- Satellite imagery makes it possible to monitor vegetation biophysical parameters and forest disturbances from space automatically using artificial intelligence
- Google’s TensorFlow AI software will be applied to Copernicus Sentinel-1, Sentinel-2, Planet and JAXA’s ALOS-2 PALSAR-2 satellite imagery
- Expected outcomes include the automated detection and labeling of different types of forest disturbances and the lost of aboveground biomass
Recent advances in computing technology, cloud computing and high-performance computing are paralleled with advanced artificial intelligence (AI) algorithms and significant investment in the European Copernicus Earth Observation programme and its Sentinel satellite missions. AI enables automatic detection of spatial patterns in environmental data such as satellite images based on training data. The paradigm of looking for spatial patterns instead of the historic focus on spectral information in satellite imagery allows the identification of the types of forest disturbances. AI can also be used to accurately estimate from space forest biophysical parameters that are difficult to measure in the ground such as aboveground biomass (Rodriguez-Veiga et al, 2017)
Machine learning / AI (Le Cun et al. 2015) have previously been applied to hyperspectral image classification (Hu et al. 2015), CORINE land cover mapping from Sentinel-1 SAR images (Balzter et al. 2015), forest biomass mapping using a combination or SAR and optical images (Rodriguez-Veiga et al, 2016), providing evidence of slavery from WorldView satellite images (Boyd et al. 2018).
This interdisciplinary studentship (National Centre for Earth Observation and Department of Mathematics) aims to explore the application of AI to operational automated forest monitoring, focusing on tropical forests (cases in Kenyan and Colombian forests). Time-series stacks of multispectral optical and SAR sensors will be input into the AI. The AI will be trained based on mesurements collected from in-situ forest inventories and visual interpretation of very high resolution images.
- How accurately can an AI be trained to identify types of forest disturbances based on satellite time-series information?
- Over and above the type of disturbance how accurately can the associated aboveground biomass loss be estimated?
Methods will be drawn from Mathematical Modelling / AI and Earth Observation / Geography. The TensorFlow AI (Abadi et al. 2016) will be implemented on the High-Performance Computing facility SPECTRE-2 at Leicester and linked with an existing 10 m resolution Sentinel-2 image processing chain developed in Python (https://github.com/clcr/pyeo). It will also be adapted to ingest Planet, Sentinel-1 and ALOS-2 PALSAR-2 imagery. Training data for the AI is available from in-situ forest inventories and interpretation of high resolution imagery. Once trained up, the AI will identify forest disturbance types and aboveground biomass loss over time. The outcomes of this analysis will be validated against recently collected forest inventory data interpretation of high resolution imagery. Experiments with different spatial filters and configurations of TensorFlow will be undertaken to optimise the detection of particularly difficult forest disturbances (e.g. gold minning, under canopy crops), and estimation of high forest biomass density levels.
Training and Skills
The National Centre for Earth Observation will provide access to its Researcher Forum, staff conferences/workshops and national-level training
The student will be trained in Sentinel data processing on the HPC facility SPECTRE-2 at Leicester. The student will take the new MSc module GY7709 (Satellite Data Analysis in Python), available in 2019-2020, and any other modules deemed suitable, dependent on the background of the student. Complementary individual training in using AI, especially TensorFlow, will be available from the Department of Mathematics. Further training will take place ‘on-the-job’ as part of the research team in NCEO.
Year 1: Literature review, refinement of research questions and work plan, liaison with and consultation of project partners, installation of AI on HPC, training in satellite data processing in Python
Year 2: Training the AI, completing test runs with optical data (Sentinel-2 and Planet) and evaluating outcomes, iteratively refining data flow and accuracy, running AI ‘experiments’
Year 3: Implementing SAR data (ALOS-2 PALSAR-2 and Sentinel-1) AI runs, evaluating and comparing results, submitting 2 papers for publication
Partners and collaboration (including CASE)
- Kenya Forest Service (KFS) - goverment agency in charge of forest inventory and forest monitoring in Kenya
- Institute of Hydrology, Meteorology and Environmental Studies (IDEAM) - goverment agency in charge of forest inventory and forest monitoring in Colombia
- Planet – satellite assembly, operations and data analytics company
- Google – London office has expressed interest in projects around AI and satellite data
Informal enquiries are welcome to Prof. Heiko Balzter, firstname.lastname@example.org