\"DETERMINATION OF SOIL COLOUR THROUGH HYPERSPECTRAL REMOTE SENSING TECHNIQUE\" TAU 416 PROJECT WORK (0+1) Submitted By Hema B (BSA-10-023) (BSA-10-038) Parashar Kumar Azad (BSA-10-051) Poonguzhali E (BSA-10-053) Project Guide

“DETERMINATION OF SOIL COLOUR THROUGH HYPERSPECTRAL REMOTE SENSING TECHNIQUE”

TAU 416 PROJECT WORK (0+1)

Submitted By

Hema B (BSA-10-023) Keerthana R.J (BSA-10-038) Parashar Kumar Azad (BSA-10-051) Poonguzhali E (BSA-10-053)

Project Guide Dr. R. JAGADEESWARAN Assistant Professor (SS & AC) Department of Remote Sensing & GIS

AGRICULTURAL COLLEGE AND RESEARCH INSTITUTE TAMILNADU AGRICULTURAL UNIVERSITY COIMBATORE – 641003

CERTIFICATE This is to certify that the project work entitled “DETERMINATION OF SOIL COLOUR THROUGH HYPERSPECTRAL REMOTE SENSING TECHNIQUE” submitted for fulfilment of the requirements for the course ‘TAU 416 – Project work (0 +1)’ to the Agricultural College and Research Institute, Tamil Nadu Agricultural University, Coimbatore is a record of bonafide work carried out by Ms. Hema. B (BSA-10-023), Ms. Keerthana.R.J (BSA-10-038), Mr. Parashar Kumar Azad (BSA-10-051), Ms. Poonguzhali .E (BSA-10-053) under my supervision and guidance.

Place: Coimbatore Date:

Dr.R.JAGADEESWARAN Asst. Professor (SS & AC) Project Guide Approved by: Chairperson: Dr. S. MAHIMAIRAJA, Dean (Agriculture) Members:

(i)

(ii)

Dr. R. JAGADEESWARAN Assistant Professor (SS & AC) Project Guide

Dr. BALAJI KANNAN Professor and Head (RS & GIS) i/c

ACKNOWLEDGEMENT We owe so many debts and gratitude, both intellectual and personal, to many people who have helped in this research endeavour. It is almost impossible to acknowledge our gratitude and debt to each of them. Thank you all, but we owe special thanks to some people who must be acknowledged. We would like to express our deep sense of gratitude to Dr.R.Jagadeeswaran, Assistant Professor (SS & AC), Department of Remote Sensing and GIS and Dr.R.Sivasamy, Professor & Head, Department of Remote sensing and GIS under whose valuable guidance this project work is completed. We greatly value their patiently advise, continuous encouragement, meticulous inputs, constructive criticism and feedbacks throughout the study as well as research period. It was indeed a pleasure to work under their guidance. Thanks to tireless efforts of Dr.Kalpana (U.G. coordinator) Assistant Professor, Department of Agronomy for his detailed review, constructive criticism and excellent advice during the preparation of this project report. Last, but not least we fervently record our heart full thanks to the Almighty, who showers blessings on us now and always. Project students: Hema B (BSA-10-023) Keerthana R.J (BSA-10-038) Parashar Kumar Azad (BSA-10-051) Poonguzhali E (BSA-10-053)

ABSTRACT Determination of soil colour is useful to characterize and differentiate soils. Routine determination of soil colour in the field is usually accomplished by visually comparing a soil sample with the Munsell Soil Colour Charts. The colour of soil can be measured in laboratory using Hyperspectral Remote Sensing Technique. Spectral reflectance measurement of 150 samples were carried out in the laboratory using spectroradiometer (model: GER1500; range: 350 to 1050nm). Colours were also visually estimated using Munsell soil colour charts. The Hyperspectral data were analysed for different indices and correlated with Munsell Soil colour measurement using correlation techniques and they showed good agreements. The overall correlation was 0.776. Thus the colour aspect of soil sample can be predicted from its spectral reflectance and this has numerous applications in remote sensing.

INDEX PAGE

Chapter No:

Title

1.

Introduction

2.

Review of Literature

3.

Materials and Methods

4.

Results and Discussions

5.

Summary and Conclusion Reference

Page No:

Chapter : 1 INTRODUCTION

Soil colour is an important soil property that is reported in all soil profile descriptions because it constitutes a useful first approximation of soil conditions and properties. Colour can be estimated with a Spectrophotometer or other mechanical device; but it is frequently done by visual inspection. The practice of describing soil colour first began in Russia, where attempts were made to form a cohesive system of soil colour identification. In America, soil colours were occasionally mentioned in reports of the early 1900’s, but no formal system was agreed upon until the 1940’s, when the work of Dorothy Nickerson and Albert H. Munsell led to the use of the colour chip system now employed. The system has led to a uniform and systematic description of soil colour employed in all current scientific literature.

Soil colour is used for both soil classification and evaluation. From colour, inferences regarding such things as reduction status (whether or not a soil remains waterlogged for long periods of time), organic matter content, iron oxide content and mineralogy are possible. For example, red, yellow, or reddish brown colours suggest the presence of oxidized iron and are indicative of good aeration and adequate drainage. Poor aeration and imperfect drainage are indicated by blue and gray soil colours, denoting reduced condition. Similarly, a dark brown soil colour is usually attributed to organic matter. Minerals can be distinguished by inspection from the differing values of redness; acid sulfate soils are frequently in the gray-green-black spectrum; and types of clays present have also been characterized by colour.

