JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.

ORG

133

Identifying Crop Type And Classifying It Using Local Binary Pattern variance
Ashwin A. Shinde, Dr.Maneesh Shrivastava
Abstract Manual survey of crop estimation takes a lot of time and are often misprice and unreliable. Crop monitoring using remote sensing (RS) enables timely, economical, reliable, and the most important comprehensive and accurate crop Identification. The aim of this research is to evaluate crop discrimination using satellite image data by following remote sensing approach. This paper illustrates the use of Local Binary Pattern Variance on satellite images to classify the land in to crop land and non-crop land and to classify different crops. The input image is enhanced first then Local Binary Pattern Variance is used to extract features from the crop images specifically extracting green colors. After identifying the LBPV pattern of each pixel (i,j) in the given image the whole texture image is represented by building a histogram showing intensity values for uniform and non uniform patterns. A texture image database of different crops is created. The texture features of the input image are then compared with texture features obtained from the image database of different crops and the different types of crops are identified. Index Terms Remote Sensing, Image Enhancing, LBPV, Histogram, Crop Identification, Crop Classification.

1 INTRODUCTION
onitoring, estimating and forecasting agricultural production are very important for the management of world / regional or local food demand and supply balance for social security.The importance of agriculture for the Indian society can hardly be over-emphasized, as its role in economy, employment, food security, national self reliance and general well being, does not need reiteration. The need for timely and reliable information on crop area and production for tactical and strategic decision making by all stakeholders in agriculture, such as producers, processors, resource managers, marketing, finance and the government is well known. With the global shift in the market economies, reliable agricultural information has gained more importance than ever before. Remote Sensing (RS) using space-borne sensors is a tool, par excellence, for applications such as crop inventory, crop production forecasts, drought and flood damage assessment, range and irrigated land monitoring and management. Crop Identification using Remote Sensing (RS) images is the basis of an agricultural monitoring system using RS (Yang 2005), to meet the required precision, visual interpretation is used in routine operations. Therefore, development of an operational Crop identification model that suits agricultural situation and quickly pro-

vides comprehensive, reliable information of crops which is an important task which the operational crop monitoring system is facing. The satellite image processing is becoming increasingly available for vegetation mapping and to decision makers for future growth and development. Remote sensing is identified as a tool to assess performance more than a decade ago [1] [2]. In the last decade, remote sensing has been increasingly identified an objective, standardized, possibly cheaper and faster methodology for crop production surveys than conventional field investigation [3] [4]. Regional state government officers in India currently process the technique of Crop Estimation by selecting farmers or village officials and interviewing them regarding their crops. Comparing these results with the previous years analysis, information is then extrapolated to generate data and predictions on a regional basis, which means that these traditional surveys are time-consuming. In addition, the information collected is often imprecise and unreliable, leading to inaccurate crop-yield forecasts and subsequent difficulties for agricultural planners and managers on both regional and national scale. Also in the case of Natural calamities such as flood or draught or pest infection, the survey to find out damage to crops takes several months and; the finical or any other Ashwin A. Shinde,Student,Department of Information Technology, Laxmi aid may not reach timely to the needy farmers. Therefore, Narain College of Technology, Bhopal,MP,India. there is need for sophisticated methods which are accurate enough for crop estimation and such surveys to Dr.Maneesh Shrivastava, Professor, Head, Department of Information Technology, Laxmi Narain College of Technology, Bhopal, MP, India. avoid delays so that timely action can be taken. Satellite imaginary use is proposed in this research. The very first thing is the image acquisition that is the source from where the images are acquired. In this paper

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

134

we used a technique of enhancement which is simple but more effective than some of the existing techniques reported earlier. Also for the process of identifying crop types and classifying them an appropriate feature extraction algorithm and classification algorithm are very important. These classification algorithms include Multiple Classifiers [5] [6], Neural Network [7] [8], Support Vector Machine [9], fuzzy classification [10] and other biologically inspired algorithms [11]. These classifications were based on various features such sub-pixels, wavelet functions and texture features etc. In this paper the use of Local Binary Pattern Variance for obtaining features from satellite images is demonstrated for land classification as crop and non-crop area. Local Binary Pattern Variance is one of the most effective feature extraction techniques for textures, as the LBP Features are believed to be rather consistent to the response of Human Vision System (HVS). It provide a comprehensive experimental evaluation of feature extraction as compared with other multi-resolution texture features indicating that the LBPV features provide the best pattern retrieval accuracy [12].

