Showing papers in "Photogrammetric Engineering and Remote Sensing in 2013"

Automatic camera calibration in close-range photogrammetry

Abstract: Automatic camera calibration via self-calibration with the aid of coded targets is now very much the norm in closerange photogrammetry. This is irrespective of whether the cameras to be calibrated are high-end metric, or the digital SLRs and consumer-grade models that are increasingly being employed for image-based 3D measurement. Automation has greatly simplified the calibration task, but there are real prospects that important camera calibration issues may be overlooked in what has become an almost black-box operation. This paper discusses the impact of a number of such issues, some of which relate to the functional model adopted for self-calibration, and others to practical aspects which need to be taken into account when pursuing optimal calibration accuracy and integrity. Issues discussed include interior orientation stability, calibration reliability, focal plane distortion, image point distribution, variation in lens distortion with image scale, colour imagery and chromatic aberration, and whether 3D object space control is warranted. By appreciating and accounting for these issues, users of automatic camera calibration will enhance the prospect of achieving an optimal recovery of scene-independent camera calibration parameters.

Predicting Surface Fuel Models and Fuel Metrics Using Lidar and CIR Imagery in a Dense, Mountainous Forest

Abstract: We compared the ability of several classification and regression algorithms to predict forest stand structure metrics and standard surface fuel models. Our study area spans a dense, topographically complex Sierra Nevada mixed-conifer forest. We used clustering, regression trees, and support vector machine algorithms to analyze high density (average 9 pulses/m 2 ), discrete return, small-footprint lidar data, along with multispectral imagery. Stand structure metric predictions generally decreased with increased canopy penetration. For example, from the top of canopy, we predicted canopy height (r 2 0.87), canopy cover (r 2 0.83), basal area (r 2 0.82), shrub cover (r 2 0.62), shrub height (r 2 0.59), combined fuel loads (r 2 0.48), and fuel bed depth (r 2 0.35). While the general fuel types were predicted accurately, specific surface fuel model predictions were poor (76 percent and 50 percent correct classification, respectively) using all algorithms. These fuel components are critical inputs for wildfire behavior modeling, which ultimately support forest management decisions. This comprehensive examination of the relative utility of lidar and optical imagery will be useful for forest science and management.

LASzip: Lossless Compression of Lidar Data

Abstract: This article describes how low flying aircraft equipped with modern laser-range scanning technology, which is also called Light Detection and Ranging (LIDAR), can collect precise elevation information for entire cities, counties, and even states By shooting 100,000 or more laser pulses per second into the Earth’s surface, these aircraft often take measurements at resolutions exceeding one point per square meter Derivatives of this data such as digital elevation models are used in numerous applications to assess flood hazards, to plan solar and wind applications, to carry out forest inventories, and to aid in power grid maintenance However, the shear amount of LIDAR data collected poses a significant challenge as billions of elevation samples need to be stored, processed, and distributed The article describes a lossless compression scheme for LIDAR in binary LAS format The compressed LAZ files are only 7 to 25 percent of the original size The encoding and decoding speeds are several million points per second Compression is streaming and decompression supports random-access On a national scale, the compression savings of storing LAZ instead of LAS can be measured in Petabyes of data, ie, data that no longer needs to be hosted, backed up, and served

Point-based Classification of Power Line Corridor Scene Using Random Forests

Abstract: The power line network, interconnecting power generation facilities and their end-users, is a critical infrastructure on which most of our socio-economic activities rely. As society becomes increasingly reliant on electricity, the rapid and effective monitoring of power line safety is critical. In particular, accurately knowing the current geometric and thermal status of power lines and identifying possible encroachments is the most important task in the power line risk management process. To facilitate this task, the correct identification of key objects comprising a power line corridor scene from remotely sensed data is the first important step. In recent years, airborne lidar has been successfully adopted as a cost-effective and accurate data source for mapping the power line corridors. However, in today’s practice, the classification of power line objects using lidar data still relies on labor-intensive data manipulation, and its automation is urgently required. To address this problem, this paper proposes a point-based supervised classification method, which enables the identification of five utility corridor objects (wires, pylons, vegetation, buildings, and low objects) using airborne lidar data. A total of 21 features were investigated to illustrate the horizontal and vertical properties of power line objects. A non-parametric discriminative classifier, Random Forests model, was trained with refined features to label raw laser point clouds. The proposed classifier showed 91.04 percent sample-weighted and 90.07 percent class-weighted classification accuracy, which indicates it could be highly valuable for large-scale, rapid compilations of corridor maps. A sensitivity analysis of the proposed classifier suggested that when compared, training with class-balanced samples improves classification performance over training with unbalanced samples, particularly with corridor objects such as wires and pylons.

