Wildfire Damage Assessment

The project investigates a fire of 370 hectares within the recovery programs and post-emergency development.

The project makes use of the relief of an area representative of the entire area affected by fire, as a "pilot" mapping to be attached in the Lombardy Region within the recovery programs and post-emergency development.

In particular, the survey is carried out using techniques high performance photogrammetric based on aerial instrumentation (UAVs) and terrestrial (GPS geodesics) to demonstrate the usefulness of this approach integrated into the return of elaborates able to highlight all the agronomic characteristics (tree species, status of degradation / damage) and geometric (measurement centimeter of distances, surfaces and volumes) suitable for evaluate and classify from the qualitative point of view e quantitative the whole incendiary phenomenon and the state vegetative.



The project examines an area of 50 hectares whose center is the Val Castellera nature reserve, in particular the area is included within the following boundaries:

• WEST from the eastern slopes of the mountains Chiusarella and Martica

• SOUTH from Val Fredda in the south near the Alpe Cuseglio,

• EST from the Valvassera Mine

• NORTH from the source of the Margorabbia river


Of the total 370 hectares of the fire, 88% fall within the areas of protections of the Natura 2000 directive, and the area taken into exam is placed in the center of the area where the fire has been stationed for longer.


The vegetative typologies on which we want to investigate are the following
• Shrubs: Brughiera, Erica
• Woods: Roverella, Beech and Chestnut




The survey was composed of the following phases:

Preliminary Survey

During this phase the team interfaced with the park ranger, in quality of expert connoisseur of the Park's path and responsible for the supervision group, for identify the best organizational solutions.
On January 31, 2019, the first 8-hour survey took place on the paths 14, in Val Castellera, Alpe Cuseglio and Minivale Valvassera.


Logistics and technical inspections

The PTS technician made several excursions to the points identified and positions different photogrammetric targets on the ground.


Topographic typing

The PTS technician carried out two points typing campaigns with GNSS instrumentation topographical on the 14th and 15th of February 2019 in the previously used hiking routes for the target positioning on the ground.
It is connected via GPRS mobile network to the local "NETGEO" geodetic network for the acquisition of geographic coordinates with millimetrical positioning accuracy RTK (Horizontal: 5 mm; Vertical: 10 mm)



The Photogrammetric Training School and Tidalis Aero team performed on 18th of February 2019 the survey with the UAV, a SenseFly EbeeRTK with RGB camera SenseFly SODA with the following performances optical.


A total of 472 frames are acquired with geographic location coordinates positioned with decimetric accuracy, with adequate overlap to guarantee the condition fundamental that every point of the final paper is visible in common between at least 5 frames.


Data Elaboration

The technician performed the processing of the data collected in workstation with industry-leading software Pix4Dmapper (

The processing consisted of the following phases:
1. Import frames detected by drone
2. Insertion of photogrammetric target coordinates and maximum processing on workstation
3. Cloud processing and finishing results on workstations




The aero-photogrammetric products resulting from the processing are shared with the Regional Park Authority Campo dei Fiori by means of 3D visualization software and GIS. 

The quality of the digital processing products is analyzed and listed in the annex “report of quality ”, which quantifies the accuracy of the results of the data entered and of the products processed.

Below is a summary of these properties:
Average pixel ground resolution (GSD): 4.90cm/pixel
Root mean square error (RMS): 0.015m (axes X, Y), 0.057m (Z axis)

Note the importance of the accuracy of the measurements accepted by the technical office, which has influence on the choice of topographic instrumentation to be used to acquire points at known coordinates in the field using photogrammetric targets.

Digital products exported during processing are the result of photogrammetric processing, that is to say the identification of homologous points between the frames which reconstructs their shape, size and position in space forming a high-density 3D dot pattern that comes next plane reprojected in a 2D orthomosaic.


The data

3D point cloud

The 3D point cloud, called "point cloud", is formed by the set of correspondence points in common between each aerial frame and is generated automatically by the algorithm of photogrammetric processing with a "stereoscopic forward intersection" technique.
They are equipped with coordinates geographical areas in which positioning accuracy is determined by the type of support topographic (see above), for which the model is insertable and stackable in any 3D digital display system from PC to perform spatial analysis, extrapolate geometries and export sections / contour lines.

In this project, over 45 million 3D points have been identified, with a dot density per cubic meter of 23.4.

Triangulated 3D mesh

The triangulated 3D mesh, also called "3D Mesh" is a digital product made up of one mathematical surface that "triangulates" all the 3D points of the point cloud describing them the trend.

Being a "vector" type product, it can be imported and modified into software CAD and GIS design for subsequent analysis or simply to divulge the result final in an interactive way with platforms from virtual reality, 3D viewers and scaled 3D prints and in color.


The 3D Mesh of the project is composed of 1.2 million of jerseys.


The orthomosaic is the leading product in photogrammetry and is formed by the mosaic two-dimensional of all the frames taken on the basis of their common points.
Being a "raster" format it lends itself to a large number of cartographic operations in GIS software for the generation of maps in false colors, interpolations and spatial analyzes linked to other tabular / geographical data associated with a survey.

The orthomosaic of the project has an area of 73 hectares, an average resolution of the pixel to soil (GSD) of 4.9 cm/pixel and is returned in real RGB colors.


Relevant features

In the visible spectrum, elements can be distinguished by simple photointerpretation soil clearly distinguishable at resolutions between 1:100 and 1:1000. 


If followed clear examples of the state of the vegetation with particular reference to:

1. Hardwoods
2. Conifers
3. Shrubs
4. Mixed strips


The state of impairment of vegetation is visible in all points where the foliage does not prevent a clear visibility both to the ground and to the crown itself, the winter period of relief lends itself particularly to highlight these features. 


The areas of intact vegetation and heath, the areas of living soil are also clearly distinguishable and the areas covered by dense scrub little affected by the incendiary phenomenon.


Digital model of surfaces

As with orthomosaics, the digital model of surfaces (DSM) is a representation two-dimensional of all frames based on their common points, which highlights on color scale the trend of the "surfaces" seen from above.

It differs from the digital model of the land in that it represents the trend of the surface of the land and does not take into account the elements above it (buildings, trees, cars, ...).


For a clear representation of surface differences is returned in color scales (eg "false RGB colors", ...) to allow you to assign a different color to each difference height (eg Blue = points with minimum altitude, Green = points with average altitude, Red = Points with maximum altitude).


The DSM of the project has an area of 73 hectares, an average resolution of the pixel on the ground (GSD) of 4.9 cm / pixel and is returned in scale of false RGB colors. 


Classification of results


A demonstration of the great potential of drone survey and post-processing software geospatial, a map was produced in which the areas detected according to one were classified coloration "in false colors" following the legend:

Blue-blue: Areas of burnt soil
Orange: Vegetative areas
Green-water: Shaded areas in crevasses

The predominance of burned areas with respect to the remaining areas is clearly visible Typical color distribution is represented by the "blue-orange" pattern, which represents visibility aerial of areas with superficial burnt soil and apparently healthy treetops. 


Combining the information from the RGB orthomosaic and those provided by the map thus classified, is It is possible to generate on the GIS software homogeneous areas of damage by maximizing visibility under foliage given by surveys in winter. 

On the contrary, during the spring rebirth it will be possible to highlight the real state of recovery vegetative through appropriate mappings combined with RGB and multispectral sensor. 


This information can be integrated into important analyzes with LIDAR sensor (laser scanner) such as:

• height of the plants
• leaf density
• volume of timber present


Finally, the validity of this integrated approach can be implemented in a mapping approach drone thermographic to monitor the presence of hydrogeological risk.