SYSTEM AND METHOD FOR POTHOLE DETECTION BASED ON ROAD AERIAL IMAGES

摘要

The present invention relates to a system and method for detecting a pothole based on a road image, and more particularly, to a system and method for detecting a pothole using deep learning to label an aerial image for each tile and extract a pothole from a tile classified as a pothole. will be. The present invention provides a framework unit for receiving a first image in which location information is stored from a user and displaying the first image and a map image according to the location information on a main page; a machine learning unit that divides the first image into tiles that are a plurality of preset second images, and classifies each tile by assigning a label in the form of a location through machine learning by machine learning; and a storage unit for storing the first image, the second image, and the tile to which the label is given, wherein a second image including the tile in which the pothole is detected from the first image is provided to the user It relates to a porthole detection system based on road images. According to the present invention having the configuration as described above, the detection speed is faster than other systems by applying video processing and machine learning technology, and it is based on a hybrid detection system, and there is an advantage of detecting a pothole in an aerial image of a road.