Sagi Filin

Neighborhood Systems for Airborne Laser Data

Analysis of common neighborhood definitions for airborne laser data, triangulation or raster-based, reveals deficiencies in modeling the measured objects. Concepts that originate from 2D data structures are used for modeling complex 3D objects and for handling datasets with different point densities. Realizing these shortcomings, this paper proposes a new neighborhood system for airborne laser data. Based on laser data characteristics the proposed systems consider, among other features, point density, layered and overhanging structures, and local surface trends. Parameters for the proposed systems are derived from theoretical and practical observations. The paper demonstrates the type of neighborhood that is established by the proposed systems, and shows that artifacts that are usually created by the common neighborhoods are avoided here, and that structures within the data that are usually masked are revealed. The paper demonstrates how subsequent applications benefit from the new system. Finally, the estimation of surface normals by the proposed systems is compared to the triangulation; results show a significant improvement in the reliability and quality of the estimation.

Orthogonal polynomials supported by region growing segmentation for the extraction of terrain from lidar data

Light Detection and Ranging (lidar) systems supply a massive amount points in 3D space with no semantic information helping knowing the objects they represent. To identify points that were reflected from the terrain, numerous algorithms have been developed in recent years. Many of them apply local operators that tend to face difficulties with complex scenes while their performance also varies between landscapes. In this paper, we present a filtering method that integrates a global approach using orthogonal polynomials with a local one that is region-based. The algorithm makes use of only a few parameters, and no fine-tuning is required between landscapes. Applying the algorithm over areas with varying topography and objects such as bridges, tunnels, and complex building, shows an improved performance compared to results obtained by others. This improvement is reflected in a lower than usual rate of misclassification errors for data acquired over different landscapes.

Robust Methods for Measurement of Leaf-Cover Area and Biomass from Image Data

Leaf-cover area is a widely required plant development parameter for predictive models of weed growth and competition. Its assessment is performed either manually, which is labor intensive, or via visual inspection, which provides biased results. In contrast, digital image processing enables a high level of automation, thereby offering an attractive means for estimating vegetative leaf-cover area. Nonetheless, image-driven analysis is greatly affected by illumination conditions and camera position at the time of imaging and therefore may introduce bias into the analysis. Addressing both of these factors, this paper proposes an image-based model for leaf-cover area and biomass measurements. The proposed model transforms color images into an illumination-invariant representation, thus facilitating accurate leaf-cover detection under varying light conditions. To eliminate the need for fixed camera position, images are transformed into an object–space reference frame, enabling measurement in absolute metric units. Application of the proposed model shows stability in leaf-cover detection and measurement irrespective of camera position and external illumination conditions. When tested on purple nutsedge, one of the worlds most troublesome weeds, a linear relation between measured leaf-cover area and plant biomass was obtained regardless of plant developmental stage. Data on the expansion of purple nutsedge leaf-cover area is essential for modeling its spatial growth. The proposed model offers the possibility of acquiring reliable and accurate biological data from digital images without extensive photogrammetric or image-processing expertise. Nomenclature: Purple nutsedge, Cyperus rotundus L. CYPRO

Correction of reflectance anisotropy: a multi-sensor approach

Quantitative mapping by means of hyperspectral remote sensing (HRS) can be hampered by reflectance anisotropy emerging in large field of view (FOV) optics, and may contain spectral radiometric distortions. This paper presents an algorithm for the rectification of reflectance anisotropy for rough terrain. A new method is offered for correction of radiometric bias caused by topography and sensing geometry. The correction of HRS data of lawn grass is demonstrated, and the method is tested on a large park area. To record elevation we used airborne laser scanning data to obtain a digital surface model (DSM). The Compact Airborne Spectral Imager (CASI) recorded reflectance of the same area. Anisotropy of reflectance was recorded by a laboratory spectro-goniometer. An analysis of the effect of correction on the normalized difference vegetation index (NDVI) shows that even moderate slopes, medium sensor FOV and high illumination conditions will result in reflectance anisotropy. Further analysis shows a clear inverse relationship between sensitivity of interpretation and spatial or spectral resolutions. We conclude with an outlook on the utilization of this method among other pre-processing tasks.

Temperature- and Radiation-Based Models for Predicting Spatial Growth of Purple Nutsedge (Cyperus rotundus)

Abstract Purple nutsedge is a troublesome C4 weed, characterized by high photosynthetic efficiency, compared to C3 weeds. As its dispersal is based on vegetative growth, accurate prediction of its growth could help in arriving at favorable management decisions. This article details the development and validation of predictive models of purple nutsedge spatial growth, based on temperature (thermal model), and temperature and radiation (photothermal model) measurements. Plants were grown in six experiments in the summers of 2008, 2009, and 2010, under different temperature and radiation conditions. Results indicate that under optimal temperatures, radiation becomes the main growth-limiting factor, and is highly related to the final leaf-cover area (R2  =  0.89). Comparison of the thermal and photothermal models showed that under all conditions, including varied temperature and radiation, the photothermal model performs significantly better, with differences in root-mean-square error values reaching up to 0.073, compared to 0.195 with the thermal model. Validation experiments confirmed the ability of the photothermal model to predict purple nutsedge spatial growth accurately. Nomenclature: Purple nutsedge, Cyperus rotundus L. CYPRO.

