A method for providing a camera-based robust object distance prediction by more precisely performing pitch calibration of a camera by fusing information acquired via V2V communication with information acquired via a camera, and the same. Equipment using

摘要

PROBLEM TO BE SOLVED: To provide a method for accurately predicting a distance even when the quality of an input image is not good. A method of using additional information acquired via V2V communication for pitch calibration, wherein an initial pitch calibration module of a computing device applies a pitch calculation operation to a reference image to perform an initial prediction. A process of generating a pitch and a process of generating a reference object detection information by applying a neural network operation to a reference image with an object detection network, an adjustment pitch calibration module selects a target object, and a target object After calculating the predicted target height of, the error corresponding to the initial predicted pitch is calculated, and the adjusted predicted pitch for the target camera is determined using the error. [Selection diagram] Figure 2