A spectral remote sensing sensor detects the reflected radiation of the Earth's surface. A remote sensing sensor detects the reflected radiation of the Earth's surface and stores it as numbers in a raster. In accordance, each area that has been detected constitutes a cell in a raster. These raster cells are called pixels. The size of an area represented in a pixel depends on the capability of the sensor to detect details.
Raster with low and high spatial resolution.
Low and high spatial resolution
The ability of a remote sensing sensor to detect details is referred to as spatial resolution. The spatial resolution is stated in metres. The more pixels are included in a remote sensing image of a certain area, the higher the spatial resolution meaning the more details can be observed.
The swipe below shops two satellite images of Bonn. You can clearly distinguish between a higher spatial resolution of 30 metres and a lower spatial resolution of 300 metres. In the image with the lower resolution, much more different objects must be included in one pixel.
Satellite images of Bonn with a spatial resolution of 30 metres and 300 metres respectively (© USGS/NASA Landsat Program).
In almost every satellite image, objects that are close together must be included in one pixel. Such pixels are called mixed pixels. The image below shows a house and a garden included in the same pixel. Due to the low spatial resolution, the colour components of both objects (brown and green) result in a brown-green mixed pixel, which is very hard to analyse. The lower the spatial resolution, the more mixed pixels and the harder it is to tell areas apart.
Formation of a mixed pixel caused by different objects in the same raster cell.
Why do only some sensors have a high spatial resolution?
We can ask the question: Why is it that not every sensor in spectral remote sensing has a very high resolution? The answer becomes clearer looking at the purpose of satellite-based remote sensing systems: If the same sensor is attached to an airplane and a satellite, the airborne sensor will have a very high resolution of e.g. 1 m, whereas the satellite-based sensor will have a low resolution of e.g. 30 m. At the same time, the satellite-based sensor detects a wider area in one single image and circles the Earth completely in only a few days. This is impossible for aircrafts!
The spatial characteristics of spectral sensors are determined by the ratio of extend and resolution. If a maximized extend is required in order to depict as great an area as possible, we have to lower our sight regarding resolution because we cannot store all these data in a sensor (fig.).
The images show four views of the same area as detected by four different sensors. The images all have the same number of pixels; but the spatial resolution of the different sensors cause different pixel sizes. Thus, in one case only one building is depicted (aerial image, 0,1 m, © ATKIS), another image shows the district of Poppelsdorf (QuickBird, 1 m, © DigitalGlobe), the third image covers the city of Bonn (Landsat, 30 m, © USGS/NASA Landsat Program) and the last image depicts a large area from Ruhrgebiet to Koblenz (MODIS, 300 m, © USGS/NASA MODIS Project)
Each remote sensing sensor produces raster image data. In turn, each raster consists of raster cells, which are also referred to as pixels. The bigger a pixel, the more objects on the surface of the earth are captured and the lower the spatial resolution of a raster image. The higher the spatial resolution, the smaller the amount of hard-to-analyse mixed pixels included in a raster image.