METHOD FOR CREATING OBSTRUCTION DETECTION MODEL USING DEEP LEARNING IMAGE RECOGNITION AND APPARATUS THEREOF

摘要

Disclosed is a method and apparatus for generating an obstacle prediction model using deep learning image recognition of the present invention. The method for generating an obstacle prediction model according to an embodiment of the present invention includes acquiring an orthogonal image of land photographed from the air, detecting obstacles in the acquired image, and classifying the obstacles into predetermined categories and learning It is configured to include the steps of constructing data and generating an obstacle prediction model to which semantic segmentation is applied based on the learning data.