	US 12,260,656 B2
	Method and apparatus for estimating size of damage in the disaster affected areas
Seong Sam Kim, Ulsan (KR); Dong Yoon Shin, Ulsan (KR); and Cheol Kyu Lee, Ulsan (KR)
Assigned to NATIONAL DISASTER MANAGEMENT RESEARCH INSTITUTE, Ulsan (KR)
Filed by NATIONAL DISASTER MANAGEMENT RESEARCH INSTITUTE, Ulsan (KR)
Filed on May 4, 2022, as Appl. No. 17/736,332.
Claims priority of application No. 10-2021-0057837 (KR), filed on May 4, 2021.
Prior Publication US 2022/0358774 A1, Nov. 10, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 20/70 (2022.01); G06T 7/00 (2017.01); G06T 7/60 (2017.01); G06V 10/764 (2022.01); G06V 10/778 (2022.01); G06V 20/10 (2022.01)

CPC G06V 20/70 (2022.01) [G06T 7/0002 (2013.01); G06T 7/60 (2013.01); G06V 10/764 (2022.01); G06V 10/7784 (2022.01); G06V 20/10 (2022.01); G06T 2207/10032 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/30181 (2013.01)]

16 Claims

1. An operation method of a server for estimating a size of damage in disaster affected areas, the operation method comprising:

acquiring at least one first disaster image;

deriving a disaster affected area from each of the at least one first disaster image, and acquiring affected area related information through labeling based on the derived disaster affected area;

training a first learning model using the at least one first disaster image and the affected area related information;

estimating damage size information of the disaster affected area in the first disaster image based on the first learning model;

acquiring a second disaster image from an external device;

deriving a disaster affected area and a disaster prone area from the second disaster image;

acquiring a plurality of disaster related information through labeling based on the derived disaster affected area and the derived disaster prone area;

assigning a weight to each of the plurality of acquired disaster related information;

inputting the second disaster image and the plurality of disaster related information to the trained first learning model;

outputting disaster affected area identification information and disaster damage type information based on the first learning model;

acquiring disaster affected area use information;

identifying an area of the disaster affected area, a type of the disaster affected area and a use of the disaster affected area based on the disaster affected area identification information, the disaster damage type information and the disaster affected area use information;

deriving disaster affected area feature information based on the identified area of the disaster affected area, the identified type of the disaster affected area and the identified use of the disaster affected area; and

identifying the damage size information based on the derived disaster affected area feature information.

10. A server for estimating a size of damage in disaster affected areas, the server comprising:

a transmitter/receiver which communicates with an external device; and

a processor to control the transmitter/receiver,

wherein the processor is configured to:

acquire at least one first disaster image,

derive a disaster affected area from each of the at least one first disaster image, and acquire affected area related information through labeling based on the derived disaster affected area,

train a first learning model using the at least one first disaster image and the affected area related information,

estimate damage size information of the disaster affected area in the first disaster image based on the first learning model,

acquire a second disaster image from the external device,

derive a disaster affected area and a disaster prone area from the second disaster image,

acquire a plurality of disaster related information through labeling based on the derived disaster affected area and the derived disaster prone area,

assign a weight to each of the plurality of acquired disaster related information,

input the second disaster image and the plurality of disaster related information to the trained first learning model,

output disaster affected area identification information and disaster damage type information based on the first learning model,

acquire disaster affected area use information,

identify an area of the disaster affected area, a type of the disaster affected area and a use of the disaster affected area based on the disaster affected area identification information, the disaster damage type information and the disaster affected area use information,

derive disaster affected area feature information based on the identified area of the disaster affected area, the identified type of the disaster affected area and the identified use of the disaster affected area, and

identify the damage size information based on the derived disaster affected area feature information.