Automatic forest health monitoring system unveiledFriday 20 Oct 2023
The innovative system was developed with collaboration partner Indufor as part of Scion’s remote sensing and GIS team’s mission to map, measure, and monitor the nation's forests. Red needle cast, an unyielding fungal-like disease that has been impacting pine trees in New Zealand since at least 2008, served as the primary case study for this system's development. The innovative system at its core combines free imagery sourced from ESA's Sentinel-2 satellites with high-resolution imagery from commercial satellites. During the last four years, extensive mapping and analysis of East Coast pine forests have been carried out using the high-resolution images to pinpoint regions impacted by red needle cast disease. These results have undergone verification by collaborating pathologists. These disease-affected regions are essential for training a multi-temporal model, which leverages more frequent but less detailed imagery from Sentinel-2. This year, the success of the monitoring system has been bittersweet for Scion researchers, who have progressively seen large swaths of East Coast forests change colour from green to brown since March 2023. This illustrates the severity of the disease’s impact as well as providing major validation of the system’s effectiveness in detecting disturbances and large-scale forest health issues. Scion team lead for remote sensing and GIS, Grant Pearse, says what sets this system apart is its ability to automatically monitor only forest areas by combining the detection routine with the high-resolution forest maps Scion is producing for their Digital Twin project. This dramatically reduces the amount of computation required. “This targeted approach ensures that alarms are only triggered when there's a real threat to the forest's health, eliminating false positives caused by unrelated changes in the landscape. It’s ability to target specific areas and bypass irrelevant data or landscape disturbances, makes it a very precise and cost-effective monitoring tool.” While the system has demonstrated its capabilities in monitoring large outbreaks, the team is keen to receive more on-the-ground intelligence from industry sources. These tips from foresters can further enhance the system's accuracy in detecting and responding to forest health threats. “Ongoing collaboration between scientists and industry players is crucial in enhancing this biosecurity tool's capabilities,” adds Pearse. “Having more accurate data means we’ll need to investigate fewer false positives. Ideally, we are looking for larger areas of trees affected by red needle cast as satellites struggle to detect a few roadside trees, for example.” The success of this system has raised hopes for its expansion to a national system once all New Zealand’s planted forests have been mapped by Scion’s team. While the focus has primarily been on disease outbreaks, the system has the potential to serve as a general disturbance detector. By targeting specific areas for monitoring, it can effectively identify and respond to various threats that might affect planted forests. Additionally, the team wants to use the system to study areas within affected forests that remain healthy, potentially due to disease resistance or other factors. This knowledge could provide valuable insights into forest resilience and contribute to more effective forest management strategies. “The benefit of this monitoring system is that it also provides an opportunity to gather more data so we can learn more and provide even greater support to industry as it looks to recover from the impact of red needle cast or other diseases in the future.” Its asset as a valuable forest protection tool means Scion is now exploring the possibility of integrating this system into New Zealand’s broader biosecurity measures. Click here to view the automated forest health monitoring system for red needle cast. A live demonstration of the system will be delivered at the Remote Sensing Cluster Group in a pre-conference workshop the afternoon before the ForestTECH 2023 NZ event runs. This will be held in the Rimu Room at Scion in Rotorua from 1pm on Monday 13 November. The full agenda for the workshop can be viewed here and the programme for the ForestTECH 2023 event running on the 14-15 November, here Source: Scion 
| ||
Copyright 2004-2024 © Innovatek Ltd. All rights reserved. |