Web-service-based Monitoring and Analysis of Global Agricultural Drought

Abstract: It is of great importance and an urgent demand to enable operational and near real-time monitoring and analysis of global agricultural drought at desirable spatial and temporal resolutions. Traditional approaches and existing systems are not able to meet the demand because of big-data and geoprocessing-modeling challenges. The latest advances in Web service, geospatial interoperability and cyberinfrastructure technologies and the availability of near real-time global remote sensing data have shown potential to address the challenges and meet the demand. This paper presents a Web service approach to building the Global Agricultural Drought Monitoring and Forecasting System ( GADMFS), an open, interoperable, and on-demand geospatial Web service system, for meeting the demand. The big-data and geoprocessingmodeling issues in providing complete agricultural drought information are resolved in GADMFS through improved data-, service- and system-level interoperability and servability. GADMFS is able to overcome major limitations of current drought information systems in the world and better support decision making with improved global agricultural drought data and information dissemination and analysis services.

An Automatic Approach to UAV Flight Planning and Control for Photogrammetric Applications

Abstract: This paper deals with the implementation of an automatic and novel approach for UAV flight planning and control based on photogrammetric principles. Software which performs multiple tasks related to aerial photogrammetry and particularized to UAV systems was developed. Specifically, the flight planning and control framework incorporates a robust geometric control that guarantees the effectiveness, precision, and reliability of image acquisition. A real case study was generated to test and validate the flight planning and control strategies developed here.

Parcel-Level Identification of Crop Types Using Different Classification Algorithms and Multi-Resolution Imagery in Southeastern Turkey

Abstract: This research investigates the accuracy of pixel- and object-based classifi cation techniques across varying spatial resolutions to identify crop types at parcel level and estimate the area at six test sites to fithe optimum data source for the identifi cation of crop parcels. Multi-sensor data with spatial resolutions of 2.5 m, 5 m and 10 m from SPOT5 and 30 m from Landsat-5 TM were used. Maximum Likelihood (ML), Spectral Angle Mapper (SAM), and Support Vector Machines (SVM) were used as pixel-based methods in addition to object-based image classifi cation (OBC). Post-classifi cation methods were applied to the output of pixel-based classifi cation to minimize the noise effects and heterogeneity within the agricultural parcels. In addition, processing-time performance of the algorithms was evaluated for the test sites and district scale classifi cation. OBC results provided comparatively the best performance for both parcel identifi cation and area estimation at 10 m and fi ner spatial resolution levels. SVM followed OBC at 2.5 m and 5 m resolutions but accuracies decreased dramatically with coarser resolutions. ML and SAM results were worse up to 30 m resolution for both crop type identifi cation and area estimation. In general, parcel identifi cation effi ciency was strongly correlated with spatial resolution while the classifi cation algorithm was a more effective factor than spatial resolution for area estimation accuracy. Results also provided an opportunity to discuss the effects of image resolution and the classifi cation algorithm independent factors such as parcel size, spatial distribution of crop types and crop patterns.

Robust Affine-Invariant Line Matching for High Resolution Remote Sensing Images

Abstract: Point-based matching methods usually hold limitation in dealing with low texture scenes. In this paper, a robust affi neinvariant lines matching method is proposed. The method commences with line segments extraction. All the extracted line segments are grouped into salient lines and general lines. Accordingly, the matching procedure includes salient lines matching and general lines matching. In salient lines matching, affi ne invariants are calculated and the matched salient line correspondences are the basis of the general lines matching. Each general line is clustered into a matched salient line according to a certain rule. Taking each salient line as the root, together with all the general lines clustered to it, a control network is constructed. Finally, the general lines matching procedure is performed between the two subnetworks whose roots are correspondences. Experimental results show that our proposed method can successfully process local distortion and improve the matching performance in low texture areas.

