Significant vegetation changes

This service detects probable changes in forest cover over a given territory and period. The service is based on the detection of change between two satellite images and more specifically on the difference in NDVI, which is an index measuring the rate of vegetation contained in each pixel. Using the difference in NDVI between two given dates and the measurement of statistical parameters, it delimits the probable changes that have occurred in a forest area between two dates. These changes may concern clear-cutting, but also areas affected by drought or disease.

The images used for this service are Sentinel-2 images with a spatial resolution of 10 or 20 m depending on the spectral bands.

eo-forestry-1

 

Forest Mapping

This service maps the forest areas in a territory (A) and identifies their typology, i.e. the differentiation between deciduous and coniferous trees (B). This typology can vary according to the forest and/or geographical context, the prerequisite being the availability of field data related to the analyzed area. The service uses Machine Learning algorithms to identify the pixels of the image and classify them as forest/non-forest (A) and deciduous/evergreen (B). A time dimension is added to the classification process to increase the accuracy of the results. Several images are thus processed over a period of one year, and the combination of the different classifications makes it possible to infer the probability of belonging to one of the classes. 

The images used for this service are Sentinel-2 images with a spatial resolution of 10 or 20 m depending on the spectral bands.

eo-forestry-2

 

Main Forest Species Mapping

This service maps the main forest species in a territory. The definition of the main tree species can vary according to the forest and/or geographical context, the prerequisite being the availability of field data related to the analyzed area. In this example the species mapped are 1) Spruce, 2) Douglas, 3) Larch, Pine and other coniferous, 4) Oak, 5) Beech, 6) other broadleaved. The service is based on Machine Learning algorithms to classify the pixels in the image and define them according to the class list. A time dimension is added to the classification process to increase the accuracy of the results. Several images are thus processed over a year, and the combination of the different classifications makes it possible to infer the probability of belonging to one of the classes. 

The images used for this service are Sentinel-2 images with a spatial resolution of 10 or 20 m depending on the spectral bands.

eo-forestry-3

 

Phenology of species and detection of anomalies 

This service illustrates the evolution of the vegetation cover of one or more forest stands within a graph. The service calculates, for each date when an image is available, the average of the Normalized Difference Vegetation Index (NDVI measures  vegetation change ) within each plot for which boundaries are provided. These values are recorded in a table that can then be exploited either by plotting the phenological curves per plot for simple visualization, or in order to compute more targeted analyses, such as for example the detection of anomalies in the vegetation cover from one year to another (drought, disease, clear cut, etc.).

The images used for this service are Sentinel-2 images with a spatial resolution of 10 or 20 m depending on the spectral bands.

eo-forestry-4