A method for cloud detection and opacity classification based on ground based sky imagery

Abstract

Digital images of the sky obtained using a total sky imager (TSI) are classified pixel by pixel into clear sky, optically thin and optically thick clouds. A new classification algorithm was developed that compares the pixel red-blue ratio (RBR) to the RBR of a clear sky library (CSL) generated from images captured on clear days. The difference, rather than the ratio, between pixel RBR and CSL RBR resulted in more accurate cloud classification. High correlation between TSI image RBR and aerosol optical depth (AOD) measured by an AERONET photometer was observed and motivated the addition of a haze correction factor (HCF) to the classification model to account for variations in AOD. Thresholds for clear and thick clouds were chosen based on a training image set and validated with set of manually annotated images. Misclassifications of clear and thick clouds into the opposite category were less than 1%. Thin clouds were classified with an accuracy of 60%. Accurate cloud detection and opacity classification techniques will improve the accuracy of short-term solar power forecasting

3D cloud detection and tracking system for solar forecast using multiple sky imagers

“…Such sky images have been successfully used to estimate observed cloud cover over a short-time horizon (Long et al, 2001;Morris, 2005;Pfister et al, 2003). By analyzing the red-green-blue (RGB) channels in sky images, researchers can determine the details of various cloud properties, such as opaqueness and thickness (Shields et al, 1993;Ghonima et al, 2012;Long et al, 2006). In addition to utilizing original RGB channels, Souza-Echer et al proposed detecting clouds in the hue-saturation-luminance (HSL) color space (Souza-Echer et al, 2006).…”

Real-time forecasting of solar irradiance ramps with smart image processing

“…Secondly, if the image is categorized as cloudy, the hybrid thresholding method (Li et al, 2011) is used to analyze red blue ratio histograms and further categorize the image as either overcast or partly cloudy. Thirdly, after the image categorization, the SACI employs Fixed threshold method (Li et al, 2011) for overcast images, clear sky library method with fixed threshold (Ghonima et al, 2012) for clear images, and clear sky library method with adaptive threshold (Otsu, 1979;Li and Lee, 1993;Li and Tam, 1998) for partly cloudy images. Examples of SACI cloud detection are shown in Fig.…”

“…Forecast models that consider cloud information are developed based on either remote sensing or local sensing. Remote sensing based models (Hammer et al, 1999;Perez et al, 2010;Marquez et al, 2013b; that use satellite images have limited spatial and temporal resolutions and therefore not appropriate for intra-hour forecast (Ghonima et al, 2012;Marquez and Coimbra, 2013;Chu et al, 2013). To perform the intra-hour hour forecast, local sensing based models are developed using high resolution local sky images captured by high-frequency sky imagers (Chow et al, 2011;Marquez et al, 2013a;Marquez and Coimbra, 2013;Urquhart et al, 2013;Quesada-Ruiz et al, 2014).…”

Streamline-based method for intra-day solar forecasting through remote sensing

“…A detailed review of current solar forecasting methods can be found in Inman et al (2013). While short term forecasts (intra-hour) became more sophisticated based on recent advances in ground based sky-imagery and modeling techniques (Chow et al, 2011;Ghonima et al, 2012;Marquez and Coimbra, 2013a;Dong et al, 2013;Handa et al, 2014), intra-day methods still lack accuracy. Several previous studies propose and evaluate methods for intra-day forecasts (Mathiesen and Kleissl, 2011;Marquez et al, 2013).…”