WATER LEVEL DETECTION METHOD USING K-MEANS CLUSTERING

摘要

PURPOSE: A river stage measuring method using a K-means clustering and a system thereof are provided to accurately measure a river stage under various environments and conditions and to change selectively change the conditions as occasion demands. CONSTITUTION: A river stage measuring method using a K-means clustering comprises next steps. Images of a boundary region of a structure and a river are photographed by using a camera. The image data of the images is collected by an image processing device(S100). A motion vector is generated by using an optical flow referred to the image data collected by the image processing device and a size of the motion vector is accumulated so that a accumulated motion histogram is formed(S200). The accumulated histogram is K-means grouped(S300). The river stage is calculated by using the K-means grouped histogram(S400). [Reference numerals] (AA) Start; (BB) End; (S100) Collecting the image structure-river of a boundary region of a structure and a river; (S200) Generating a motion accumulated histogram by using the image data; (S300) K-means clustering using the motion accumulated histogram; (S400) Calculating a water level of the river by using the clustered histogram