A digital image begins as an analog signal. Through computer data processing, the image becomes digitized and is sampled multiple times. The critical characteristics of a digital image are spatial resolution, spectral resolution, radiometric resolution, contrast resolution, noise, and dose efficiency. These depends upon satellite orbit configuration and sensor design. Different sensors have different resolutions.
Spectral resolution describes the ability of a sensor to define fine wavelength intervals. The narrowest spectral interval that can be resolved by an instrument. Spectral resolution (spectral capability) also refers to the number of wavebands within the EM spectrum that an optical sensor is taking measurements over.
Radiometric resolution can be defined as the ability of an imaging system to record many levels of brightness. Radiometric resolution refers to the range in brightness levels that can be applied to an individual pixel within an image, determined on a grayscale. E.g., Sentinel-2 sensor MSI is a 12 bit sensor imaging with 4.096 levels.
Spatial resolution of an image corresponds to the size of the minimum area that can be resolved by the sensor.
Temporal resolution, also referred to as the revisit cycle, is defined as the amount of time it takes for a satellite to return to collect data from exactly the same location on the Earth. Imageing of the exact same area at the same viewing angle a second time is temporal resolution.
The spatial resolution of an image corresponds to the size of the minimum area that can be resolved by the sensor.
Due to the different techniques of acquisition of passive and active sensors, the spatial resolution is determined for both sensor types differently.
For passive sensors, the spatial resolution depends on their instantaneous field of view (IFOV), which determines the area of the Earth’s surface that is viewed at one particular moment in time by one detector element. The size of this area is called resolution cell and characterizes the spatial resolution of the sensor. Depending on the spatial resolution, whole features of the Earth’s surface can be detected homogeneously in one or several resolution cells. For features smaller than the spatial resolution, the average reflected radiation of all features within a resolution cell is recorded, leading to so-called mixed-pixels.
For imaging active systems, the spatial resolution is dependent of both the length of the transmitted pulse in looking direction and the width of the radiation beam or the antenna width in flight direction.
In all cases, the spatial resolution indicates the level of detail observable in an image. Usually, one distinguishes between coarse (low), moderate (medium) and fine (high and very high) resolution, whereby the use of this denomination is often context-dependent. Sensors with coarse resolution can only detect large features, but they usually cover a much larger area than high-resolution sensors, which can provide detailed information on small objects such as individual buildings, trees or cars, but for much smaller areas. Coarse spatial resolution mean in general a resolution cell larger than 250 m and a scene extent of several thousands of kilometers (>1000 km). Moderate resolution sensors have a spatial resolution of 30 m to 80 m, and a coverage of approximately 200 km in a single acquisition. Sensors showing spatial resolutions from 5 m or 6 m are high-resolution sensors, with a spatial coverage up to approximately 20 km. Sensors with a resolution cell’s width of less than 1 m are considered as very-high-resolution sensors.
Low resolution sensors are appropriate for the analysis of broad-scale phenomena such as ocean color or cloud patterns. Medium resolution sensors are rather used for regional analysis such as land cover change and phenological response of vegetation. High-resolution sensors are particularly useful for object detection.