Part 4—Measure Areas using Thresholding
Step 1 – Download Landsat Triptych Image
- Go to the NASA Earth Observatory search box and enter the terms Aral Sea, 1973, and 2000 in the Search field, click the link for the article entitled "The Shrinking Aral Sea."
- Alternately click this link to take you directly to the Earth Observatory Newsroom article, entitled The Shrinking Aral Sea.
- Download the three-part image that appears at the beginning of the article. This triptych or "triple image" shows 3 Landsat images of the northeastern portion of the Aral Sea. The images have a spatial resolution of 250 m per pixel.
- NOTE: Download the relatively small image (JPG) that appears with the article. Don't attempt to work with the full-size image available on the page!
Step 2 – Set Scale
- In ImageJ, open the landsat_aral_triptych.jpg file. Compare these images with the upper-right portion of the MODIS images to see which portion of the Aral Sea is shown in the triptych images. Close the MODIS images.
- Choose Analyze > Set Scale... and enter information to show that 1 pixel represents a known distance of 250 m. Alternatively, if you want to read your results in km, you could set 1 pixel to represent a distance of 0.25 km.
NOTE: If the image takes an abnormally long time to open, you may have grabbed the large version of the file. Delete that file and go back to Step 1 above to download the smaller file.
Step 3 – Automatically Select and Measure Pixels that Represent Water
In these images, pixels that represent water are much darker than pixels that represent land. If the color information for each pixel is converted to grayscale, the dark pixels representing water will have low values and the lighter pixels that represent land will have higher values. As the ranges of pixel values that represent land and water are unique, you can manipulate ImageJ to highlight the pixels that represent water in each image, and measure the area it covers.
NOTE: the techniques described here do not work as well on the MODIS images that you used in the previous section, as there is not enough contrast in the MODIS images.
NOTE: the techniques described here do not work as well on the MODIS images that you used in the previous section, as there is not enough contrast in the MODIS images.
- Choose Image > Type > 8-bit to convert the image to grayscale. This action converts each pixel's color information into a brightness measurement.
- Choose Image > Adjust > Threshold.... The Threshold dialog window allows you to highlight pixels in an image that have values within a range you define. Adjust the sliders so that the dark, low-value pixels that represent water turn red, but those that represent land don't change. Close the Threshold dialog window without clicking any of the buttons.
- To measure the area of water that you highlighted in each of the three images, select the Rectangular Selections tool in the ImageJ toolbar and drag a rectangle over just the top image of the triptych.
- Choose Analyze > Set Measurements... and click the Area and Limit to Threshold checkboxes. This sets a preference to measure only the highlighted pixels within the rectangular selection you made.
- Choose Analyze > Measure to measure the area covered by water in the image. Results will be reported in whatever units you set. You do not need to save your previous results. You may need to choose Window > Results to bring the Results window forward.
- Put your cursor inside the rectangular selection and click and drag the selection down to center it on the second image, then repeat the measurement.
- Move the selection down over the third image and measure again. Record your results for all three images on paper. Include the units of measure you used when you set the scale.
Step 4 – Interpret Your Results
Prepare a quantitative description of how the area of the Aral Sea visible in these three images changed from 1973 to 2000.
For instance, you could try the following:- graph the sea's area versus time
- create an annotated image to show areas that were exposed between successive images
- compare the amount of land that was exposed in successive images to familiar features such as a state, national park, or city.
The Going Further section of this chapter offers several suggestions for research ideas in which you can apply the technique of setting a scale and making measurements in new situations.