Automated Inorganic Pigment Classification in Plastic Material Using Terahertz Spectroscopy

Abstract

This paper presents an automatic classification of plastic material’s inorganic pigment using terahertz spectroscopy and convolutional neural networks (CNN). The plastic materials were placed between the THz transmitter and receiver, and the acquired THz signals were classified using a supervised learning approach. A THz frequency band between 0.1–1.2 THz produced a one-dimensional (1D) vector that is almost impossible to classify directly using supervised learning. This paper proposes a novel pre-processing of 1D THz data that transforms 1D data into 2D data, which are processed efficiently using a convolutional neural network. The proposed pre-processing algorithm consists of four steps: peak detection, envelope extraction, and a down-sampling procedure. The last main step introduces the windowing with spectrum dilatation that reorders 1D data into 2D data that can be considered as an image. The spectrum dilation techniques ensure the classifier’s robustness by suppressing measurement bias, reducing the complexity of the THz dataset with negligible loss of accuracy, and speeding up the network classification. The experimental results showed that the proposed approach achieved high accuracy using a CNN classifier, and outperforms 1D classification of THz data using support vector machine, naive Bayes, and other popular classification algorithms.

Review of Bioplastics Characterisation by Terahertz Techniques in the View of Ensuring a Circular Economy

“…Sarjaš et al used THz spectroscopy and convolutional neural networks (CNNs) to automatically classify inorganic pigments of plastic materials. They proposed a method to preprocess 1D THz data into 2D data, which is efficiently processed using convolutional neural networks [217]. Recycling bioplastics is sometimes made more difficult by the presence of additives such as pigments, which are present in almost every finished plastic product [218].…”

Interval-based sparse ensemble multi-class classification algorithm for terahertz data

“…Zhang and Li et al [13] used PCA to reduce the dimensionality of original THz spectral information, and then employed SVM, decision tree (DT), and RF to discriminate herbal medicines. Sarja et al [14] proposed a classification method for plastic inorganic pigments based on THz spectroscopy and convolutional neural networks (CNN). Huang and Cao et al [15] used THz spectroscopy to inspect mouse liver injury.…”

Fast FMCW Terahertz Imaging for In-Process Defect Detection in Press Sleeves for the Paper Industry and Image Evaluation with a Machine Learning Approach

“…On the other hand, the large sleeve area with a relatively limited number of round defects, which show with good contrast in the acquired terahertz images, constitutes a promising situation for the automated image processing and defect detection by machine learning (ML) approaches [ 39 ]. Application of ML techniques to terahertz measurements has been reported many times before, however, mostly in terms of direct application of the ML methods to the quite complex terahertz signals (in pulsed time-domain systems or continuous-wave systems [ 40 ]) and employing various sophisticated ML concepts such as artificial neural networks (ANNs) [ 39 , 41 ], random forests, support vector machines (SVMs) and many others (see [ 42 ] and references therein). There exist only few examples where ML is applied on the acquired terahertz images in an image processing sense, in which a direct evaluation of the image content itself is performed, rather than the measured signals.…”