International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 03 Issue: 11 | Nov -2016 www.irjet.net p-ISSN: 2395-0072 © 2016, IRJET | Impact Factor value: 4.45 | ISO 9001:2008 Certified Journal | Page 182 Orthogonalise Digital Morphological Features Using Principal Component Analysis Vandana R. Patil 1 , Ramesh R. Manza 2 1 Assistant Professor, Department of MCA Engineering, KKWIEER, Maharashtra, India 2 Associate Professor, Department of CS & IT, Dr. BAMU, Maharashtra, India ---------------------------------------------------------------------------***---------------------------------------------------------------- Abstract - Plants play an important role in preserving earth ecology and maintaining healthy atmosphere. Most of the food sources are from plants. So it is necessary to establish a system to extract the features of plant. As there are large variety of plant species available in the universe . There are various parts of plant which can be chosen for feature extraction. In this research we choose shape of leaf to extract the features. Every plant leaf has variety of features to extract. This research reduces large number of digital morphological features to five principal components. This research describes procedure to transform raw image to a preprocess image for feature extraction. Further five geometrical features are extracted from preprocessed image and from geometrical features twelve digital morphological features are extracted. Due to large number of digital morphological features it is difficult to proceed for further for plant identification. These twelve digital morphological features are orthogonalise to five principal components using principal component analysis. Key Words : Morphological feature, plant leaf classification, feature extraction, digital morphological features, principal component analysis. 1.INTRODUCTION Plants play an important role in our ecosystem. They are useful for human beings and animals too. Plants provide food, medicine, oxygen and many important substances for living and non living thing. Plants having very broad range of applications in agriculture and medicine. So it is necessary to know which plants are useful or harmful to human being to save the life of human being. Due to the effect of global warming large number of rare plant species are at the boundary of extinction. But it is important and difficult task to recognize the plant species. Plants recognition is useful in agriculture and medicine as well a biological diversity research. It is also useful in tea , cotton and other industries. So it is necessary to develop a database by information technology as soon as possible. To recognize the plant species researcher needs to extract the various features of plant species. To proceed with this, features can be extracted from various parameters of plant like leaf, fruit, flower or stem . In this research we choose shape of leaf parameter to extract the feature. The aim of this research is to transform the digital morphological features of leaf to principal components. As the number of features extracted from leaf are in large number , so it is difficult for classifier to proceed with this to classify the plant species from plant kingdom. The finding of this research is useful for plant classification purpose. 2. PREVIOUS WORK Global feature and local descriptor are two categories for features of leaf, as stated by Shabanzade et al(2011)[1]. According to C. L. Lee , S. Y. Chen [2] past research in recognizing objects can be broadly classified into two categories : a) contour based and b) region based approaches. The disadvantage of the contour based feature is the difficulty on finding the correct curvature points. Based on the contour of leaf, features were extracted to differentiate species. However contour of leaves have variation even in the same species. For plants identification purpose Wu, et al[3] used shape slimness, defined as ratio of length to width of leaves, shapes roundness, defined as ratio of area of leaf image and perimeter of leaf contour, and shape solidity, defined as ratio of the internal area connecting to valley points and the external area connecting the top points. A paper by Ji- Xiang Du, Xiao- Feng Wang, Guo-Jun Zhang [4] introduce how to extract digital morphological features. The features are extracted from the contours of leaf. The digital morphological features(DMF) generally include geometrical features(GF) and invariable moment features(MF). A paper by Cholhang Im, Hirobumi Nishia and Tosiyasu L. Kunii[5], a method that normalizes shapes of leaves is presented using symmetry of each leaflet with respect to its vein. According to Najjar and Zagrouba [6] had used region based feature for proposed method in order to classifying the leaf. According to C.S Sumathi and A.V.Senthil Kumar in plant