A Framework for the Registration and Segmentation of Heterogeneous Lidar Data

Abstract: This article describes how Light Detection and Ranging (LIDAR) has been established as a mainstream tool for the acquisition of three dimensional point data over the past few years. Besides the conventional mapping missions, LIDAR has also proved to be very effective for a wide range of applications such as forestry, urban planning, structural deformation analysis, and reverse engineering. In the context of a national dataset, it is safe to assume that multiple laser scanners are under different conditions in order to collect data. Current registration and segmentation algorithms assume homogeneity in the local point density and accuracy, which is an invalid assumption that cannot be tolerated. As a consequence of the wide range of LIDAR sensors that are currently available, it is becoming crucial to develop algorithms for the registration and segmentation of LIDAR data with significantly varying characteristics, for example, varying point density and accuracy. A methodology for the optimal registration and segmentation of heterogeneous LIDAR data is presented in this article. An example of integrating airborne and terrestrial laser scans is also presented, which is followed by a discussion of the pros and cons of the integration process.

The Effects of Data Selection and Thematic Detail on the Accuracy of High Spatial Resolution Wetland Classifications

Abstract: Accurate wetland maps are of critical importance for preserving the ecosystem functions provided by these valuable landscape elements. Though extensive research into wetland mapping methods using remotely sensed data exists, questions remain as to the effects of data type and classifi cation scheme on classifi cation accuracy when high spatial resolution data are used. The goal of this research was to examine the effects on wetland mapping accuracy of varying input datasets and thematic detail in two physiographically different study areas using a decision tree classifi er. The results indicate that: topographic data and derivatives signifi cantly increase mapping accuracy over optical imagery alone, the source of the elevation data and the type of topographic derivatives used were not major factors, the inclusion of radar and leaf-off imagery did not improve mapping accuracy, and increasing thematic detail resulted in signifi cantly lower mapping accuracies i.e., particularly in more diverse wetland areas.

Developing Efficient Procedures for Automated Sinkhole Extraction from Lidar DEMs

Abstract: Sinkhole detection in karst areas is usually difficult through remote sensing image interpretation. We present an efficient approach to extract mature sinkholes from lidar DEM. First, an adaptive Wiener filter (AWF) and hierarchical watershed segmentation (HWs) are applied to identify all local depression or potential sinkholes. Second, a hole-filling algorithm is applied to the potential sinkholes, and nine spatial features are extracted. Finally, the random forest classifier is used to select true sinkholes from all potential sinkholes. Our results show that this approach is efficient for detecting mature sinkholes from lidar data, and it can be used for risk assessment and hazard preparedness in karst areas.

Generation and Dissemination of a National Virtual 3D City and Landscape Model for the Netherlands

Abstract: This paper describes the generation and dissemination of a national three-dimensional (3D) dataset representing the virtual and landscape model. The 3D model is produced automatically by fusing a two-dimensional (2D) national objectoriented database describing the physical landscape and the national high-resolution height model of the Netherlands. Semantic constraints are introduced to correctly model 3D objects. Three areas from different regions in the Netherlands have been processed in order to develop, improve, and test the automatic generation of a national 3D city and landscape model. Specific attention has been paid to exceptional cases that may occur in a nationwide dataset. Based on the test results, the Kadaster, the national agency in the Netherlands responsible for the production of nation wide geo-information, decided that it is feasible to produce a national 3D city and landscape model that fulfills the specifications that were defined as part of this study. Future research is identified to make the results further ready for practice.

Change Detection and Deformation Analysis in Point Clouds

Abstract: The paper outlines a method to compare two digital surfaces of the same rock face to detect major changes resulting from detached rocks and deformations. A terrestrial laser scanning survey is used for data gathering. After georeferencing, if the cliff has a complex morphology, a 3D segmentation algorithm is applied to split the whole rock surface into more subregions with an almost planar structure. In each subregion the raw point cloud is resampled on a regular grid and multitemporal differences are analyzed. Anomalies in differences, which should be very close to zero if no geometric variations have occurred, are identified with the following purposes: (a) localizing gross changes due to rock detachments, (b) removing global rigid-body displacements, and (c) understanding local cliff deformations. In the case where the rock face is covered by vegetation, this has to be filtered out, e.g., by visual inspection of RGB images co-registered to the point cloud. This paper also describes a procedure to carry out vegetation filtering in automatic way from the analysis of near-infrared images captured by a camera integrated to laser scanner. The application of the full processing pipeline has been tested on a real case study located in the Italian pre-alpine area. Here, after filtering some vegetation, a total rock fall volume of 0.15 m3 was detected on a cliff of about 375 m2 and within a period of six months.

