Imagery

View and change symbology of a satellite image

​Link to download data for this section

Satellite imagery intro

Optical medium-resolution imagery

Imagery is usually supplied by MAGIC as an RGB synthesis. Add a satellite image to the map and style it.

1. Drag image file LC09_L1GT_220108_2024-04-07_BGRNIR.tif into the project. This is a Landsat-9 medium-resolution image (30m/pixel), it has 4 spectral bands (1: Blue; 2: Green; 3: Red, 4: Near Infra Red) and 16-bit depth pixels.

2. Go to Image Properties and then to the Symbology tab. To create a true-color composite, set for Red Band 3, for Green - Band 2, and for Blue - Band 1. Click OK. In general, this creates an acceptable image representation.

Image Symbology

3. Zoom into Rothera Point, it may seem dark. For this case, use the following stretch method. Go to layer Properties - Symbology. Expand Min/Max Value Settings section. Change Statistics extent from Whole Raster to Current extent.

What has just happened?

An image could be represented as a histogram - a graph showing distribution of pixels with different brightness within an image. Brightness is just a pixel value: low values correspond to dark pixels (rock, shadows, water), high values - to bright pixels (snow, clouds).

At the histogram of the Landsat image of Rothera first group of peaks (about 5000 - 6000) indicates water, and the second group (around 9000) - snow and clouds.

When new image added to map view, QGIS by default stretches the histogram between 2 and 98 percentiles of the histogram, cutting out the most bright and dark pixels. This usually works well with cloud/snow/water-free images. In this case, the default style leaves point of interest in darkness. But it is possible to recalculate the histogram just for the visible area and adjust the histogram to it.

4. Some sensors have a Panchromatic band (of broad spectral range, usually from blue to yellow or red, band with higher resolution), which allows for enhancing image resolution and preserving original colors. To compare the original and enhanced (pan-sharped) images, add the file LC09_L1GT_220108_2024-04-07_BGRNIR_PSH.tif to the map. Apply the same symbology as in step 3. Compare the level of detail.

Optical high-resolution imagery

5. Low- and medium resolution images are crucial for sea-ice, snow coverage assessment, and monitoring matters. But, for studying an area in more detail, high-resolution and very-high resolution data is needed. Add WV2_2020-02-16_RGB.tif image to the map view.

SAR low-resolution imagery

6. Add S1A_IW_GRDH_1SSH_20240409T082739_0494_S_1.8bit.jp2 image to map view. This is a Sentinel-1 image captured in HH-polarization 2 days after the Landsat image above. Snow-covered surfaces, dry soil, man-made objects, and thick ice usually look bright on SAR images, while water, thin ice, and wet soil - are much darker. SAR usually acquires images from a low angle, this creates shadows in mountain areas, as no signal back-scattered from them. You may also notice the Foreshortening and Layover effects.Compare this image with the more recent one S1A_IW_GRDH_1SSH_20240807T082737_E354_S_1.8bit.jp2.