Deep learning based road crack detection device using black box image and road crack detection method and computer program for executing the same

摘要

The present invention relates to a deep learning-based road crack detection device and a road crack detection method for detecting road cracks using a vehicle black box image, A road crack detection apparatus according to an embodiment of the present invention includes: a data collection unit that is photographed by a vehicle black box and collects a plurality of image data including location information; an image extracting unit for extracting a plurality of reference images based on the moving distance information of the vehicle calculated by using the location information included in the collected plurality of image data; A patch image generation unit that selects the reference image extracted from the image extraction unit in units of pixels of a certain size to classify classes for each of road cracks, lanes, and undamaged roads, and generates a patch image for each of the classified classes ; a model training unit for training each patch image generated by the patch image generation unit to extract road cracks using a neural network model; And, when a plurality of image data including location information is input from an arbitrary vehicle black box, a plurality of reference frames are generated based on the vehicle moving distance information calculated using the location information included in the input plurality of image data. and a crack detection unit that extracts, extracts an image of a road surface from the plurality of extracted reference frames, and detects road cracks by applying a neural network model trained from the model training unit to the extracted image.