Developing an Object-based Hyperspatial Image Classifier with a Case Study Using WorldView-2 Data

Abstract: Recent advancements in remote sensing technology have provided a plethora of very high spatial resolution images. From pixel-based processing designed for low spatial resolution data, image processing has shifted towards object-based analysis in order to adapt to the hyperspatial nature of currently available remote sensing data. However, standard object-based classifiers work with only object-level summary statistics of the reflectance values and do not sufficiently exploit within-object reflectance pattern. In this research, a novel approach of utilizing the object-level distribution of reflectance values is presented. A fuzzy Kolmogorov-Smirnov based classifier is proposed to provide an object-to-object matching of the empirical distribution of the reflectance values of each object and derive a fuzzy membership grade to each class. This object- based classifier is tested for urban objects recognition from WorldView-2 data. Results indicate at least 10 percent increase in overall classification accuracy using the proposed classifier in comparison to various popular object- and pixel-based classifiers.

The influence of multi-season imagery on models of canopy cover: A case study

Abstract: Quantifying tree canopy cover in a spatially explicit fashion is important for broad-scale monitoring of ecosystems and for management of natural resources. Researchers have developed empirical models of tree canopy cover to produce geospatial products. For subpixel models, percent tree canopy cover estimates (derived from fi ne-scale imagery) serve as the response variable. The explanatory variables are developed from reflectance values and derivatives, elevation and derivatives, and other ancillary data. However, there is a lack of guidance in the literature regarding the use of leaf-on only imagery versus multi-season imagery for the explanatory variables. We compared models developed from leaf-on only Landsat imagery with models developed from multi-season imagery for a study area in Georgia. There was no statistical difference among models. We suggest that leaf-on imagery is adequate for the development of empirical models of percent tree canopy cover in the Piedmont of the Southeastern United States.

Tweeting up a storm: the promise and perils of crisis mapping

Abstract: Crisis mapping is an inter-disciplinary field that aggregates crowd-generated input data, such as social media feeds and photographs, with geographic data, to provide real-time, interactive information in support of disaster management and humanitarian relief. This article, published in the October 2013 issue of Photogrammetric Engineering & Remote Sensing, provides a brief overview of the emerging legal and ethical issues within crisis mapping.

It Used To Be Dark Here

Abstract: Nighttime satellite imagery from the Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) has a unique capability to observe nocturnal light emissions from sources including cities, wild fires, and gas flares. Data from the DMSP OLS is used in a wide range of studies including mapping urban areas, estimating informal economies, and estimations of population. Given the extensive and increasing list of applications a repeatable method for assessing geolocation accuracy would be beneficial. An array of portable lights was designed and taken to multiple field sites known to have no other light sources. The lights were operated during nighttime overpasses by the DMSP OLS and observed in the imagery. An assessment of the geolocation accuracy was performed by measuring the distance between the GPS measured location of the lights and the observed location in the imagery. A systematic shift was observed and the mean distance was measured at 2.9 km.

3D Tree Reconstruction from Simulated Small Footprint Waveform Lidar

Abstract: Lidar-based 3D tree reconstruction enables the retrieval of detailed tree structure; however, many existing methods are based on high-density discrete return lidar datasets. In this paper, we propose the use of small footprint waveform lidar data to achieve branch-level tree reconstruction for both leaf-off and leaf-on conditions. The DIRSIG simulation environment was used for algorithm validation purposes. Leaf-off data served as reference, and leaf-on reconstruction for a particular tree resulted in an average branch length difference of 0.07 m and an average angular difference of approximately 6 degrees for both tilt and azimuth angles. Compared to in situ methods this approach may be used by an airborne system for accurate estimation of forest biomass, forest inventory, land degradation, etc. in large scale applications. Furthermore, since this approach can also be applied on leaf-on trees, the tree skeleton characterization eventually can be conducted year round and will be less dependent on seasonal changes.