68 57 54

51 55 22

56 15 72

Threshold

1 1 0

0 0

1 0 1

Binary Code: - 10101001

Fig.1 Illustration of basic LBP Operator

g5 g4 g3

g6 gc g2

g7 g0 g1

LBP8,1= S(g0-gc)20+S(g1-c)21 S(g2-gc)22+S(g3-gc)23 S(g4-gc)24+S(g5-gc)25 S(g6-gc)26+S(g7-gc)27

Fig.2 The LBP operator of a pixels circular neighborhoods with r=1, p=8

Bilinear interpolation method is used for a sampling point does not fall in the center of the pixel. Let LBP p,r denote the LBP feature of a pixel s circularly neighborhoods, where r is the radius and p is the number of neighborhood points on the circle .From Fig.2 we can write p-1 LBP p, r = s(gi-gc)2i , S(x)= {1
If x0 (1)

2 LBPV
Local Binary Pattern (LBP) is a simple yet very efficient texture operator which labels the pixels of an image by thresholding the neighborhood of each pixel and considers the result as a binary number. Due to its discriminative power and computational simplicity, LBP texture operator has become a popular approach in various applications. It can be seen as a unifying approach to the traditionally divergent statistical and structural models of texture analysis. Perhaps the most important property of the LBP operator in real-world applications is its robustness to monotonic gray-scale changes caused, for example, by illumination variations. Another important property is its computational simplicity, which makes it possible to analyze images in challenging real-time settings. Local Binary Pattern was introduced by Timo ojala [13].The standard version of the LBP of a pixel is formed by thresholding the 3X3 neighborhood of each pixel value with the center pixels value. Let gc be the center pixel gray level and gi (i=0,1,..7) be the gray level of each surrounding pixel.Fig.1 illustrate the basic LBP operation. If gi is smaller than gc , the binary result of the pixel is set to 0 otherwise set to 1. All the results are combined to get 8 bit value. The decimal value of the binary is the LBP feature. Bilinear interpolation method is used for a sampling point does not fall in the center of the pixel. Let LBP p,r denote the LBP feature of a pixel s circularly neighborhoods, where r is the radius and p is the number of neighborhood points on the circle.

P=0 0 otherwise Each such type of LBP code generates uniform patterns in the image.The concept of uniform patterns is introduced to reduce the number of possible bins. Any LBP pattern is called as uniform if the binary pattern consists of at most two bitwise transitions from 0 to 1 or vice versa. For example if the bit pattern 11111111(no transition) or 00110000 (two transitions) are uniform where as 10101011 (six transition) are not uniform. The uniform pattern constraint reduces the number of LBP pattern from 256 to 58 and it is very useful for crop detection [10]. LBP variance (LBPV) is proposed to characterize the local contrast information into the one-dimensional LBP histogram. The experimental results on representative databases show that the proposed LBPV operator and global matching scheme can achieve significant improvement. The LPBV descript or proposed offers as solution to the above problems of LBP;R=VARP;R descriptor. The LBPV is a simplified but efficient join t LBP and contrast distribution method. As can be seen in Equation (2), calculation of the LBP histogram H does not involve the information of variance VARP;R. That is to say, no matter what the LBP variance of the local region, histogram calculation assigns the same weight 1 to each LBP pattern. Actually, the variance is related to the texture feature. Usually the high frequency texture regions will have higher variances and they contribute more to the discrimination of texture images [8]. Therefore, the variance VARP; R can be used as an adaptive weight to adjust the contribution of the LBP code in histogram calculation. The LBPV histogram is computed as:

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

135

N M W (LBP p,r(i, j),k) k 0 LBPV p, r (k) = i=1 j=0 W(LBP p,r (i,j),k,) ={VARP,R(I,J) 0
varp,r(I,j),k=0

(2)

(3)

otherwise

The paper is organized as follows: Section 3 contains the image enhancement, Section 4 contains a feature extraction using Local Binary Pattern Variance, Section 5 Segmentation, Section 6 Land classification and Section 7 explains Crop Identification part, followed by the conclusions drawn from the work with the proposed system. The proposed flow can be given as below: TABLE 1 RESEARCH METHODOLOGY FLOW Image Acquisition Image Preprocessing Image Resizing Image Enhancement Extraction of Features using LBPV Image Segmentation (region of interest) Land Classification & Crop Identification
Fig.3 Input Crop Land Image

3 IMAGE ENHANCEMENT
Image enhancement techniques are designed to improve the quality of an image as perceived by a human being to improve the interpretability of the information present in images. Image enhancement is one of the most interesting and important issues in digital image processing field. The main purpose of image enhancement is to bring out details that are hidden in an image, or to increase the contrast in a low contrast image .Image enhancement produces an output image that subjectively looks better than the original image by changing the pixels intensity of the input image. Generally, image enhancement enlarges the intensity differences among objects and background. The input crop land image shown in Fig.3 is being enhanced to improve its visibility and interpretability for human visibility system. The Enhancement procedure is basically followed for removing unwanted noise and errors from the image. In Fig.4 we show the enhanced version of the image in Fig.3. As can be seen from the Fig. 4 that green colored portions of the given crop land image are enhanced to the extent so that they are clearly visible according to the human vision.

Fig. 4 Enhanced Crop Land Image

4 FEATURE EXTRACTION
In Crop Type classification appropriate feature selection from satellite images is important. Local Binary Pattern Variance is an advanced feature extraction algorithm which extracts the green color from the crop images in order to classify it later by comparing the extracted part with that of the template patterns stored in the texture database. As shown in Fig. 5 specifically green color is extracted using Local Binary Patten code. While the green color is being extracted using LBPV, GETMAPPING method is used which returns a mapping table for LBP codes. After this, numbers of patterns in the resulting LBP code are

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

136

generated, uniform well as rotation invariant patterns are generated to complete features extraction part. The LBP and VAR feature extractors are first reviewed. To address the limitation of VAR, the LBPV is then proposed, as LBP ignores sensitive features of the image such as edges and boundaries. Finally, the matching dissimilarity metric in this work is presented. LBPV is a texture operator which characterizes the spatial structure of the local image texture. Given a central pixel in the image, a pattern number is computed by comparing its value with those of its neighborhoods as shown in equation 1.

est which in turn will then be compared with the texture patterns stored in the texture database if a match is found then particular crop type will be identified.

