ADS-B track and video fusion scene aircraft detection method and device

By constructing an alignment and registration framework for the integration of ADS-B trajectory and video in remote and medium-sized airports, combining the conversion alignment and registration of cross-modal data, the problem of difficulty in identifying target aircraft by simple video surveillance is solved, and the accurate detection of targets in wide-area large-scene videos is achieved, and the support capabilities of automatic monitoring and control of airport scenes is enhanced.

CN119723958AActive Publication Date: 2025-03-28BEIHANG UNIV
12 Cites 1 Cited by

Patent Information

Application Number
CN202411660612.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-20
Publication Date
2025-03-28
Estimated Expiration
2044-11-20

AI Technical Summary

Technical Problem

In remote small and medium-sized airports, it is difficult to accurately identify target aircraft simply by relying on video surveillance, especially when the aircraft is located in relatively remote locations, resulting in insufficient awareness of scene safety situations and inefficient regulatory operation guarantees.

Method used

By constructing an alignment and registration framework for ADS-B trajectory and video fusion, combining the conversion alignment and registration of cross-modal data, a feature set for cross-modal cross-attention is constructed, and a fusion model based on attention mechanism is designed to detect targets that cannot be recognized by a single image monitoring source.

Benefits of technology

It realizes accurate detection of targets that cannot be identified by single surveillance sources in wide-area large-scene airport surveillance videos, and enhances the support capabilities of automatic surveillance and control of airport scenes.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

According to the ADS-B track and video fusion scene aircraft detection method and device, complementarity of different monitoring sources can be fully utilized, targets which cannot be recognized by a single monitoring source in a wide-area large-scene airport monitoring video can be accurately detected, and powerful support is provided for subsequent automatic monitoring and control of an airport scene. The method comprises the steps that (1) conversion alignment and registration facing airport cross-modal data are carried out, an alignment and registration framework is constructed based on associated key indexes of three data sources including airport scene monitoring ADS-B, monitoring videos and flight plans, and the framework comprises initial screening matching pair construction, rough alignment fragment acquisition and frame-by-frame target point alignment; (2) according to the registered cross-modal data, constructing a feature set oriented to cross-modal cross attention, and designing a fusion model based on an attention mechanism; and (3) inputting a cross-mode fusion result into a target detection model, wherein the target detection model detects a target which cannot be recognized by a single image monitoring source.
Need to check novelty before this filing date? Find Prior Art