ENVIRONMENTAL INFORMATICS GEOINFORMATION PRODUCTS		BROCKMANN CONSULT
	BEAM Java Tutorial
	INDEX\| SET-UP\| EX 1\| EX 2\| EX 3a\| EX 3b\| EX 4\| OUTLOOK

Exercise 3a

How to create a new data product and write it as BEAM-DIMAP format

Create a new product object

Create a new class Ex3 in the package org.esa.beam.basics. If you forgot how to create a new class in IDEA, step back to exercise 1. This time we start with a new instance of the Product class directly:

We create a product with a scene raster size of 512 times 512 pixels. We name the product product_ex3a, set the new product type EX3A. The product instance is assigned to the local vaiable product:

Create a new band and add it to the product

After we've created the product instance, we now create a new instance of the Band class. Band also must have a name, a data type and the scene raster dimension. Note that all bands in a product must have the same raster size as specified in the product's constructor:

We give the band the name bame_ex3a and set its data type to 32-bit floating point. Then the band is added to the product:

Set the band's raster data with computed pixel values

The are a few methods to set the pixel values of the band. We now use the setRasterData method, this is the direct approach which sets the entire raster data buffer with a single call:

The setRasterData method expect a parameter of type ProductData:

The ProductData class has some handy factory methods, we use the one which takes an array of 32-bit floating point values:

Since the setRasterData method want to set the entire raster data of the band, we have to size the array to contain width times hight elements:

The array of floats is the array we want to fill with data, we therefore introduce the variable pixels (Ctrl+Alt+V) from our statement, so that we can assign values to each element:

We now loop through all pixels of the raster and set an artificial value. In this exercise it does not matter which value you assign, I've simple choosed a formula which will produce an image I understand. From exercise Exercise 2 we learned that BEAM uses flat, row-order arrays, that's why I use the expression i = y * w + x again to compute the raster index:

Write the data product as BEAM-DIMAP format

As you might have expected, the ProductIO class provides a counterpart to the readProduct method. Not very surprising, its name is writeProduct. There are a few variants, we use here the one which takes a product object, a file path and the format name. As file path we use the product name plus the extension .dim because our output format will be BEAM-DIMAP:

Again, an IOException has to be handled, and finally with dispose the product object in order to close underlying I/O streams and to get rid of allocated resources associated with the product instance:

Inspect data product output and open it in VISAT

Run the program (Shift+F10). If no errors occured, the program should have written the BEAM-DIMAP product product_ex3a.dim to your IDE project directory. Now start VISAT from the start menu. In VISAT, goto the File menu and select Open. Point the file selection dialog to your project directory and select the product product_ex3a.dim. Double-click the band band_ex3a: in the Product View in order to open an image view for it:

Ok, we must confess this has not much to do with remote sensing, but it shows the minimal product file structure well. We are now going to examine the file structure. After pressing the Synchronize button in IDEA, we can see the product header product_ex3a.dim appearing in the Project pane:

Double-click product_ex3a.dim and assign the type XML-Files in the Register New Extension dialog:

Examine the BEAM-DIMAP XML-header in the editor:

In the Project pane you will also see the data directory product_ex3a.data of the BEAM-DIMAP product. product_ex3a.img contains the binary image data and product_ex3a.hdr the image header in ENVI format:

The ENVI header can also be viewed in the IDE because its a text file: