System Of Agricultural Product Trademark Retrieval Based On Content

With the rapid development of the rural economy and agricultural products mar ketization, agricultural product trademark is playing a more and more important role in the agricultural economy, consequently, the problem about the agriculture intellect ual property becomes more and more serious. In response to "serving for issues of ag riculture, farmer and rural area by the use of the agricultural product trademark" of t he government, the thesis take the agricultural product trademark as the research obj ect to conduct a research on the trademark retrieval, the main work is done as follows:Firstly, this paper mainly aims at the shape features of trademark images of agricultural products in the research. This paper takes the boundary contour features of the image as the whole shape characteristics of trademark image of agricultural products, the local pixels distribution characteristics as the local shape characteristics of trademark image of agricultural products, uses the combination of regional characteristics and the boundary characteristics to realize the agricultural product trademark image retrieval based on the content.Secondly, the trademark image is divided into text trademarks, graphic trademarks, combined trademark, the thesis puts forward the three kinds of agricultural product trademark retrieval method. Fourier descriptor is used in the description of global shape feature of three kinds of agricultural product trademark image to represent boundary features in the method. In the local shape feature description, sum of sub image object pixels, the ratio of sub image object pixels sum and the background pixel sum, and the ratio of the sub image object pixels sum in the whole object pixels sum respectively represent the regional characteristics of combined trademark, graphic trademarks, text trademarks. Suited retrieval of feature classification can improve the retrieval efficiency of trademark. Finally, the paper takes 5000 agricultural product trademark images as the sample, use the method of artificial classification to establish trademark image database, use the advanced programming language MATLAB to realize the content-based agricultural products brand retrieval system, and analyze the performance of the system through the distortion similarity experiment and visual similarity experiment. The experimental results show that the content-based trademark retrieval algorithm has better retrieval performance in the paper. The content-based trademark retrieval system which is realized by this algorithm can complete the retrieval of the agricultural products trademark image quickly and effectively.