Interactive Image Segmentation Method Based On Candidate Boundary Points

With the continuous development of radio technology,the radio service has been continuously expanded,providing many conveniences for people's production and life.At the same time,the increasing number of radio services and the increasingly complex electromagnetic environment have brought challenges to radio spectrum monitoring and management.Problems such as tight spectrum resources and illegal occupation have become the focus of spectrum monitoring and management.In order to solve the problems faced by radio regulation,a lot of work has been carried out around spectrum data monitoring and analysis.Compressed storage of massive spectrum data is very important.As an important technology in the field of artificial intelligence,pattern recognition is also widely used in spectrum monitoring and spectrum data compression,and has achieved good results.As an important part of pattern recognition technology,pattern extraction has a great influence on the effect of pattern recognition,but there is almost no research on pattern extraction of spectrum data.In this paper,by analyzing the characteristics of spectrum data and existing clustering algorithms,the extraction of spectrum data patterns and application of pattern recognition of spectrum data are studied.The main work of this article are as follows:(1)An improved k-means clustering algorithm is proposed.In this algorithm,there is no need to specify the number of clusters,and an appropriate initial cluster center is automatically selected in the data set.The algorithm has also been improved in iterative optimization,which can quickly converge and achieve better clustering results.The improved algorithm is very suitable for cluster analysis of spectrum data,and can better meet the needs of spectrum data pattern extraction.(2)Pattern extraction of spectrum data based on the clustering algorithm is provided.By analyzing the characteristics of spectrum data and channel data,several frequency division methods are proposed.On the basis of frequency division,an improved clustering algorithm is used to perform cluster analysis on the sampled data of the channel,and the cluster center is used as the mode of the channel.(3)The application of spectrum data mode is analyzed.The spectrum data mode is applied to the compressed storage of the spectrum data.In this application,the mode matching method is used to obtain the channel monitoring data matching mode,and the serial number of the matching mode is used as the stored data,which effect achieved is better than other methods.