Fig.7 Segmented Part (Region of Interest)

4 LAND CLASSIFICATION
It can be seen from Fig. 5 that the green colored part can be termed as crop land and the brown colored part can be termed as the non-crop land.

5 CROP IDENTIFICATION
The texture database is used in the experiments which consist of different texture classes. A query pattern is any one of the image in the database. This image is then processed to compute the LBPV feature vector. The distance d (i, j), where i is the query pattern and j is a pattern from the database, is computed. The distances are then sorted in increasing order and the closest set of patterns are then retrieved and grouped together for identification of crop type.

6 CONCLUSION
Fig. 5 Extracted Green Colored Part

The extracted green colored features and its intensity values of uniform patterns in the image can be shown by a stem plotted histogram.

In this research, Local Binary Pattern Variance method is used. A texture based classification method will be used for the classification of land as well as for the classification of different crops and then the method can be applied for the monitoring of the different crops throughout the year.

ACKNOWLEDGMENT
I would like to thank my guide Mr. Maneesh Shrivastava in guiding me writing this paper.

REFERENCES
[1] Abernethy, C.l.; and Pearce, G.r. 1987. Research needs in third world irrigation, Hydraulics Research Limited, Wallingford, England. Saha, S.k., 1999. Crop yield modelling using satellite remote sensing and GIS-current status and future prospects, Processing Geo informatics-Beyond 2000, an international conference on Geo informatics for natural resources assessment, monitoring and management held at IIRS Dehradun during 9-11 March, 1999. Singh, Randhir and Ibrahim, AEI, 1996.,Use of spectral data in markov chain model for crop yield forecasting, Jr. Indian Socy. of Remote Sensing 24 No.3, 145-152. Bauman, B.A.M., 1992, Radiometric measurements and crop yield forecasting-some observations over Millet and Sorghum experimental plots in Mali, International Journal of Remote

[2]

Fig.6: Histogram of Uniform Pattern Intensity Values.

5 SEGMENTATION
The green colored extracted image will be segmented into different parts according to the region of interest. Purpose of image segmentation is to divide the image into some meaningful regions. In segmentation part we will be selecting specific extracted green colored parts of our inter[3]

[4]

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 4, APRIL 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

137

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

Sensing, 9(10):1539-1552. Defries, R.S., Chan, J.C.W., Multiple Criteria for Evaluating Machine Learning Algorithms for Land Cover Classification from Satellite Data, RSE (74), No. 3, 2000, pp. 503-515. Steele, B.M., Combining Multiple Classifiers. An Application Using Spatial and Remotely Sensed Information for Land Cover Type Mapping, Remote sensing of environment (74), No. 3, 2000, pp. 545- 556. 0102 Ji, C.Y., Land-Use Classification of Remotely Sensed Data Using Kohonen Self-Organizing Feature Map Neural Networks, PhEngRS (66), No. 12, December 2000, pp. 1451-1460. Kavzoglu, T., Mather, P.M., Pruning artificial neural networks: an example using land cover classification of multi-sensor images, JRS(20), No. 14, September 1999, pp. 2787. Huang, C., Davis, L.S., Townshend, J.R.G., An assessment of support vector machines for land cover classification, JRS(23), No. 4, February 2002, pp. 725-749. Omkar . S.N. Senthilnath . J. et al. 2008 Crop Classification using Biologically-inspired Techniques with High Resolution Satellite Image, - Journal of the Indian Society of Remote Sensing, 36 (2). pp. 175-182. Islam, Z., Metternicht, G., The Performance of Fuzzy Operators on Fuzzy Classification of Urban Land Covers, PhEngRS (71), No. 1, January 2005, pp. 59-68. B.S. Manjunathi and W.Y. Ma, Texture Features for Browsing and Retrieval of Image Data, IEEE Transactions on pattern analysis and machine intelligence, Vol. 18,No.8, August ,1996 Timo Ahonen, Matti Pietikainen and T.Maenpaa, Multiresulation gray scale and rotation invariant texture classification with local binary pattern, IEEE Transactions on pattern analysis and Machine Intelligence Volume 24, No 7, pp 971-987, 2002.

Ashwin A. Shinde BE (IT) 2009 passout.Currently working as Lecturer in Dr. Babasaheb Ambedkar College of Engineering and Research, Nagpur, Maharashtra, India. Presented 1 paper in International conference ICECT 2012, Kanyakumari.Current research area interest in Image Processing, Databases. Dr.Maneesh Shrivastava Professor, Head, Department of Information Technology, Laxmi Narain College of technology,Bhopal, Madhyapradesh,India.