Jul 09, 2020

New satellites are being launched monthly with most designed to record and monitor vegetation change. The increased temporal resolution (allowing daily revisit) represents an important shift towards continuous monitoring of resources. The ability to monitor the same location repeatedly (cloud-permitting) enables the detection of subtle changes in vegetation vigour and identification of trends. The real analytical efficiencies are accomplished by leveraging off cloud computing architecture which hosts and serves petabytes of historical and recently acquired images on-demand. With data held in this environment there is no need to individually review, download, or process and analyse satellite imagery as was the norm in the recent past.

Indufor’s resource monitoring team have developed a Continuous Plantation Monitoring System (CPMS) that leverages off both free and commercial satellites (such as Planet and Maxar) to provide timely and accurate information. The CPMS enables the monitoring of harvesting and plantation development across large areas. The rationale is that the increased frequency of satellite overpasses (temporal frequency) at high resolution adds a further dimension that enables planning managers to more efficiently allocate field resources, resulting in a more structured approach to monitoring and the area update process. The Canopy Index (CI) is one of the algorithms routinely applied to satellite data. The main output is a colour-coded map that identifies any unusual deviations from expected benchmark values, like areas impacted by foliar diseases, or as in the example to identify plantation gaps.

The data layers produced can be readily displayed using web-enabled dashboards as a way to communicate resource updates to project proponents or investors. Alternatively, to save time and resources by allowing foresters responsible for conducting field inspections to quickly validate harvest areas, pinpoint areas of un-mapped change, disease or failed areas.

The example below compares a CI output (right) to a high resolution Maxar WorldView subset. Here the CI was generated using open access Sentinel-2 data with a resolution of 10 m (compared to the 1.8 m of the WorldView). Despite the relative coarseness of the input data, the CI agrees well with the contemporaneous WorldView scene.