New Approaches for Estimating the Local Point Density and its Impact on Lidar Data Segmentation

Abstract: This article describes how Light Detection and Ranging (LIDAR) systems have been created for the rapid collection of high density three-dimensional (3D) point cloud data over the past few years. The advent of these systems has reduced the cost and increased the availability of accurate 3D data for diverse applications such as terrain mapping, transportation planning, emergency response, 3D city modeling, heritage documentation, forest parameter estimation, flood hazard mapping, and coastal management. Usually, the original LIDAR point cloud does not comprise semantic information about the type and characteristics of reflecting surfaces. Therefore, the data that has been collected should be processed to extract useful information for the applications mentioned above, such as ground and non-ground classification, Digital Terrain Model (DTM) generation, and building hypothesis generation.

A Flexible Method for Zoom Lens Calibration and Modeling Using a Planar Checkerboard

Abstract: This paper presents a fl exible method for zoom lens calibration and modeling using a planar checkerboard. The method includes the following four steps. First, the principal point of the zoom-lens camera is determined by a focus-of-expansion approach. Second, the infl uences of focus changes on the principal distance are modeled by a scale parameter. Third, checkerboard images taken at varying object distances with convergent image geometry are used for camera calibration. Finally, the variations of the calibration parameters with respect to the various zoom and focus settings are modeled using polynomials. Three different types of lens are examined in this study. Experimental analyses show that high precision calibration results can be expected from the developed approach. The relative measurement accuracy (accuracy normalized with object distance) using the calibrated zoom-lens camera model ranges from 1:5 000 to 1:25 000. The developed method is of signifi cance to facilitate the use of zoom-lens camera systems in various applications such as robotic exploration, hazard monitoring, traffi c monitoring, and security surveillance.

Influence of a Dense, Low-Height Shrub Species on the Accuracy of a Lidar-Derived DEM

