V2V METHOD FOR PROVIDING ROBUST OBJECT DISTANCE ESTIMATION BASED ON CAMERA BY PERFORMING PITCH CALIBRATION OF CAMERA MORE PRECISELY WITH FUSION OF INFORMATION ACQUIRED THROUGH CAMERA AND INFORMATION ACQUIRED THROUGH V2V COMMUNICATION AND DEVICE USING THE SAME

摘要

Disclosed is a method of improving the accuracy of object distance prediction based on the target camera by more precisely performing pitch calibration of a target camera using additional information obtained through vehicle to vehicle (V2V) communication. That is, (a) the computing device causes (i) a process for causing an initial pitch calibration module to apply a pitch calculation operation to the reference image to generate an initial predicted pitch and (ii) an object detection network to: applying a neural network operation to the reference image to perform a process of generating reference object detection information; and (b) the computing device causes the adjusted pitch calibration module to (i) select a target object, (ii) calculate a predicted target height of the target object, and (iii) determine the corresponding initial predicted pitch. after calculating the error, (iv) using the error to determine an adjusted prediction pitch for the target camera.