Semi-Automated Delineation of Stands in an Even-Age Dominated Forest: A LiDAR-GEOBIA Two-Stage Evaluation Strategy

Abstract

Regional scale maps of homogeneous forest stands are valued by forest managers and are of interest for landscape and ecological modelling. Research focused on stand delineation has substantially increased in the last decade thanks to the development of Geographic Object Based Image Analysis (GEOBIA). Nevertheless, studies focused on even-age dominated forests are still few and the proposed approaches are often heuristic, local, or lacking objective evaluation protocols. In this study, we present a two-stage evaluation strategy combining both unsupervised and supervised evaluation methods for semi-automatic delineation of forest stands at regional scales using Light Detection and Ranging (LiDAR) raster summary metrics. The methodology is demonstrated on two contiguous LiDAR datasets covering more than 54,000 ha in central Idaho, where clearcuts were a common harvesting method during the twentieth century. Results show good delineation of even-aged forests and demonstrate the ability of LiDAR to discriminate stands harvested more than 50 years ago, that are generally challenging to discriminate with optical data. The two-stage strategy reduces the reference data required within the supervised evaluation and increases the scope of a reliable semi-automatic delineation to larger areas. This is an objective and straightforward approach that could potentially be replicated and adapted to address other study needs.

Stand Characterization of Eucalyptus spp. Plantations in Uruguay Using Airborne Lidar Scanner Technology

“…According to Haara and Haarala [12], from a remote sensing perspective, stands delineation is a segmentation problem. Optimal segmentation should maximize intra-segment homogeneity and inter-segment heterogeneity [13]. Most unsupervised evaluation methods involve the calculation of intra-segment and inter-segment uniformity and dissimilarity for each segment, and the subsequent aggregation of these values into a unique comprehensive value [14].…”

Automated Delineation of Microstands in Hemiboreal Mixed Forests Using Stereo GeoEye-1 Data

“…Metric wVarNorm CHM characterises the height variance of the pixels belonging to the same microstand [21]. The upper bound of wVarNorm CHM is equal to 1 when the variance of the segment is equal to the variance of the whole segmented image, while the lower bound is 0 when all pixel values within a segment are the same.…”

“…Supervised, direct polygon overlap metrics included OS, US, D, and C B . Oversegmentation (OS), undersegmentation (US), and summary score (D) explained in [21,35] were used as area-based metrics to characterise the overlap between microstand polygons delineated by the image analyst and the segmentation workflow. Thus, the values close to zero indicate higher accuracy, but those close to one indicate low accuracy.…”

“…The algorithm includes the region-growing segmentation by minimizing the spectral and spatial heterogeneity of the segments. Several datasets, including orthophotos [20], LiDAR-based feature images [21], and optical and LiDAR data fusion [15], can be processed using this method. The region-growing approach is also employed in other studies [22,23].…”

“…The higher accuracy of those forest inventory variables hints at a more efficient segmentation [23,26]. The direct accuracy assessment includes evaluating the quality of the segments using reference data such as stands delineated by experts (supervised assessment) [21] or segment homogeneity criteria [17] (unsupervised assessment).…”