Detection of Sinkhole Hazards using Airborne Laser Scanning Data

Airborne laser scanning technology is primarily perceived as a means for gathering detailed three-dimensional information about the surface and objects on it. The dense 3D data contain information about surface features and geohazards, some of which are of subtle form. Geohazards are usually embedded within the terrain, and scarcely form distinct shape-transition to their surroundings; therefore their detection is challenging. We address in this paper detection of subtle terrain features and demonstrate it on collapse sinkholes. Collapse sinkholes are surface depressions whose formation has severe effect on the environment and on regional development. We present an autonomous model for their extraction and characterization. Sinkholes within the studied regions appear in various size, forms, from their embryonic to a well developed formation. The level of sinkhole detection is high, and as demonstrated, the model performs well under varying landforms and surface texture, with little influence on the correctness of the extracted sinkholes. As the results show, features of approximately 20 cm deep can be identified and separated from their surroundings in the data.

Reconstruction of Complex Shape Buildings from Lidar Data Using Free Form Surfaces

Building reconstruction from lidar data offers promising prospects for rapid generation of large-scale 3D models autonomously. Such reconstruction requires knowledge on a variety of parameters that refer to both the point cloud and the modeled buildings. The complexity of the reconstruction task has led researchers to use external information to localize buildings and assume that they consist of only planar parts. These assumptions limit the reconstruction of complex buildings, particularly those having curved faces. We present in this paper a detection and reconstruction model that considers the point cloud as the only information source and supports the reconstruction of general shape surfaces. Nonetheless, since many of the buildings are composed of planar faces, we maintain the planar based partitioning whenever possible and model non-planar surfaces only where needed. This way, standard models are extended to support free-form roof shapes without imposing artificial models. In addition to the free-form surface extension, we demonstrate the effect of imposing geometric constraints on the reconstruction as a means to generate realistic building models.

Extraction of Objects from Terrestrial Laser Scans by Integrating Geometry Image and Intensity Data with Demonstration on Trees

Terrestrial laser scanning is becoming a standard for 3D modeling of complex scenes. Results of the scan contain detailed geometric information about the scene; however, the lack of semantic details still constitutes a gap in ensuring this data is usable for mapping. This paper proposes a framework for recognition of objects in laser scans; aiming to utilize all the available information, range, intensity and color information integrated into the extraction framework. Instead of using the 3D point cloud, which is complex to process since it lacks an inherent neighborhood structure, we propose a polar representation which facilitates low-level image processing tasks, e.g., segmentation and texture modeling. Using attributes of each segment, a feature space analysis is used to classify segments into objects. This process is followed by a fine-tuning stage based on graph-cut algorithm, which considers the 3D nature of the data. The proposed algorithm is demonstrated on tree extraction and tested on scans containing complex objects in addition to trees. Results show a very high detection level and thereby the feasibility of the proposed framework.

Elimination of systematic errors from airborne laser scanning data

The paper discuses the elimination of systematic errors from airborne laser scanning data. Systematic errors affect the reliability of the data as well as subsequent processes; their elimination can almost be regarded as mandatory. The paper discusses and models the errors that can be found in airborne laser scanning systems and then analyzes their effect on the reconstructed surface. Following is a presentation of an error recovery model that is based on natural surfaces. The error recovery model is then extended to cope with the adjustment of laser swaths. Some of the topics that are addressed concern the tie and control of information and their incorporation into the model and automation of the process. The paper concludes with a demonstration of the adjustment.

Effect of Tuber Density and Trifloxysulfuron Application Timing on Purple Nutsedge (Cyperus rotundus) Control

Abstract Herbicides are the basis for conventional management of purple nutsedge, one of the worlds most troublesome weeds. However, as concern rises over their environmental impact, farmers are being required to reduce herbicide usage. Herbicide efficacy is strongly affected by weed growth stage and density at application, and when herbicides are applied under optimal conditions, low rates can provide maximal control efficacy (CE). Therefore, this study aimed to determine the time window for control of purple nutsedge using a low rate of herbicide, based on an effective degree days (EDDs) model, at low (one tuber) and high (10 tubers) densities. Two experiments were performed under field conditions, in the summers of 2009 and 2010. Rate of 3.75 g a.i. ha−1 trifloxysulfuron was applied once on each of five individual application dates. The growth of both treated and untreated plots was evaluated by means of leaf cover area (LCA) and biomass, which were then used to establish the time window for control. Results showed differences in both growth parameters between low and high tuber densities. The high-density patches reached LCA and fresh biomass values of 1,367 g and 1.12 m2, respectively, compared to 604 g and 0.69 m2, respectively, in the lower density patches. The favorable control periods based on biomass and LCA for the lower density patches were set to later dates than those for the higher density patches, 626 EDD compared to 483 EDD for biomass, and 786 EDD compared to 502 EDD for LCA, respectively. Although differences between the biomass- and LCA-based favorable control periods were observed at both tuber densities, the computed linear relations between the two growth parameters enabled adjusting them and setting the appropriate control period. Nomenclature: Trifloxysulfuron; Purple nutsedge, Cyperus rotundus L.; CYPRO.