Application And Research On Digital Meadow Management Using Content-based Image Retriveal

Meadow is an agricultural land.It is a must for animal husbandry.Digital management on meadow needs image processing.And the way we get the image we need is to retrieval.Sure we can recognize the image,but when there is a significant number of images in database,how can we find the image easily?We can use content-based image retrieval.It can improve the efficiency of the whole digital management system.So using content-based image retrieval on digital meadow management is meaningful.From the beginning of 70 20th century, the research on image retrieval had begun, which was mainly text-based image retrieval (Text-based Image Retrieval, referred to as TBIR), using the manner described in the text description of the image features, such as paintingsauthor, era, genre, size and so on. To 90 years after the emergence of the semantic content of the image, such as color, texture, layout analysis in image retrieval, that is, content-based image retrieval (Content-based Image Retrieval, referred to CBIR) technology. CBIR (Content-based Retrieval, referred to as the CBR) is part of CBR (Content-based Retrieval), CBR also include motion video, audio and other forms of multimedia information retrieval technology.In the theory of CBR,text-based image retrieval or content based image retrieval,both mainly include three aspects:on one hand, the analysis and transformation of user needs, to form the index so that the database can retrieved; On the other hand, the collection and processing of images resources, feature extraction, analysis and the indexing to build the image index database are also important; the last one is the similarity algorithm to calculate the user demo and the similarity index of the size of records in the database, extract the records that meet the threshold as a result, in accordance with similarity descending order output.Firstly this paper introduces the most popular technology of content-based image retrieval.Secondly it describes in detail the application of color features in the field of content-based image retrieval.Then through the analysis of fixed block partition and the ring partition we find the disadvantages of them.And using common color histogram brings high dimension feature vector.So this paper proposes filling main color in segmentation area using image segmentation technology in content-base image retrieval.Using image segmentation technology keeps the image's integrity and continuity.Using main color fits the characteristic of human eyes' vision feature,also it reduces the dimension of feature vector.Experiments shows that this method is a practical retrieval strategy,and has a lot of advantages in recall and precision.Finally, the paper designs and implements a set of content-based image retrieval system in digital meadow management.We build this system either for the purpose of proofing the way of filling main color in segmentation area using image segmentation technology in content-base image retrieval or for the further research on the 863 project--a resource-saving research on the digital meadow management in the northeast of China.