Abstract: Airborne lidar provides an effective platform for collecting elevation data. However, the accuracy of lidar-derived digital elevation models (DEMs) can be adversely affected by natural conditions as well as methods used to process the data. Using a lidar dataset from a mountainous region of southwest Idaho, this study extends previous assessments of DEM accuracy with a focused investigation of a specific dense, low-height shrub species (Ceanothus velutinus). Bare-earth elevations were collected using survey-grade CPS and compared to lidar­ derived elevations to assess DF:M accuracy. Results suggest that the magnitude of elevation error varied depending on morphological characteristics of ceanothus, terrain slope, and filtering parameters used to process the lidar data. When using optimal filtering parameters, root mean square error (RMSE2 ) was largest in areas of ceanothus cover, ranging from 0.17 to 0.26 m in slopes <25° and 0.28 to 0.37 m in slopes ?25°. An examination of lidar returns found that ceonothus obstructed laser pulse penetration and few returns reached the ground surface. In areas of ceanothus cover, we conclude that the obstruction of the ground surface contributed to filtering errors, which resulted in mislabeled ground returns and decreased accuracy in bore-earth DEMs. These results have implications for the use of lidar-derived DF:Ms in areas of ceonothus throughout western North America, and in ecosys­ tems with similar dense shrub cover. Introduction and Background Airborne laser scanning, also referred to as lidar (Light Detection and _Ranging), is an active remote sensing technol­ ogy capable of collecting detailed three-dimensional in forma­ tion about the Earth 's surface. Small-footprint, d iscrete-return systems have the abi lity to penetrate surface vegetation and yield multiple returns from the canopy and underlying ter­ rain. The dense collection of elevation data makes lidar an attractive data source for the production of high-resolution digital elevation models (DEMs) used in many geographic information system (G1s] applications. While the vertical accu­ racy of many lidar systems is commonly quoted as ~0.15 m (Baltsavias, 1999). such accuracy is typically only achievable under the most ideal circumstances (Hodgson and Bresnahan, 2004). Several studies (Reutebuch et al., 2003; Hodgson and Bresnahan, 2004; Su and Bork, 2006; Raber et al. , 2007; Bater and Coops, 2009; Spaete et al., 2011) have analyzed the Samuel B. Gould, Nancy F. Glenn , Temuulen T. Sankey, and Lucas P. Spaete are with the Department of Geosciences, Boise Center Aerospace Laboratory, Idaho State University, 322E. Front Street, Boise, ID 83702 (goulsamu@isu.edu). James P. McNamara is with the Department of Geosciences, Boise State University, 1910 University Drive, Boise, ID 83725. accuracy of lidar-derived DEMs and found natural conditions such as land -cover and terrain slope, as well as lidar poin t processing, to be significant influential factors on accuracy. Previous studies investigating the influence of land-cover (Hodgson et al. , 2003; Su and Bork, 2006) have categorized vegetation according to similar characteristics (e.g., low grass, h igh grass, shrub, coniferous, deciduous, etc.], which provide generalized error predictions, but may overlook unique inter­ actions between the laser beam and a particular species of vegetation. Analyzing error within specific types of vegetation is necessary to understand and quantify the accuracy of lidar­ derived ground surfaces in semiarid mountainous ecosystems, where dense, broadleaf shrubs coincide with sparse shrub­ steppe and mixed-forest communities. Morphological characteristics of vegetation such as ceanothus (Ceanothus velutinus) present unique challenges for the accurate generation of DEMs using lidar data. Few openings in the canopy are large enough for lidar pulses (~0.20m diameter ) to pass through and the oval-shaped, waxy texture of the leaves (Figure 1) highly reflect and attenuate laser beam irradiance. As a result, the dense stand characteris­ t ics and leaf orientation of ceanothus can prevent lidar pulses from reaching the ground surface, which potentially leads to mistakenly-labeled ground returns and subsequent error in DEM elevations . Low-height vegetation can also be problematic because of the small elevation differences between the top of the vegetation and underlying ground surface. Two lidar ech­ oes can only be discriminated if their distance is larger than half of the p ulse length (Beraldin, 2010), which is commonly on the order of 3 to 6ns or 1 to 2 m (Liu, 2008). As a conse­ quence, the reflected energy of a pulse from the vegetation becomes comingled with the reflected energy from the ground surface, and the system is unable to discern more than one return in short vegetation (Schmid et al., 2011). A well-documented characteristic of observed eleva­ tion error in lidar data is the relationship with terrain slope (Maling 1989). As slope increases, \"apparent\" error may also increase due to horizontal error in the lidar observation . Spaete et al. (2011) examined the effects of slope and vegeta­ tion on the accuracy of a lidar-derived DEM and calculated potential vertical error (PYE) using the average slope for each of their vegetation categories. They found a low PVE (0.0260.037 m) in slopes <10° and a higher PYE (0.075 to 0.12 m) in slopes >10°. Root mean square error (RMSE2 ) values for slopes >10° were roughly twice than those for slopes <10°, suggest­ ing that PYE likely contributed to the overall RMSE2 • Hodgson and Bresnahan (2004) and Su and Bork (2006), respectively, Photogrammetric Engineering & Remote Sensing Vol. 79, No. 5, May 2013, pp. 421431. 0099-1112/13/7905-421/$3.00/0 © 2013 American Society for Photogrammetry and Remote Sensing PHOTOGRAMMETRIC ENGINEERING & REMOTE SENS ING May2013 421 Figure 1. Photograph of Ceanothus velutinus in the natural envi ronment and magnified inset of ceano­ thus leaves. Inset map identifies states and provinces in the western US and Canada where ceanothus can be found. found similar results, with RMSE2 on slopes >25° twice those found on relatively flat areas and RMSE2 of slopes > 10° twice those found on slopes <2°. In nearly all lidar applications, height-fi ltering of the 30 point cloud is a necessary process to determine which lidar returns are from the ground surface and which are from non-ground features. Classifying the lidar data is a crit ical step for DEM generation, but the development of methods to accurately separate returns has proven to be difficult. A common assumption of many height-filteri ng algorithms is that the lowest return in a specified neighbor­ hood represents the ground surface (Meng et al. , 2010). This method can be problematic in areas of complex vegetation and varying topography where the lowest return may not truly represent the ground surface. Many fi ltering algorithms have been proposed with none performing equally well for all landscapes (Forlani and Nardinocchi, 2007). Thus, an understanding of the fi ltering method used and its ability to accurately separate returns within specific vegetation and varying terrain is necessary for a viable DEM accuracy assessment. Several studies (Glenn et al. , 2 011 ; Mitchell et al., 2011; Sankey and Bond, 2011; Spaete et al., 2011) have used the same height-filtering method as imple­ mented in this research determining optimal parameters based on study-specific accuracy. Al though previous studies have examined the accuracy of lidar-derived DEMs across varying terrain and land cover (Reutebuch et al., 2003; Hodgson and Bresnahan, 2004; Bater and Coops, 2009; Spaete et al., 2011). relatively few have focused on quantifying the error for an individual species of vegetation (Su and Bork, 2006). Furthermore, little research has investigated the effects in semiarid mountainous land­ scapes where complex topography (slopes >25°) and dense, low-height shrubs are abundant. Our study investigated the influence of ceanothus (Ceanothus ve/utinus). a native broadleaf evergreen shrub, on the accuracy of a lidar-derived DEM. With populations of ceanothus found in several moun­ tainous ecosystems throughout the western US and Can ada (Figure 1). such an investigation has important implications for lidar applications at a regional scale. This investigation may also hold relevance for ecosystems where similar but unrelated species of vegetation exist and interact with lidar in much the same manner. The objectives of this study were to: (a) determine if ceanothus introduces elevation error in derived DEMs, (bl quantify the vertical accuracy for derived OEMS within combinations of ceanothus cover and terrain slope, (cl explore the influence of filtering parameters on DEM accuracy, and (d) determine lidar pulse penetration by analyzing the vertical distribution of returns within ceano­ thus canopy. 422 M a y 2013 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENS ING

