Abstract
Through analyzing the annual volatility of cloud detection thresholds and integrating the band characteristics of HJ-1B,the authors proposed a new algorithm of cloud detection for HJ-1B image based on dynamical thresholding according to the spectral standard deviation anomaly.Using image registration,band math,linear regression and error analysis,the authors acquired the cloud abnormal regions which could be used to remove cloud pixels from the image.The results show that this approach can detect cloud pixels over different periods and in different scenes successfully,thus promoting the use of HJ-1B data and improving the precision of image classification.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
