Joyce Fry

Automated cloud and shadow detection and filling using two-date Landsat imagery in the USA

A simple, efficient, and practical approach for detecting cloud and shadow areas in satellite imagery and restoring them with clean pixel values has been developed. Cloud and shadow areas are detected using spectral information from the blue, shortwave infrared, and thermal infrared bands of Landsat Thematic Mapper or Enhanced Thematic Mapper Plus imagery from two dates (a target image and a reference image). These detected cloud and shadow areas are further refined using an integration process and a false shadow removal process according to the geometric relationship between cloud and shadow. Cloud and shadow filling is based on the concept of the Spectral Similarity Group (SSG), which uses the reference image to find similar alternative pixels in the target image to serve as replacement values for restored areas. Pixels are considered to belong to one SSG if the pixel values from Landsat bands 3, 4, and 5 in the reference image are within the same spectral ranges. This new approach was applied to five Landsat path/rows across different landscapes and seasons with various types of cloud patterns. Results show that almost all of the clouds were captured with minimal commission errors, and shadows were detected reasonably well. Among five test scenes, the lowest producers accuracy of cloud detection was 93.9% and the lowest users accuracy was 89%. The overall cloud and shadow detection accuracy ranged from 83.6% to 99.3%. The pixel-filling approach resulted in a new cloud-free image that appears seamless and spatially continuous despite differences in phenology between the target and reference images. Our methods offer a straightforward and robust approach for preparing images for the new 2011 National Land Cover Database production.

Quantifying Urban Land Cover Change Between 2001 and 2006 in the Gulf of Mexico Region

We estimated urbanization rates (2001–2006) in the Gulf of Mexico region using the National Land Cover Database (NLCD) 2001 and 2006 impervious surface products. An improved method was used to update the NLCD impervious surface product in 2006 and associated land cover transition between 2001 and 2006. Our estimation reveals that impervious surface increased 416 km2 with a growth rate of 5.8% between 2001 and 2006. Approximately 1110.1 km2 of non-urban lands were converted into urban land, resulting in a 3.2% increase in the region. Hay/pasture, woody wetland, and evergreen forest represented the three most common land cover classes that transitioned to urban. Among these land cover transitions, more than 50% of the urbanization occurred within 50 km of the coast. Our analysis shows that the close-to-coast land cover transition trend, especially within 10 km off the coast, potentially imposes substantial long-term impacts on regional landscape and ecological conditions.

An efficient method for change detection of soil, vegetation and water in the Northern Gulf of Mexico wetland ecosystem

Mapping and monitoring wetland ecosystems over large geographic areas based on remote sensing is challenging because of the spatial and spectral complexities of the inherent ecosystem dynamics. The main objective of this research was to develop and evaluate a new method for detecting and quantifying wetland changes in the Northern Gulf of Mexico (NGOM) region using multitemporal, multispectral, and multisensor remotely sensed data. The abundance of three land- cover types (water, vegetation, and soil) was quantified for each Landsat 30 m pixel for 1987, 2004, 2005, and 2006 using a regression tree algorithm. The performance of the algorithm was evaluated using an independent reference data set derived from a high-resolution QuickBird image, and several statistics including average error (AE), relative error (RE), and the Pearson correlation coefficient (r). For per-pixel percentage estimation, the AE is under 10% for water prediction, 9.5–11.4% for vegetation, and 9–11.1% for soil. The correlation coefficients between predicted and reference data range from 0.90 to 0.96 for water, from 0.80 to 0.89 for vegetation, and from 0.79 to 0.86 for soil. The high accuracy achieved by this method is attributed to the high quality of training data and the rigorous calibrations applied to multisensor and multitemporal satellite imagery. Based on the multitemporal estimation of the three land-cover components, spatial and temporal changes of the land-cover types from 1987 to 2006 were quantified and analysed. The study demonstrates that the method provided useful information on the abundance and changes of the key land-cover types in the NGOM region where long-term disturbances and episodic events occurred. Such information is valuable for monitoring land and vegetation loss and recovery processes, and for understanding possible drivers of the coastal wetland evolution in the region.