Recently, I needed to order some imagery over an area in Bolivia. I searched a few provider databases and discovered that GeoEye had the best coverage in my area. Below, my area of interest is shown by the red square. The blue boxes outline the images in my area, and the yellow, highlighted ones are the 4 images that were the best choices—they overlapped my area, had the least cloud cover, and were acquired within the last few years.
When I got my images, they were delivered as 7 images. The top grouping contained 4 images and the bottom grouping contained 3 images. What happened was, the two scenes on the top and the two on the bottom were actually each collected at the same time. To make the data small enough to be a reasonable size for FTP, they were tiled.
As I continue in the next part, I’m just going to discuss the top row (top 4 tiles). The first thing I did was open the Catalog window and examine all the images. I draged the Multispectral product for each scene from the Catalog window to display it in ArcMap.
I ended up with the following image.
I checked the metadata to be sure all these tiles were the same scene. You can do this by reading the text or xml files that come with the data, or open the layer properties in ArcMap (or right-click the raster product in the Catalog window) and click the Key Metadata tab. I verified that each scene was taken on the same date and time, so I know they’re all part of one scene.
Next, I selected all 4 images in the Image Analysis window and clicked the Mosaic button.
This created a fifth layer that combines the four selected layers (I can remove the selected four layers). The resultant layer now behaves as one image.
Of course I wanted to get rid of the black areas on the edges. So I selected the layer in the Image Analysis window and clicked the Add Function button.
This opens the Raster Function Editor, where I clicked the top-most function and inserted the Mask function.
In the Mask function, I set the NoData Interpretation value to All – so all pixels must be the same to mask it out–and I entered 0 for each band.
This gave me a new layer with the black area masked out.
From here, I saved the layer so I don’t have to go through all these steps again and continued to examine my new imagery.