Correction visibility estimation method based on three-spectrum real-time imaging and polynomial regression

Abstract

The invention discloses a correction visibility estimation method based on three-spectrum real-time imaging and polynomial regression, and belongs to the technical field of atmospheric visibility detection, wherein the method comprises the steps: constructing a three-spectrum real-time imaging optical system by using three near-infrared cameras, three spectroscopes, three optical filters corresponding to near-infrared imaging wavebands and three lens groups; enabling a three-spectrum real-time imaging system to image a selected building target in a foggy day to obtain three near-infrared images imaged in different wavebands in the same scene; further deriving a visibility estimation formula based on near-infrared three-spectrum imaging and distance in the foggy weather; and obtaining multiple pairs of standard visibility U and estimated visibility V, then, establishing a visibility correction formula by using a polynomial regression equation, and correcting the subsequently obtained estimated visibility V to obtain corrected visibility Vx. The method utilizes three-spectrum real-time imaging and polynomial regression to overcome the defect that in the prior art, a large number of training samples are needed to optimize the performance of the neural network, so that the method is low in visibility measurement error in the foggy weather.

Description

technical field [0001] The invention belongs to the technical field of atmospheric visibility detection, and in particular relates to a modified visibility estimation method based on three-spectrum real-time imaging and polynomial regression. Background technique [0002] Visibility refers to the maximum distance that a person with normal vision can recognize the target object from the background, and the unit is usually meters. The main factors affecting visibility are fog and haze. When visibility is low, it will bring danger to driving safety and aviation and other fields. With the advancement of technology, visibility is becoming more and more important in the fields of life, military, environment and climate. Therefore, how to estimate visibility easily and accurately is an important research direction. [0003] The patent "Urban Building Relative Distance Estimation Method Based on Dual Spectral Extinction under Fog" (Patent Application No. 202010292894.1, Applicati...

AI Technical Summary

Problems solved by technology

The disadvantages of this method are: a large number of training samples are needed to optimize the performance of the neural network, and there is still a large error in predicting visibility using the linear regression model

Images