A Review of Hyperspectral Remote Sensing and its Application for Water

Pollution

Introduction
Water, the hydrosphere, covers approximately 71% percent of the earth. It consists of
ocean, river, lake, marsh, glacier, snow, groundwater, air moisture, and so on. Water
environment, closely linked with human beings life, is facing serious problems of
pollution and eutrophication. Water is one of the most valuable and essential
resources that form the basis of all life.
Over the last decade, the increased spectral and spatial resolution of remote sensing
equipment has promoted the development of new methods for water bodies
monitoring. The advantages of the

new physical approaches, with respect to

empirical methods, include the capability of fully deploy hyperspectral data, the
reduction of the amount of laboratory and in-situ measurement. This new methods
require very rigorous data processing. The measured signal, in fact, has to be as
similar as possible to the real signal to be interpreted on the bases of a physical
model.
Hyperspectral systems have made it possible for the collection of several hundred
spectral bands in a single acquisition, thus producing many more detailed spectral
data. However, with the advances in hyperspectral technologies practical issues
related to increased sensor or imager costs, data volumes and data-processing costs
and times would need to be considered especially for operational modes.
Water pollution is one of the major threats to public health in the world. Water
pollution refers to harmful substances released into surface or ground water, either
directly or indirectly. water pollutants can originate, for example from waste water
stabilization ponds, sludge lagoons, barnyard runoff, septic tank leaching fields or
seepage pits, pit privies and the deep well disposal of certain industrial wastes or
treatment plant effluents
This article is structured as follows. Section 2 presents the overall overview of the
hyperspectral technology. Section 3 presents the application of hyperspectral on the

water pollution types. Section 4 includes current problems of the hyperspectral and
recommendations for the future improvements. Section 5 closes the paper with
conclusions.

Overview of the hyperspectral technology

The trend in the of development of remote sensing has been, with the increase of the
spectral resolution, to move from the panchromatic multispectral to the hyperspectral,
and then to the ultraspectral. Sensor developments include a new generation of highresolution commercial satellites that will provide unique levels of accuracy in spatial,
spectral and temporal attributes.
In the airborne remote sensing system, the hyperspectral sensor is already in place
as one of the basic systems. Two new kinds of hyperspectral sensors, PHI and
OMIS, were designed specifically for hyperspectral applications. In addition, a small
hyperspectral digital camera system (HDCS) with limited number but narrow band
was also implemented for environmental and agricultural monitoring.
With the

development and perfection of the hyperspectral remote sensing

technologies, hyperspectral remote sensing has been the major technique applied in
many studies. Now with commercial airborne hyperspectral imagers such as CASI
and Hymap and the launch of satellite-based sensors such as Hyperion,
hyperspectral imaging is fast moving into the mainstream of remote sensing and
applied remote sensing research studies Hyperspectral images have found many
applications in water resource management, agriculture and environmental
monitoring (Smith, 2001a). For hyperspectral sensors have become available to
provide both high spatial and high spectral resolution with high signal/noise ratio. Due
to the sufficient spectral features such as spectral reflectance with wavelength it
provides, hyperspectral data plays an important role in the different fields.
To obtain data of a higher spectral resolution compared to multispectral data,
hyperspectral sensors on board satellites or airborne hyperspectral imagers are used
(Smith, 2001b).

Hyperspectral remote sensing imagers acquire many, very narrow, contiguous

spectral bands throughout the visible, nearinfrared, mid-infrared, and thermal infrared
portions of the electromagnetic spectrum. Hyperspectral sensors typically collect 200
or more bands enabling the construction of an almost continuous reflectance
spectrum for every pixel in the scene
Brownfields, refineries, tank farms, pipeline corridors, oil fields, and other industrial
sites can have surface deposits of hydrocarbon-based materials. Sophisticated
VNIR-SWIR hyper-spectral sensor is ideal for detecting soils and surfaces that have
been impacted by oil-based materials. Some researchers have found that the
detection of oil spills in soil is related with the concentration of light hydrocarbons in
the soil and in the air. Light hydrocarbons tend to evaporate fairly quickly, therefore is
a time constrain with this way of detection. On the other side, the more hydrocarbons
get evaporated from soil or the water the less environmental damage in can cause.
Tests

have shown that hydrocarbons in soil and plastics are characterized by

absorption maxima at wavelengths of 1730 and 2310nm.

Hyperspectral Analysis
The sensor used for the study of the Patuxent River oil spill is the Airborne Imaging
Spectroradiometer for Applications (AISA) sensor system. AISA hyper-spectral
imaging sensor can measure up to 55 spectral bands of information; has an airborne
DGPS (Differential Global Positioning System - to measure aircraft position); and an
INS (Integrated Navigation System - to combine the DGPS and an IMU (Inertial
Measurement Unit) - to measure aircraft attitude. AISA is a solid-state, push-broom
instrument of small size, which makes it perfect for use in aircrafts. The instrument
can be mounted on a plate that is compatible with a standard aerial camera mount,
and has the flexibility of selecting the sensor's spatial and spectral resolution
characteristics. AISA is capable of collecting data within a spectral range of 430 to
900 nm, and up to 286 spectral channels within this range. Current operational
collection configurations for the AISA hyperspectral sensor covers a range from 10 to
70 spectral bands, this will depend on the aircraft speed, altitude, and the 1212
specific mission goals.

Hyperspectral systems have made it possible for the collection of several hundred
spectral bands in a single acquisition, thus producing many more detailed spectral
data. However, with the advances in hyperspectral technologies practical issues
related to increased sensor or imager costs, data volumes and data-processing costs
and times would need to be considered especially for operational modes.

Problems

of

the

hyperspectral

and

recommendations

for

the

future

improvements
. In this specific study (oil in water) since the signature can be easily misidentify as
water, hyper-spectral imagery can help to obtain a more detail spectrum to be able to
separate between pure water and oil-water

Conclusions
The use of hyper-spectral imagery to detect oil spills in water has a lot of advantages
in the field. It can be use to monitor oil facilities and therefore prevent worst scenarios
when a leak in the facility is found. Also can be use to help planning the cleanup of
the area, by quickly identifying the affected areas and possible path of the spill to be
one step ahead.