Sentinel 2

Sentinel-2 is an optical resolution satellite system operated by the European Space Agency (ESA). The keyword "optical" here refers to the satellite primarily using light—visible or infrared radiation—for its observations. The multispectral instrument (MSI) of the Sentinel-2 satellite is a remote sensing device or sensor capable of measuring and recording various wavelengths or spectral ranges to gather information about the Earth's surface, atmosphere, and other objects. Sentinel-2 images are quickly and conveniently available in the Satiladu environment and also in our satellite data portal.

Sentinel-2 products consist of pellets or tiles obtained from a single data collection. A granule is the smallest indivisible unit of a product that contains all possible spectral channels. For Level-1B, the granules are small images created on the basis of the sensor geometry, recorded by the sensors of the twelve MSI (Multispectral Imager) detectors. The size of each granule is 23 × 25 km².

The following graph (Figure 3) provides us with an overview of the spectral regions collected by the Sentinel-2 satellite. We can see that there are a total of 13 channels, represented in the graph with different illustrative colors. A channel is a specific range of wavelengths within which certain types of information are collected or recorded. For example, channels 2, 3, and 4 represented in dark blue, green, and red in the graph are specialized in the visible light spectrum, perceived by the human eye as blue, green, or red, respectively. Channel 1, marked in light blue, or ultraviolet-blue, is a part of the light spectrum with shorter wavelengths than visible blue light and is not perceivable by the human eye. Similarly, the channels marked in light blue and orange (5, 6, 7, 8, 8a, 9, 10, 11, 12) cannot be seen by the eyes in the near-infrared and shortwave infrared regions. Nevertheless, the technology on satellites enables the collection of this information, and through various processing methods, we can visually represent values measured in the non-visible spectrum to the human eye.

 

Figure 3. The spectral ranges collected by Sentinel-2 (source: Freie Universität Berlin).

The following table (Table 1) shows the spatial resolution and physical details of Sentinel-2.

	Channel	Sentinel-2A		Sentinel-2B
Spatial resolution		Average wavelength	Bandwidth	Average wavelength	Bandwidth
(m)		(nm)	(nm)	(nm)	(nm)
10	2	492,4	66	492,1	66
	3	559,8	36	559	36
	4	664,6	31	664,9	31
	8	832,8	106	832,9	106
20	5	704,1	15	703,8	16
	6	740,5	15	739,1	15
	7	782,8	20	779,7	20
	8a	864,7	21	864	22
	11	1613,7	91	1610,4	94
	12	2202,4	175	2185,7	185
60	1	442,7	21	442,2	21
	9	945,1	20	943,2	21
	10	1373,5	31	1376,9	30

Table 1. Sentinel-2 spatial resolution summary table (translated; source: ESA SentiWiki).

The atmosphere influences light in various ways, including scattering, absorption, and reflection, all of which can impact the quality and accuracy of data collected by satellite sensors. Optical sensors are highly sensitive to both the intensity and direction of light. Excessively strong light, such as direct sunlight, can result in oversaturation, causing areas of the image to appear overly bright or white, leading to a loss of detail and potential distortion. Conversely, low light conditions can diminish the sensor’s ability to capture fine details.

The angle of incoming light also significantly affects how objects are represented in the sensor’s field of view. For instance, at low sun angles, long shadows may obscure important features or distort their appearance. Similarly, light reflecting off surfaces, such as water (glint), can interfere with the sensor’s ability to accurately measure light and capture precise data.

To mitigate these challenges, the Sentinel satellites’ orbits and observation schedules are meticulously designed. Sentinel-2 satellites operate in a sun-synchronous orbit, ensuring they pass over each location on Earth at the same local solar time. This design provides consistent lighting conditions for observations, minimizing shadow variability and enhancing the comparability of data over time. While shadows cannot be completely eliminated, the Sentinel satellite system is optimized to reduce their impact, ensuring data remains as accurate and reliable as possible.

However, certain weather conditions, such as clouds, fog, or haze, obstruct light transmission and limit the sensor’s ability to collect ground information in affected areas (Figure 4). Additionally, optical sensors are unable to penetrate dense objects like thick vegetation. These limitations should be considered when analyzing satellite imagery.

Figure 4: Sentinel-2 satellite image partially covered by clouds with Lake Pihkva in the centre, covered by ice.

In Sentinel-2 imagery, clouds often appear as thin, wavy formations that can range from translucent to opaque (Figure 5). Their presence in satellite images (Figure 6) can significantly impact the quality and usability of the data by obscuring or distorting views of the Earth's surface. This interference complicates the interpretation and analysis of imagery. For instance, cloud shadows may create misleading artifacts, such as false positives when monitoring forest fires or analyzing agricultural patterns.

Figure 5. Semi-transparent clouds and haze; an area with cloud shadow in the upper left and smaller clusters of cumulus clouds on the right.

Figure 6. RGB satellite image showing opaque clumps of clouds with dark cloud shadows, Lake Võrtsjärv on the left..

Although Sentinel-2 satellites do not penetrate clouds, the clouds themselves can serve as valuable sources of information. Their presence or absence can reveal changes in precipitation patterns and microclimates within a region. Monitoring cloud patterns and distribution aids in mapping area shading and analyzing the spectral distribution of sunlight, which directly influences vegetation health and growth. While optical sensors have inherent limitations, their wide range of applications across various fields underscores their utility, provided their characteristics and constraints are considered.

Atmospheric correction in the context of Sentinel-2 refers to the process of mitigating the effects of atmospheric interference on satellite imagery. This involves intricate calculations and modeling to estimate atmospheric impacts and compensate for them in the imagery. However, this process can sometimes introduce artifacts or distortions. These arise because accurately modeling and compensating for all atmospheric components and dynamic conditions is challenging. For instance, some correction methods may overcompensate for specific wavelengths, resulting in reduced contrast or color distortion. Complex atmospheric conditions, such as dense clouds or aerosols, further complicate correction efforts, occasionally leaving residual inaccuracies or distortions in the images (Figure 7). It is crucial to recognize these potential errors and assess the corrected imagery carefully to ensure reliability and accuracy for specific applications.

The raw satellite data undergoes processing to generate Level-1C (L1C) products, which include radiometrically corrected and geometrically aligned data, as well as metadata on satellite and solar angles. Subsequently, L1C products are processed into Level-2A (L2A) products, which involve atmospheric correction. A widely used tool for this purpose is Sen2Cor, which applies atmospheric correction algorithms to reduce the influence of atmospheric effects on the imagery. Once corrected, the European Space Agency (ESA) makes these L2A products available to users through its data distribution platforms, ensuring they are ready for analysis and application.

Figure 7: RGB image with a relatively strong contrast difference in the centre as a straight line, due to over-compensation for atmospheric correction.