TY - JOUR
T1 - Modeling and detection of heat haze in computer vision based displacement measurement
AU - Luo, Longxi
AU - Feng, Maria Q.
AU - Wu, Jianping
AU - Bi, Luzheng
N1 - Publisher Copyright:
© 2021
PY - 2021/9
Y1 - 2021/9
N2 - Computer vision has become widely applied for structural displacement monitoring. However, heat haze is one of the major challenges. Image distortions caused by heat haze in hot weather can result in displacement errors. Therefore, a comprehensive study of properties of heat haze-induced distortions and displacement errors is conducted. Firstly, an image distortion estimation method is proposed for estimating heat haze-induced image distortions. Secondly, displacement errors due to heat haze are analyzed. A heat haze error model is formulated to describe the properties of heat haze errors, and the explicit effect of the environmental factor of temperature on the heat haze error model. Thirdly, a heat haze detection method is proposed to enable detection of heat haze's influence on vision-based displacement sensors by extracting features from distortion measurements and applying a classification algorithm. Field tests in hot weather and experiments with dark heaters for introducing heat haze are conducted for validations.
AB - Computer vision has become widely applied for structural displacement monitoring. However, heat haze is one of the major challenges. Image distortions caused by heat haze in hot weather can result in displacement errors. Therefore, a comprehensive study of properties of heat haze-induced distortions and displacement errors is conducted. Firstly, an image distortion estimation method is proposed for estimating heat haze-induced image distortions. Secondly, displacement errors due to heat haze are analyzed. A heat haze error model is formulated to describe the properties of heat haze errors, and the explicit effect of the environmental factor of temperature on the heat haze error model. Thirdly, a heat haze detection method is proposed to enable detection of heat haze's influence on vision-based displacement sensors by extracting features from distortion measurements and applying a classification algorithm. Field tests in hot weather and experiments with dark heaters for introducing heat haze are conducted for validations.
KW - Computer vision
KW - Heat haze detection
KW - Heat haze error model
KW - Image distortion estimation
KW - Structural displacement measurement
UR - http://www.scopus.com/inward/record.url?scp=85108316298&partnerID=8YFLogxK
U2 - 10.1016/j.measurement.2021.109772
DO - 10.1016/j.measurement.2021.109772
M3 - Article
AN - SCOPUS:85108316298
SN - 0263-2241
VL - 182
JO - Measurement: Journal of the International Measurement Confederation
JF - Measurement: Journal of the International Measurement Confederation
M1 - 109772
ER -