An Automatic Approach to UAV Flight Planning and Control for Photogrammetric Applications: A Test Case in the Asturias Region (Spain)

Abstract: This paper deals with the implementation of an automatic and novel approach for unmanned aerial vehicle (UAV) flight planning and control based on photogrammetric principles. Software which performs multiple tasks related to aerial photogrammetry and particularized to UAV systems was developed. Specifically, the flight planning and control framework incorporates a robust geometric control that guarantees the effectiveness, precision, and reliability of image acquisition. A real case study was generated to test and validate the flight planning and control strategies developed here.

Towards Integration of GLAS into a National Fuel Mapping Program

Abstract: Comprehensive canopy structure and fuel data are critical for understanding and modeling wildland fire. The LANDFIRE project produces such data nationwide based on a collection of field observations, Landsat imagery, and other geospatial data. Where field data are not available, alternate strategies are being investigated. In this study, vegetation structure data available from GLAS were used to fill this data gap for the Yukon Flats Ecoregion of interior Alaska. The GLAS-derived structure and fuel layers and the original LANDFIRE layers were subsequently used as inputs into a fire behavior model to determine what effect the revised inputs would have on the model outputs. The outputs showed that inclusion of the GLAS data enabled better landscape-level characterization of vegetation structure and therefore enabled a broader wildland fire modeling capability. The results of this work underscore how GLAS data can be incorporated into LANDFIRE canopy structure and fuel mapping.

Registration of Optical Images with Lidar Data and Its Accuracy Assessment

Abstract: Photogrammetry and lidar are two technologies complementary for 3D reconstruction. However, the problem is that the current registration methods of optical images with lidar data cannot satisfy all the requirements for the fusion of the above two technologies, especially for close-range photogrammetry and terrestrial lidar. In this paper, we propose a novel method for registration of optical images with terrestrial lidar data, which is implemented by minimizing the distances from the photogrammetric matching points to terrestrial lidar data surface, with the collinearity equation as the basic mathematical model. One advantage of this method is that it requires no feature extraction and segmentation from the lidar data. Another advantage is that non-rigid deformation caused by lens distortion can be eliminated through the use of bundle adjustment similar to self-calibration. In addition, experiments with two different data sets show that this method cannot only eliminate the influence of certain gross errors, but also offer a high accuracy of 3 mm to 5 mm. Therefore, the proposed registration method is proved to be more effective, accurate, and reliable.

A Combined Object-and Pixel-Based Image Analysis Framework for Urban Land Cover Classification of VHR Imagery

Abstract: This paper aims at exploiting the advantages of pixel-based and object-based image analysis approaches for urban land cover classification of very high resolution (vHR) satellite imagery through a combined object- and pixel-based image analysis framework. The framework starts with segmenting the image resulting in several spectral and spatial features of segments. To overcome the curse of dimensionality, a wavelet ­ based feature extraction method is proposed to reduce the number of features. The wavelet-based method is automatic, fast, and can preserve local variations in objects' spectral/ spatial signatures. Finally, the extracted features together with the original bands of the image are classified using the conventional pixel-based Maximum Likelihood classifica­tion. The proposed method was tested on the WorldView-2, QuickBird, and Ikonos images of the same urban area for comparison purposes. Results show up to 17 percent, 10 percent, and 11 percent improvement in kappa coefficients compared to the case in which only the original bands of the image are used for WV-2, QB, and IK, respectively. Furthermore, the objects' spectral features contribute more to increasing classification accuracy than spatial features.

