Study On Parameters Extraction In Ionogram Based On Image Processing

In recent years, natural disasters around the world are serious, such asearthquakes, tsunamis and so on. Researchers find that these natural disastershappened with abnormal fluctuations in the ionosphere, so the study of the earthionosphere becomes a hot field in the space probe. In the study of the ionosphere, itneeds many of the parameters of the ionosphere which are mainly from artificialmeasurement process, however, with the increase and enhance of the ionospheric datafor real-time requirements in ionospheric prediction and detection, artificial measureefficiency cannot meet the current needs, which need to explore other ways to extractionospheric parameters and improve the efficiency of ionospheric sounding.In other countries, studies of ionospheric parameters that are automaticallyextracted are conducted by many research institutions, mainly in the United StatesLowell University Center for Atmospheric Research, the Italian National Institute forEarth Physics and Volcanology, the U.S. space environment and the Australiancompany ionospheric prediction service center. The main process of these studies is asfollows: Firstly, get the formation echo signal reflected from the ionosphere, thenfilter out the interference echo signal that is useful for forming a continuous trace, andfinally use mathematical methods to extract ionosphere parameter. Our research onthe automatic extraction of ionospheric parameters is relatively short time, mainlyfrom China Radio Institute and researchers in Harbin Institute of Technology. Thesestudies are mainly with some experience in the model of the ionosphere, usemechanism to form the echo signal, and extract the parameter by the method ofextrapolation or fitting the ionosphere trace.Based on the shape characteristics of hierarchical analysis methods combinedwith existing method of ionospheric parameter extraction, we propose a parameterextraction method based on image processing and recognition. By this method ofautomatic extraction, we study and explore the ionospheric parameters. The main tasks include:(1) By the way of the vertical integral projection to locate the effective frequencydomain, divide the target image region, and filter out the interference of non-targetarea information. Then we use the dividing line positioning method to separate the Eand F layers combined with ionospheric trace layered characteristics, providing thebasis for subsequent stratification parameter extraction.(2) Extract the parameters of the F layer. Interfering signal is weaker than thestrength of the desired signal according to the characteristics of high frequencyionospheric figure. By the gray stretch, between-class variance method and minimumfiltering approach, we filter out the ionospheric high frequency noise and isolatedpoints. Then by rotating projection integral and connected components labelingmethod high frequency secondary reflection of F layer is removed effectively and theF layer of the main trace is extracted. Finally use the skeleton and the limited windowmethod to extract F layer parameters.(3) Extract the parameters of the E layer. Firstly, perform morphologicalprocessing for the E layer image tracing and fix the breaking point of the trace. Thencombined with the shape characteristics of E layer, Es is discriminating. Finally withartificial measure of a priori knowledge of the E layer, parameter is extracted.