreme learning machine model has high classification accuracy and impressive stability. The average classification accuracy of the proposed model is 98.89%, whereas the accuracy of kernel extreme learning machine is only 91.08%.