Range Image Techniques for Fast Detection and Quantification of Changes in Repeatedly Scanned Buildings

Abstract: A u g u s t 2 0 1 3 1 evolution that require a more subtle analysis of the measured scene. Several deformation analysis approaches for designated objects have been discussed in recent years. An approach that employs terrestrial laser scanning data of a lock gate was presented by Schäfer et al. (2004). This method enabled deformation detection through the construction of a uniform and regular grid from each scattered point cloud. Instead of simple thresholding of differences, Lindenbergh and Pfeifer (2005) proposed a method to detect deformation below the nominal single point measurement accuracy using statistical deformation analysis based on the comparison of laser points and fitted planar segments between different epochs. Similarly, Kenner et al. (2011) presented a detailed spatial (by superficially extending the movement) and temporal (the number of epochs) analysis to detect changes smaller than the single point accuracy, which was applied in the monitoring of erosive processes in a rock wall. Gosliga et al. (2006) implemented deformation analysis on a bored tunnel by means of terrestrial laser scanning that included an analysis of both deformation between different epochs and deformations of the fitted tunnel with respect to the design model. GirardeauMontaut et al. (2005) carried out a comparison study by using the Hausdorff distance as a measure for changes. They noted three issues to be addressed: point sampling variations between scans, computational costs of the Hausdorff distance, and real change discrimination. Zeibak and Filin (2007) also analyzed the potential artifacts that may affect the detection, which include resolution and object pose variations, occlusion and scanner-related artifacts (e.g., regions of no reflectance, range limitations, and noise). They proposed an approach for change detection through the use of TLS data in which the comparison is conducted through a mere image subtraction between the range images of a reference scan and the analyzed scan transformed into the reference frame to eliminate the artifacts mentioned above. As this method assumes that the transformation parameters are known, the reliability of image subtraction depends greatly on the registration accuracy between reference and analyzed scans. In general, the challenges of change detection of buildings sampled by terrestrial laser scans comprise the irregular point distribution, the varying scale and resolution within the scene (depending on depth), the large data volume in each scan and falsely detected changes caused by factors such as the occlusion of objects or object parts, “no-reflectance” regions, range limits, noise, and errors in the registration Abstract This paper proposes a new method for the detection and quantification of changes in buildings using terrestrial laser scanning data from different epochs. A refined registration process is implemented that utilizes an optimized version of the Iterative Closest Point (ICP) algorithm, which implements the search of adjacent points in terms of their scanning angles. For detecting changes, a novel 2d angular difference histogram is proposed to first determine point segments representing building parts from the raw scattered scans. Afterwards, Hausdorff distance-based change detection is innovatively integrated into the optimized ICP process to improve the efficiency of the entire algorithm. The detected changes are quantified in the final step by determining the total planar surface area of the changed facade. This approach is tested and illustrated on two real datasets. The change quantification results show that the accuracy of the changed area quantification is in the order of square centimeters.

Moving Toward Consistent ALS Monitoring of Forest Attributes across Canada

Abstract: As airborne laser scanning (ALS) gains wider adoption to support forest operations in Canada, the consistency and quality of derivative products that support long-term monitoring and planning are becoming a key issues for managers. The Canadian Consortium for Lidar Environmental Applications Research (C-CLEAR) has supported almost 200 projects across Canada since 2000, with forest-related studies being a dominant theme. In 2010 and 2011, field operations were mobilized to support 13 ALS projects spanning almost the full longitudinal gradient of Canada’s forests. This paper presents case studies for seven plus an overview of some best practices and data processing workflow tools that have resulted from these consortium activities. Although the projects and research teams are spread across Canada, the coordination and decade of experience provided through C-CLEAR have brought common methodological elements to all. It is clear that operational, analytical and reporting guidelines that adhere to community accepted standards are required if the benefits promised by ALS forestry are to be realized. A national Lidar Institute that builds upon the C-CLEAR model and focuses on developing standards, guidelines, and certified training would address this need.