Algorithm Research Of Extracting Boundaries Of Sample From Magneto-acoustic Signal Based On

In recent years, study on electrical characteristics of biological tissue had attracted much attention from experts in many fields. Research conclusion has indicted that not only large differences existed in electrical characteristics between different biological tissues, but also that electrical characteristics of tissues and organs varied in a large range with functional condition. So imaging of the electrical characteristics can reflect human physiological and pathological condition and their diversification. Nondestructive detection of electrical characteristics of biological tissue is significant for clinical diagnosis, especially for early diagnosis of cancer.With Visual C++as the software platform, this paper aimed to analyze characteristics of magneto-acoustic signal, reconstruct the boundaries of peaks and correspond the inner and outer boundaries of the copper loop with the boundaries of peaks.In experimental research, the signal collecting control system is introduced firstly and the control flowchart is given. In the preliminary experiment, the signal collecting control system performs well and cooperates with the signal generating system. We counted the steps needed when the stepper motor rotated a round and then calculated the precious minimum degree per step which is very important to precision of peak boundary reconstruction. According to forerunners’conclusion, sine signal is chosen as the exciting signal. When the probe collects signal at different angle, the amplitude of peaks differs. When the probe is perpendicular to the current direction, the amplitude is max. Even little divergence will cause lager attenuation which infers that the vibration source vibrates at a particular direction. When the distance between the probe and the sample become large, the amplitude of the peaks becomes smaller. Research shows that the attenuation is linear in a particular range.In the copper experiment, we collect200signals (335°) with the help of step motor. We divide one typical signal into three parts. The first part contains only random noise; major component of the second part is electromagnetic interference noise which generates in the probe; the major component of the last part is useful boundary signal. We calculate the Fourier Transform of the three parts respectively and analyze distribution of the frequency spectrum. At last, we find the range where the useful signal hides, and then we chose proper filtering function. In the restored signal, by means of "extreme value-monotonous contrarily in two sides" signal peaks are detected and by means of "tracking trend of peaks", boundaries of peaks are extracted. Then diameter and circumference of boundaries are calculated. According to diameter and circumference, we correspond the inner and outer boundaries of the copper loop with the peak boundaries and analyze the deviation.In this paper we do experimental research and correspond the copper boundaries to peak boundaries, which accumulate experimental data for further magneto-acoustic signal analysis and boundary reconstruction research.