Research On ECG Data Extraction Based On OpenCV

With more and more attention to heart disease in people’s mind, the traditional medical diagnosis in ECG (electrocardiogram) can’t satisfy people’s demands. Image processing to extract the data message and making remote intelligent diagnosis has became a tend gradually. The widespread popularity of smartphones creats the basic hardware conditions for the image acquisition and upload. Paper ECG is a valuable asset of the raw data, and it’s an important premise to extract the ECG curve image from the ECG paper. So how to get accurate ECG data quickly attracts more and more attention and becomes a hot research topic gradually.In order to reduce the error in the process of ECG acquisition, Scratchable latex for mobile phones was used in this paper. Image tilt problem exists in the process of photographing with mobile phones. To solve the problem, Hough transformation which includes Radon transformation is used to complete the tilt correction based on line detection principle. And the principle of the corner detection and the rectangle detection is introduced, which is the basis for solving the irregularity distortion of the ECG.It is necessary to get ECG curve for extracting the data message from ECG. This paper proposed a genetic image segmentation algorithm based on Otsu which is in terms of the principle of threshold segmentation to complete separation of ECG curve with background grid fast and accurately. Effective refinement of the ECG curve can be achieved with the combination of skeleton algorithms and mathematical morphology theory based on a principle that a cross-pixel corresponds to vertical pixel when extracting curve data. To solve the errors caused by irregularity distortion when photographing, grid distortion correction can be used to recover the ECG curve irregularity distortion correction using ratio method.In this paper, a series of image processing researches are carried out in OpenCV programming environment based on VC++6.0.The main task is to make a research about the problems in collecting and coordinate extracting from ECG, to make a scientific extraction of ECG information and to achieve certain results of this study.