Nonlinear atmospheric correction based on neural network for infrared target radiometry

Guoqing, 杨国庆 Yang; Zhou, 李周 Li; Chen, 赵晨 Zhao; Yi, 余毅 Yu; Yan-feng, 乔彦峰 Qiao; Fengyun, 何锋赟 He

doi:10.3788/irla.24_2019-0413

Cited by 2 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…11 Furthermore, Yang Guoqing adopted the neural network method to correct the model calculation values of atmospheric transmittance and path radiation nonlinearly; lastly, the measurement error of infrared radiation was reduced to 6.45%. 12 However, the blackbodies adopted to simulate reference and target in the mentioned methods were largely small in size, and the experiments were performed under the laboratory conditions of the horizontal atmospheric environment, covering the limitations of short observation distances and low verification temperatures. In practical tasks, the target observed is generally long-distance and high-temperature.…”

Section: Introductionmentioning

confidence: 99%

Analysis on the correction method of atmospheric transmittance close to the ground

Ge²,

Li³

et al. 2023

AOPC 2022: Atmospheric and Environmental Optics

View full text Add to dashboard Cite

Atmospheric transmittance can critically affect the accuracy of measuring infrared characteristics exhibited by targets. On the whole, the existing measurement of atmospheric transmittance has complied with the engineering calculation results of the MODTRAN software by applying several vital parameters (e.g., temperatures, air pressures and water-vapor content). In general, the error of such a method exceeds 20%, and it is significantly impacted by local weather. In this study, a ratio correction method was adopted to decrease the error in measuring atmospheric transmittance. The correction factor was determined by comparing the directly measured value from the infrared images of reference blackbody at different temperatures with the calculated value of the MODTRAN. Subsequently, the correction factor could be exploited to correct atmospheric transmittance. The experiment for measuring infrared radiation was performed, and the radiance inversion error was reduced by more than 10% after the correction of atmospheric transmittance. Furthermore, the correction factor calculated from LWIR images could be extrapolated to other bands. Besides, the inversion accuracy of the infrared radiation characteristics significantly increased. Thus, the multi-band applicability of the correction method was verified.

show abstract

Section: Introductionmentioning

confidence: 99%