Development of a remote sensing based coniferous regeneration monitoring system

Sustainable forest management depends on timely regeneration of disturbed forests. During the regenerating stage many factors can substantially alter the composition and structure of the developing forest. Monitoring at this time is critical to ensure that desirable forests are established and provide for future needs. This research evaluates the potential of digital remote sensing for monitoring regeneration. Automated methods for detecting and delineating individual conifer crowns and for estimating the abundance of woody deciduous competing vegetation are developed using leaf-off imagery (i.e. when leaves have not flushed). Detection and delineation were tested on several sites with different regeneration characteristics. Results show that errors less than 20% can be achieved if viewing geometry is controlled (view angles < 15°) and the minimum tree size (i.e. crown diameter) to be detected is greater than 30 cm with ∼ 6 cm resolution imagery. Modeling of deciduous competition abundance showed that a useful relation existed between extracted image features and measurements made in the field. With the methods developed, a automated processing system for regeneration monitoring is possible.