Multi-scale air-ground parameter automatic calibration method based on monocular vision system

Abstract

The invention provides a small aircraft multi-scale air-ground parameter automatic calibration method based on a monocular vision system. A self-made calibration board formed by splicing dark paper boards serves as a calibration tool, through a series of processing algorithms such as binarization, edge extraction, contour searching, contour minimum enclosing rectangle calculation, and noise filtering, feature points belonging to the calibration board are extracted from acquired image data, the ratio of an actual distance to a pixel number is obtained, the points are connected by using a smooth curve when the flying height of the aircraft measured by a barometer is combined, and finally, an air-ground multi-parameter relation function curve is obtained. The innovation point of the invention is that the dark paper boards are adopted as the calibration device, online calibration on the flying height of the aircraft, ground object geometric parameters and the image resolution which are three correlated air-ground parameters can be quickly realized, and the foundation is laid for information fusion of multiple sensors.

Description

technical field [0001] The invention belongs to the technical field of autonomous environment perception of drones, and relates to a multi-scale air-ground parameter automatic calibration method for a small aircraft based on an airborne monocular vision system. Background technique [0002] The visual sensor is one of the main devices for UAVs to realize environmental perception, and it has the function of collecting images in real time. UAVs can realize aerial photography, target detection, tracking, and scene understanding based on airborne monocular vision. To realize the above functions of UAVs, automatic calibration of multi-scale air-ground parameters based on monocular vision must first be completed. Literature (WangIH, DobrokhodovVN, KaminerII, et al. "Onvision-based target tracking and range estimation for small UAVs". Proceedings of AIAA Guidance, Navigation and Control Conference. 2005) proposed a real-time detection and tracking algorithm for ground moving target...

AI Technical Summary

Problems solved by technology

This method can accurately extract the spatial parameters of the target object, but due to the large mass and low cost performance of the binocular vision system, it is not suitable for small near-ground aircraft with limited load

Images