This project will develop a 360º solution applied to the power infrastructure asset management field. This solution integrates the acquisition of information through flights with RPAS, the analysis of this information and its integration in predictive models based on Machine Learning/Deep Learning techniques allowing the model to be increasingly precise and intelligent, with the ability to automatically analyse and update results online. The objective is to provide Business Intelligence performance information (m3/km of logging) that will allow the sector to propose new work schedules and optimise asset management.

The development lines on which the project is built are those that the market is demanding:

  • Infrastructure inspection. Work will be carried out on the development of specific algorithms based on Machine Learning, so that, making use of the information collected by different sensors and the current extensive HEMAV database, it will be possible to detect any anomaly in the infrastructure. Work will also be carried out on processing the data produced by the thermographic cameras to detect hot spots. This typology will be useful in the photovoltaic case.
  • Vegetation control. At this point, the main objective is to calculate the periodicity with which the felling and pruning of the vegetation must be carried out. To achieve this, a model with the following phases will be proposed:
    1. Vegetation mapping with a high level of detail. This phase is of vital importance because it will allow to know what type of vegetation exists at each point with a high level of detail. The mapping will be based primarily on data from the LiDAR sensor, and it will serve as a basis for modelling the vegetation growth rates.
    2. Vegetation growth modelling. Accurately predicting vegetation growth allows you to plan the best time to intervene to minimise risks. Using the data provided by the LiDArR technology or photogrammetry (point clouds), it will be possible to measure the height of the vegetation; and, together with diachronic data, an accurate estimate of the growth rate will be made. The weighted overlap method shall be used to discriminate between the information collected and to make the predictions. In addition, dendrometry data collection shall be incorporated for those trees which are highly dangerous or have a higher proportion of occurrence.
    3. Map of the network in 3D. The objective of this phase is to elaborate a three-dimensional representation of the entire network to be inspected, and thus be able to project the expected growth rates in order to determine the frequency of intervention.
    4. Calculation of the intervention frequency. At this point, and with all the information gathered from the previous phases, the maintenance interventions to be carried out will be planned. For this purpose, aspects such as: legal requirements regarding minimum distances must be taken into account, as well as the growth rate together with the particular orography of each point (making use of the representation resulting from phase 3).
  • Morphology. Using the same technologies that are applied for vegetation control, the necessary mechanisms will be developed to obtain the real distances between all the elements of each infrastructure. From a 3D model it will be possible to detect which spans do not respect the safety distance regulations (cable-wire, cable-ground, cable-other obstacles, etc). In addition, the high precision of this tool will allow the detection of defects or deviations in the structure, such as a high tension tower inclined by structural deformation.

Leave a Comment

Your email won't be published.

Main Headquarters


Edificio Ibercenter

Plaza Carlos Trías Bertrán 4 28020 Madrid

(+34) 914 184 546


ESA (European Space Agency) Edificio RDIT

Esteve Terradas 1 08860 Castelldefels

(+34) 932 202 063

Ciudad de México

Avenida Prado Norte 125,

Lomas – Virreyes, Lomas de Chapultepec III Secc,

11000 Ciudad de México, CDMX,


World Trade Center

Strawinskylaan 381