Recognition Method Research Of The Water Bridge Based On Image Segmentation

Automatic target recognition is a challenging task in the field of computer vision and artificial intelligence. In recent years, the automatic identification of man-made objects in natural scenery has brought many scholars’ extensive concern, and it has important application value in the field of military and civilian. Bridge is a typical representative of man-made objects, it is an important kind of artificial buildings, and it is also an important traffic artery. The automatic identification for bridge has become a hot research focus, but there is still no widely applicable methods.This paper proposes a rapid bridge recognition method based on Otsu image segmentation on the research of bridge recognition method based on Mean Shift image segmentation.In this paper, the research contents are summarized as follows:(1) The research to the basic principle and theory of the bridge recognition method. The bridge recognition method based on Mean Shift image segmentation is studied. Firstly, the image is segmented by Mean Shift algorithm. Secondly, the segmentation result is transformed into binary image, then the river area is extracted based on clustering. Thirdly, the connected river area is extracted by applying morphological operations on river area image. Fourthly, by comparing the river area image and connected river area image, the candidate bridges are extracted. Finally, identifying the location of the bridges according to the intersections between river center line and bridges, then identify the bridges.(2) This paper proposes a kind of rapid bridge recognition method based on Otsu image segmentation after improving the above method. Firstly, the color image is gray and segmented based on Otsu threshold. Secondly, the water area image is acquired after applying dilation and erosion operations on segmentation image, then the connected water area image is acquired after applying erosion and dilation operations on water area image. Thirdly, by comparing the water area image and connected water area image, then extract the bridges. Finally, positioning the location of the bridges, then the bridges have been identified. By comparing two methods on the recognition effect and recognition time respectively, and the analysis according to the experimental results are proposed. The experimental results show that the improved method is faster and more practical than the former.