The Landsat 7 Automated Cloud Cover Assessment (ACCA) algorithm recognizes clouds by passing through the scene data twice (Irish, 2000). This two pass approach differs from the single pass algorithm employed for Landsat-5. The algorithm is based on the premise that clouds are colder than Earth surface features. While true in most cases, temperature inversions do occur. Unexpected cloud cover calculations will occur in these situations.

The first pass through the data is designed to trap clouds and only clouds. Twenty-six different filters are deployed for this purpose. Omission errors are expected. The pass one goal is to develop a reliable cloud signature for use in pass two where the remaining clouds are identified. Commission errors, however, create algorithm havoc and must be minimized. Three class categories result from pass one processing - clouds, non-clouds, and an ambiguous group that is revisited in pass two.

After pass one processing, descriptive statistics are calculated from the cloud category using band 6. These include mean temperature, standard deviation, and distribution skew. New band 6 thresholds are developed from these statistics for use during pass two. Only the thermal band is examined during this pass to capture the remaining clouds. Image pixels that fall below the new threshold qualify. After processing, the pass one and two cloud cover scores are compared. Extreme differences are indicative of cloud signature corruption. When this occurs, the pass two results are ignored and the cloud score reverts to the pass one result.

During processing, a cloud cover mask is created. After the two passes through the data, a filter is applied to the mask to fill in cloud holes. This filtering operation works by examining each non-cloud pixel in the mask. If 5 of the 8 neighbors are clouds then the pixel is reclassified as cloud. The final cloud cover percentage for the image is calculated based on the filtered cloud mask.

The images in Figure 12.3 demonstrate an example of the improved power the Landsat 7 ACCA algorithm has over the Landsat 5 algorithm. In this case, clouds are successfully separated from the desert terrain below.

	This desert area in Sudan (upper left, bands 5, 4, 2) represents terrain where the heritage algorithm misclassified rocks as clouds due to their high reflectance in band 5. This can be viewed in the cloud mask to the left. Above is the cloud mask generated using the Landsat 7 algorithm. Here the clouds are captured while the rocks are recognized for what they are.
Figure 12.3 Landsat 5 and 7 Cloud Cover Mask Comparison

Back to Chapter 12

Last Update: July 24, 2006