Forest inventory based on LiDAR mobiles and A.I.
GO Monte Digital group ( Dielmo 3D, University of Castilla La Mancha and Naturtec) has developed an automated forest analysis software based on AI and LiDAR mobile data.
This specific software, called “AID-FOREST” (Artificial Intelligence for Digital Forest) allows to interpret mobile or terrestrial mobile point cloud data; transforming this information into dendometric and dasometric parameters automatically. As a result, inventories and its accuracy are improved, while costs are significantly reduced.
This new methodology is a game-changer for forestry, allowing foresters not only to optimize and improve the way data is collected and analysed on daily basis; but also helping to reduce environmental risks and to generate economic wealth in the sector.
3D forest structure
With LiDAR technology we can move from a two-dimensional information availability concept in pixels, to a three-dimensional concept with a format called VOXELS (volumetric pixels), where it is possible to know the percentage of points arranged in each voxel of space.
In the case of forests, the three-dimensional structure has implications for the spread of forest fires, the richness of wild communities, etc. In addition, other functional aspects of forests, such as productivity, are related to the structure of forest cover.
The availability of the three-dimensional structure of the forest makes it possible to establish a classification of fuel models of X classes, which favours the differentiation of vegetation structures with or without continuity of fuel between the shrub layer and the canopy layer. These classes can also be grouped into others.
DIELMO has developed proprietary algorithms for the estimation of fuel models based on the 3D forest structure and we have the ability to customize the algorithms to use this 3D forest structure along with other parameters to improve these models. In addition, we can automate our LiDAR forestry analysis and applicatoins in a production line to process large volumes of data.