Plant Leaf Image Segmentation Algorithm In The Complex Context Of Applied Research

Plants are important means of livelihood and production of human beings.It is important and essential for people to recognize and classify them at an efficient and fast manner.At present, many researchers have paid attention to plant classification using technology of computer image processing and pattern recognition.Object segmentation and extraction of digital image is a main branch causing more attention in digital image processing and computer vision field.For reasons that many algorithms can be used to segment and extract object area,combined with its widely use in computer vision,pattern recognition,it causes more and more researchers' attention. Image segmentation is the key step from image processing to image analysis,also foundation for further image analysis.Only after the proper image segmentation,image analysis can be made possible.Concerning images of plants may have different background,some is simple,some is complicate,we studied different image segmentation algorithms for different kinds of image that with different background.Using the proper image segmentation algorithm,we can get region of interest(ROI) in plant image,from which we can extract shape,color,or texture features for classifying plant.In this thesis,we do research on image segmentation algorithms on different kind of leaf images,especially the study and application of image segmentation algorithms on leaf image with complicate background.After image segmentation,we extracted shape,color, texture feature for classifying plant.In this thesis,we established a plant leaf image recognition system using plant leaves to recognize plant.Our research works are summarized as follows:1.Image segmentation:In this paper,we do research on some classic image segmentation algorithms,such as threshold segmentation,edge detect;considering leaf image with complicate background,we studied Snakes,Watershed algorithm,and used these algorithms for segmenting leaf images.2.Feature extraction:We extracted shape,color,texture features of region of interest (ROI) in the leaf image after image segmentation.We defined formulations of these feature and how to get these features are also introduced in detail.After the definition of feature,we extract 15 shape features,3 color features and 4 texture features for classifying plant.3.Classifying and recognizing of plant image:Considering the number of patterns and the dimensionality of pattern space are both very large,we studied some classifiers which are common used in image engineering,such as:Nearest Neighbor classifier,Nearest Feature Line classifier,and Back Propagation neural network and Support vector machine.Considering image feature data's character,we use support vector machine(SVM) as the classifier to process these shape,color,and texture feature data.4.Using VC ++ 6.0 to develop a plant leaf image recognition system.