DETECTION OF SINGLE STANDING DEAD TREES FROM AERIAL COLOR INFRARED IMAGERY BY SEGMENTATION WITH SHAPE AND INTENSITY PRIORS

Abstract. Standing dead trees, known as snags, are an essential factor in maintaining biodiversity in forest ecosystems. Combined with their role as carbon sinks, this makes for a compelling reason to study their spatial distribution. This paper presents an integrated method to detect and delineate individual dead tree crowns from color infrared aerial imagery. Our approach consists of two steps which incorporate statistical information about prior distributions of both the image intensities and the shapes of the target objects. In the first step, we perform a Gaussian Mixture Model clustering in the pixel color space with priors on the cluster means, obtaining up to 3 components corresponding to dead trees, living trees, and shadows. We then refine the dead tree regions using a level set segmentation method enriched with a generative model of the dead trees’ shape distribution as well as a discriminative model of their pixel intensity distribution. The iterative application of the statistical shape template yields the set of delineated dead crowns. The prior information enforces the consistency of the template’s shape variation with the shape manifold defined by manually labeled training examples, which makes it possible to separate crowns located in close proximity and prevents the formation of large crown clusters. Also, the statistical information built into the segmentation gives rise to an implicit detection scheme, because the shape template evolves towards an empty contour if not enough evidence for the object is present in the image. We test our method on 3 sample plots from the Bavarian Forest National Park with reference data obtained by manually marking individual dead tree polygons in the images. Our results are scenario-dependent and range from a correctness/completeness of 0.71/0.81 up to 0.77/1, with an average center-of-gravity displacement of 3-5 pixels between the detected and reference polygons.

[1]  W. Eric L. Grimson,et al.  A shape-based approach to the segmentation of medical imagery using level sets , 2003, IEEE Transactions on Medical Imaging.

[2]  Lindi J. Quackenbush,et al.  AN AUTOMATED OBJECT-BASED APPROACH TO DETECT SIREX-INFESTATION IN PINES , 2012 .

[3]  M. Harmon,et al.  Ecology of Coarse Woody Debris in Temperate Ecosystems , 1986 .

[4]  R. Reulke,et al.  Remote Sensing and Spatial Information Sciences , 2005 .

[5]  C. Tucker.  Red and photographic infrared linear combinations for monitoring vegetation , 1979 .

[6]  Marco Heurich,et al.  IDENTIFYING STANDING DEAD TREES IN FOREST AREAS BASED ON 3D SINGLE TREE DETECTION FROM FULL WAVEFORM LIDAR DATA , 2012 .

[7]  R. Schlaepfer,et al.  Spruce snag quantification by coupling colour infrared aerial photos and a GIS , 2004 .

[8]  Tony F. Chan,et al.  Active contours without edges , 2001, IEEE Trans. Image Process..

[9]  Hailemariam Temesgen,et al.  A comparison of selected parametric and imputation methods for estimating snag density and snag quality attributes , 2012 .

[10]  Lindi J. Quackenbush,et al.  Active contour and hill climbing for tree crown detection and delineation. , 2010 .

[11]  O. Faugeras,et al.  Statistical shape influence in geodesic active contours , 2002, 5th IEEE EMBS International Summer School on Biomedical Imaging, 2002..

[12]  Jonas Fridman,et al.  Amount, structure, and dynamics of dead wood on managed forestland in Sweden , 2000 .

[13]  Rachid Deriche,et al.  A Review of Statistical Approaches to Level Set Segmentation: Integrating Color, Texture, Motion and Shape , 2007, International Journal of Computer Vision.

[14]  Xavier Bresson,et al.  Segmentation for Hyperspectral Images with Priors , 2010, ISVC.

[15]  James E. Vogelmann,et al.  Comparison between two vegetation indices for measuring different types of forest damage in the north-eastern United States , 1990 .

[16]  Daniel Cremers,et al.  Efficient Kernel Density Estimation of Shape and Intensity Priors for Level Set Segmentation , 2005, MICCAI.

[17]  Dirk Pflugmacher,et al.  Distinguishing between live and dead standing tree biomass on the North Rim of Grand Canyon National Park, USA using small-footprint lidar data , 2009 .

[18]  Linda S. Heath,et al.  National inventories of down and dead woody material forest carbon stocks in the United States: Challenges and opportunities , 2008 .

[19]  Douglas J. King,et al.  Mapping dead wood distribution in a temperate hardwood forest using high resolution airborne imagery , 2009 .

[20]  Jay Gao.  Remote Sensing Of The Environment An Earth Resource Perspec Five, 2 nd Edition , 2006 .

[21]  J. R. Jensen.  Remote Sensing of the Environment: An Earth Resource Perspective , 2000 .

[22]  Daniel Cremers,et al.  Kernel Density Estimation and Intrinsic Alignment for Shape Priors in Level Set Segmentation , 2006, International Journal of Computer Vision.

[23]  Adrian E. Raftery,et al.  Bayesian Regularization for Normal Mixture Estimation and Model-Based Clustering , 2007, J. Classif..

[24]  Cuizhen Wang,et al.  Using Landsat images to detect oak decline in the Mark Twain National Forest, Ozark Highlands , 2007 .

[25]  Marco Heurich,et al.  Spatio-temporal infestation patterns of Ips typographus (L.) in the Bavarian Forest National Park, Germany , 2013 .

[26]  S. H. Lee,et al.  DETECTION OF THE PINE TREES DAMAGED BY PINE WILT DISEASE USING HIGH SPATIAL REMOTE SENSING DATA , 2006 .

[27]  Marco Heurich,et al.  Object-orientated image analysis for the semi-automatic detection of dead trees following a spruce bark beetle (Ips typographus) outbreak , 2010, European Journal of Forest Research.