Using machine-learning to monitor vegetation height and cover, and other environmental impacts, of operations using data from active and passive sensors in orbit.


Canopy height and coverage are a good indicator of forest carbon content. Sensors on orbit will be used to map canopy height and coverage. Spatial and temporal resolution of products will vary depending on suitable technology and methodology and their applicability to the area of interest (depending on longitude, latitude, location characteristic, forest density, etc..)

The spatial/pixel resolution of the variables of interest could be between 20/25 m (and up to 1000 meters), depending on sensors and retrieval methodology

Instead of relying on manual review of data, or on site inspection, the system will analyse the data constantly and raise an alert when limits are exceeded so that appropriate action can be taken.

The system features include:
- automatic detection of canopy height and alert when the height exceeds 2m within a 20m corridor
- assessment of impact of operations on carbon content by measuring spatial and temporal changes in vegetation height and percentage cover
- multi-sensors approach to guarantee spatial continuity of observations and leveraging international considerations,
- processing and aggregation of data from latest sensors,
- usage of state of art methodologies published in scientific literature, based on machine learning.
- frequency of data analysis can be optimised to balance needs against cost of data storage and processing

In addition to vegetation height, monitoring other operational and environmental impacts such as fire, flood and water course changes, land displacement, methane concentration, and deforestation can also be monitored, depending on regional parameters and sensor type.

In order to meet the monitoring needs, the sensor selected will depend on the the longitude and latitude of assets spread. For example, canopy height can be measured using Light Detection and Ranging (LiDAR) and Synthetic Aperture Radar (SAR) interferometric data. However, LiDAR data can be scarce in some regions such as the Indian tropics,

In addition, Interferometric SAR data from commercial satellites at high spatial resolution are costly and high temporal de-correlation makes freely available Sentinel-1 interferometric data mostly unsuitable for tropical forests.

If LiDAR measurements are not optimal for the area of interest, a spatial continuity of canopy heights/coverage indicators can be achieved by utilizing a multi-sensory approach, combining biophysical parameters and detected signal features with machine learning techniques. Sentinel-2 (Multi-Spectral Instrument) derived biophysical parameters and Sentinel-1 (SAR) interferometric coherence are proposed as a new method for estimating canopy height.

An image attached below provides an example of vegetation health assessment (in this case the Normalized Differential Vegetation Index (NDVI)), highlighting the health and stress of vegetation in an area of interest. In this example, vegetation canopy height and cover will be measured and displayed, with spatial and temporal variability, visualized for selected points.

Alerts, for such variables, are set in proximity of assets, depending on the selected thresholds and criteria.

Key Benefits

Remote monitoring of vegetation height and other environmental indicators, next to widely dispersed and remote linear assets, means on-site in-person inspections can be substantially reduced or eliminated.

Automatic assessment of the sensor data, using machine-learning technology, means that asset owners are not relying on manual, tedious, time consuming analysis of the data by valuable staff.

In addition to vegetation height, the same approach can be applied to potential asset risks and asset environmental impacts (Flood, Land movement, vegetation stress, etc).

Data acquisition and assessment frequency can be optimised, balancing the cost of data storage and processing against the type of risk, environmental impact and the value of near real-time monitoring.


Remote sensor data analysis can be used to monitor:
- Biomass estimates
- Carbon off-set
- Deforestation monitoring
- Variation of the ecosystem and its habitat
- Right of way
- Assets encroachment
- Land use management
- Site selection

Please additional material available on DNVGL web site:
- Multiple environmental impacts and risk monitoring, see here:
- Satellite-based remote sensing for energy infrastructure, see here:

Purchase a license for full unlimited access to all innovation profiles on LEO

  • Direct connection to thousands of more innovations
  • Access to market Experts and Universities
  • Filter relevant solutions into your own dedicated Network