Traditional and Innovative Approaches for Detecting Hazardous Liquids

Traditional and Innovative Approaches for Detecting Hazardous Liquids

Ebru Efeoglu, Gurkan Tuna
DOI: 10.4018/978-1-7998-6870-5.ch020
OnDemand:
(Individual Chapters)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

In this chapter, traditional and innovative approaches used in hazardous liquid detection are reviewed, and a novel approach for the detection of hazardous liquids is presented. The proposed system is based on electromagnetic response measurements of liquids in the microwave frequency band. Thanks to this technique, liquid classification can be made quickly without pouring the liquid from its bottle and without opening the lid of its bottle. The system can detect solutions with hazardous liquid concentrations of 70% or more, as well as pure hazardous liquids. Since it relies on machine learning methods and the success of all machine learning methods depends on provided data type and dataset, a performance evaluation study has been carried out to find the most suitable method. In the performance evaluation study naive Bayes and sequential minimal optimization has been evaluated, and the results have shown that naive Bayes is more suitable for liquid classification.
Chapter Preview
Top

Background

Hazardous liquids, many of them are very exothermic and/or corrosive, can cause considerable and mostly irreversible damage to organs by contact or inhalation even at low concentrations (Marrs, Maynard, & Sidell, 1996). Some of these can cause considerable damage to property or human health, even in low concentrations and quantities. Chemical warfare agents are toxic enough to cause an instant damage when inhaled or when in contact with the skin (Marrs, Maynard, & Sidell, 1996). Some highly volatile materials can be deployed easily just by opening their containers. The level of damage associated to a concealed hazardous liquid depends on the type of material, deployment method, area, hazard level, concentration and properties at room temperature (Marrs, Maynard, & Sidell, 1996).

Key Terms in this Chapter

Non-Contact Detection: A process used to remotely detect a material without a physical contact.

Patch Antenna: A patch antenna is a type of radio antenna, which can be mounted on a flat surface.

Microwave: Microwave is a form of electromagnetic radiation with wavelengths ranging from about one meter to one millimeter and with frequencies between 300 MHz and 300 GHz.

Classification Algorithm: An algorithm that maps the input data to a specific category.

Cross-Validation: It provides information about how well a classifier generalizes and is generally used to evaluate machine learning models on a limited data sample.

Safety Data Sheet: A safety data sheet is a document that provides information relating to occupational safety and health for the use of various substances and products.

Liquid Explosive Detection System: A device used to detect explosive substances in liquid form, generally used at airports.

Complete Chapter List

Search this Book:
Reset