Soil colour is determined in the field by visually matching the samples with colour chips in standard Munsell Soil Colour charts. The Munsell colour system utilizes a descriptive system of hue, value and chroma, which limits the establishment of any quantitative relationships between changes in soil colour and soil constituents. Further, its accuracy and precision are determined by many factors including the incident light, the moisture content in the soil sample, the condition of the colour chart surfaces and the skill of the person making the match.

For example, a complete colour description reads 10YR 4/3. Such a notation translates to: a hue of 10YR, a value of 4, and a chroma of 3. Hue is the dominant spectral or rainbow colour and is described by such notations as 10YR (yellow red), 7.5YR (more red, less yellow), 2.5Y (yellow), etc. Each page in the Munsell colour book is a different hue. Value is defined as the relative blackness or whiteness, the amount of reflected light, of the colour. The value designation is found on the left side of the colour book, and increases from the bottom (0 = pure black), to the top (10 = pure white). The chroma notation is the purity of the colour or the amount of a particular hue added to gray. The chroma designation is located at the bottom of each page of the colour book and increases from left (grayest) to right (least gray or brightest).

The colour of soil is also measured by instruments such as photoelectric tristimulus colourimeters and Spectrophotometer and Spectroradiometers. The photoelectric colorimeters have source–filter-photodetector combinations that simulate the CIE Standard Observer functions. Spectrophotometers have wavelength isolation systems, such as gratings, prisms, or system of filters that provide the true reflectance spectrum of the sample; however, the wavelength band pass and wavelength accuracy varies widely the simpler to sophisticated

models. On the otherhand Spectroradiometers measures the reflected light energy as a function of wavelength. And the amount of energy reflected either in the natural light or artificial light is depends on the properties of the soil. Fernandez and Schulze (1987) calculated soil colours from their reflectance spectra. Escadafal et al (1989) measured spectral reflectance of soils and computed chromaticity co-ordinates expressed in RGB (Red, Green & Blue) notation, which were strongly correlated with the soil reflectance measured in the corresponding spectral bands of Landsat sensors. The spectral reflectance data obtained with Spectrophotometers are converted to C.I.E tristimulus values manually or with the help of appropriate software in external or built in computers.

Although the soil colour is commonly described using Munsell soil colour system, it has disadvantages viz., absence of direct mathematical conversions and require large look up tables to make continuous transformations. Hence, the Munsell system is mostly used for categorial qualifications of colour, which makes it less useful for numerical and statistical analysis (Melville and Atkinson, 1985). Such calculations have already been done for a light of the c-illuminant type and are available under the form of tables, which allow to go from one system to the other ( Wiszecki and Stiles, 1982).

With these background knowledge the present study was carried out to determine the soil colour through instrumentation technique (Spectroradiometer) with following objectives: i.

Determination of soil colour through Munsell Soil Colour chart as well as Spectroradiometer.

ii.

Establishing relationship between Munsell Soil Colour and Spectroradiometer data.

Chapter : 2 REVIEW OF LITERATURE Colour is widely used for soil characterization in the field and for soil classification. Soil colour is determined in the field by visually matching the samples with colour chips in standard Munsell colour charts. The Munsell colour system utilizes a descriptive system of hue, value and chroma, which limits the establishment ofany quantitative relationships between changes in soil colour and soil constituents. Further, its accuracy and precision are determined by many factors including the incident light, the moisture content in the soil sample and of the condition of colour chart surfaces and the skill of the person making the match. Schulze (1987) calculated soil colours from their reflectance spectra and concluded that spectrophotometer gave increased accuracy and precision which enabled to quantify small differences in soil colour. Melville and Atkinson Stoner (1979) observed the same dark red Munsell colour designation of 2.5 YR 3/6 from three different soils because of their proximity to the same Munsell colour chip. Shield et al (1968) measured soil colour spectrophotometrically at moisture levels ranging from air- dryness to field capacity and found a significant decrease in value with increasing moisture. Fernandez (1985) reviewed the procedures adopted for measurement of soil colour by Munsell soil colour charts and emphasized the need for the use of appropriate masks to eliminate the effects of disturbing contrasts. The manner in which light interacts with objects has been described by several workers (Hunter, 1975; Judd and Wyszecki ). In soil a small fraction of the incident light is reflected spectrally and a major portion of the beam penetrates into the soil mass where it encounters many surface of minerals and organic particles. The colour of the light scattered

from the soil mass results from the ability of the different components to absorb more light in some wavelength than others (Torrent and Barron, 1993). The colour of an object such as soil is highly dependent on its reflectance properties in the visible spectrum. Renaud Mathieu et al (1998) studied the relationships between the soil colour and simulated reflectance values for the Landsat TM and SPOT HRV satellites. Mattikalli (1997) developed a method called optimal rotational transformation technique to maximize the correlation between soil colour and transformed reflectance. Post et al (2000) measured soil colours with a Minolta Chroma Meter and spectral reflectance curves from 0.45 to 0.9 µm (measured in 0.1-µm increments) with a multispectral radiometer. All measurements were made on