Bowel Sounds Acquisition System Design And The Signal Analysis Methods Research

Bowel sounds is one of the important physiological signals of human body, it reflects the states of intestines and is one of the important indices of clinic diagnosis for intestinal diseases.Under the normal and pathological states,the intestines of human body can general sounds with different characters for the different movements of intestines.When the body takes intestine illnesses or with intestinal obstruction especially, the times of bowel sounds increase, the volume add and the tone rise, we can usually detect abnormal sounds such as metal sound or sharp sound. If we distinguish the sounds accurately, the results will provide crucial diagnosis information for clinic. Although bowel sounds study was well underway at home and abroad, the clinical applications of bowel sounds diagnosis are restricted largely by the characteristics of weak signal, complicated noise, no periodicity and randomicity. For the moment, the doctors are using the traditional stethoscope to detect the bowel sound in the clinic all the same.To overcome the disadvantages in the bowel sounds detection such as strong subjectivity, low exactitude and unscientific detections, we design a multi-channel bowel sounds acquisition system based on C8051F340 MCU to detect the signals accurately.The system includes hardware and software, the hardware system comprise multi-channel microphones, amplifier circuit, filter circuit, voltage ascend circuit,MCU circuit,electricity supply circuit and so on, the software system is composed of the downer computer programs and the upper computer programs. The system works as follows, microphones detect the bowel sounds firstly and transform them into electrical signals, and then the signals will be amplified, filtered and advance the voltage level to satisfy the level requested by A/D converter. After the signals come into the MCU circuit, the A/D converter embedded in the MCU transform the analog signals into digital ones, then the digtal signals will be transmitted to the upper computer and be displayed, memorized for further analysis at the same time.When the acquisition system debug was done, we got signals from healthy person and patient with intestinal obstruction apart with the system, and then analyzed them contrastively. The signal analyses include the noise removing, feature extractions and automatic classifications and so on. In this paper, we select the method of adaptive noise cancellation and ICA to eliminate the noise merged in the main signals respectively. The adaptive noise cancellation can get rid of the environmental noise triumphantly and the ICA is superior in the blind source separation. To extract the features of bowel sounds effectively, we discuss several methods including wavelet, normalized average Shannon energy and Welch PSD, all of which are promising for feature extractions and automatic classifications of bowel sounds. The qualitative and quantitative results show the features of vary kinds of bowel sounds have statistical difference. It will be prospective to distinguish the different kinds of bowel sounds using the numerical value and proper classification methods.Detection and analysis of bowel sounds have important value in the diagnosis and cure of gastrointestinal diseases, the experimental results demonstrate that the detection system can get the signal effectively and the presented algorithms have potential applications in detecting various intestinal diseases.