Bon Dewitt

A Novel Approach to Terrestrial Lidar Georeferencing

Abstract This paper describes a novel method for determining theabsolute orientation of lidar point clouds using GPS measure-ments from two antennas firmly mounted on the opticalhead of a lidar scanner. The solution is linear and isderived from the non-linear georeferencing model by exploit-ing the properties of the skew-symmetric matrix. Simulationand real world experimentation using our prototype suggesta precision of about 0.05° ( 1 mrad) for the three Eulerattitude angles. The method can help alleviate problemsassociated with the conventional technique and can allowfor an increased number of practical applications forgeoreferenced terrestrial lidar. Introduction The current conventional methods for georeferencingterrestrial lidar point clouds utilize reflective targets withknown coordinates. Typically, a coordinate transformationfrom the scanner’s own coordinate system ( SOCS ) to mappingor global coordinate system ( GLCS ) can be solved using aminimum of three scanned points with known

A new approach for pass-point generation from aerial video imagery.

This paper presents a novel approach for automatically finding conjugate points between video images collected by a small autonomous unmanned aircraft. Our approach introduces the idea of saving the resampled patch from successive least-squares matching epochs and using them as templates for subsequent images. Tests show that this method is superior to using the first image as a template for all subsequent matching attempts. We show how the algorithm performs in terms of retention of points on successive images, distribution of points on the images, and utility when used for bundle adjustments in comparison with the conventional method of using the first image as a template. Our proposed method is able to match points on an average of 2.7 times as many images before failure compared with using the conventional method. This leads to stronger geometrical configuration, higher redundancy, and ultimately, significantly better bundle adjustment solutions.

Classifying spatially heterogeneous wetland communities using machine learning algorithms and spectral and textural features

Mapping of wetlands (marsh vs. swamp vs. upland) is a common remote sensing application.Yet, discriminating between similar freshwater communities such as graminoid/sedge from remotely sensed imagery is more difficult. Most of this activity has been performed using medium to low resolution imagery. There are only a few studies using high spatial resolution imagery and machine learning image classification algorithms for mapping heterogeneous wetland plant communities. This study addresses this void by analyzing whether machine learning classifiers such as decision trees (DT) and artificial neural networks (ANN) can accurately classify graminoid/sedge communities using high resolution aerial imagery and image texture data in the Everglades National Park, Florida. In addition to spectral bands, the normalized difference vegetation index, and first- and second-order texture features derived from the near-infrared band were analyzed. Classifier accuracies were assessed using confusion tables and the calculated kappa coefficients of the resulting maps. The results indicated that an ANN (multilayer perceptron based on back propagation) algorithm produced a statistically significantly higher accuracy (82.04 %) than the DT (QUEST) algorithm (80.48 %) or the maximum likelihood (80.56 %) classifier (α<0.05). Findings show that using multiple window sizes provided the best results. First-order texture features also provided computational advantages and results that were not significantly different from those using second-order texture features.

A tool for rapid post-hurricane urban tree debris estimates using high resolution aerial imagery

Abstract Coastal communities in the southeast United States have regularly experienced severe hurricane impacts. To better facilitate recovery efforts in these communities following natural disasters, state and federal agencies must respond quickly with information regarding the extent and severity of hurricane damage and the amount of tree debris volume. A tool was developed to detect downed trees and debris volume to better aid disaster response efforts and tree debris removal. The tool estimates downed tree debris volume in hurricane affected urban areas using a Leica Airborne Digital Sensor (ADS40) and very high resolution digital images. The tool employs a Sobel edge detection algorithm combined with spectral information based on color filtering using 15 different statistical combinations of spectral bands. The algorithm identified downed tree edges based on contrasts between tree stems, grass, and asphalt and color filtering was then used to establish threshold values. Colors outside these threshold values were replaced and excluded from the detection processes. Results were overlaid and an “edge line” was placed where lines or edges from longer consecutive segments and color values within the threshold were met. Where two lines were paired within a very short distance in the scene a polygon was drawn automatically and, in doing so, downed tree stems were detected. Tree stem diameter–volume bulking factors were used to estimate post-hurricane tree debris volumes. Images following Hurricane Ivan in 2005 and Hurricane Ike in 2008 were used to assess the error of the tool by comparing downed tree counts and subsequent debris volume estimates with post-hurricane photo-interpreted downed tree counts and actual field measured estimates of downed tree debris volume. The errors associated with the use of the tool and potential applications are also presented.

Accuracy Evaluation of Terrestrial LIDAR and Multibeam Sonar Systems Mounted on a Survey Vessel

AbstractThis research provides a performance test of terrestrial light detection and ranging (LIDAR) and multibeam echo sounder scanners integrated with Global Navigation Satellite System/Inertial Navigation System positioning and orientation on a survey vessel platform. To measure the accuracy of the data, experiments were designed to allow the LIDAR and sonar scanners to acquire scans of a control target that extended above and below the water surface. The scans were acquired under normal and induced conditions expected in a marine survey environment, such as variations in speed, range, motion, and orientation. SD, root-mean-square error (RMSE), and mean were computed across all data sets for each experiment. Horizontal RMSE values of 0.06 and 0.03 m were achieved for the LIDAR and sonar data, respectively. Vertical RMSE results of 0.04 m were found for both data types. These results were comparable with previous mobile mapping research involving similar systems. Contributing uncertainty and error sourc...

Using satellite imagery and LIDAR data to corroborate an adjudicated ordinary high water line

Determination of the ordinary high water line (OHWL) has been and continues to be an important issue. The OHWL defines the separation of sovereignty lands and private ownership on non‐tidal water bodies. Determination of OHWL is conducted on a case‐by‐case basis in Florida through court challenges. A judge makes the decision on where the line exists based upon several criteria—including remote sensing data. This study investigated the possibility of using various remote sensing technologies to provide an efficient and accurate means of determining OHWL for a lake in central Florida. Landsat Enhanced Thematic Mapper (ETM) satellite imagery was compared with the higher resolution imagery IKONOS and Light Detection And Ranging (LIDAR) imagery in order to determine the waters edge and location of vegetation communities that may be correlated with OHWL. It was found that ETM imagery could be used only for mapping vegetation community transition zones and that this zone provided limited insight to OHWL. IKONOS imagery, on the other hand, was more promising for land cover mapping, but requires further study in order to draw general conclusions regarding its application to OHWL. LIDAR data provided the best results for determining OHWL, but also need further study over a larger area in order to draw final conclusions.

Evaluating the potential of multi-view data extraction from small Unmanned Aerial Systems (UASs) for object-based classification for Wetland land covers

Unmanned Aerial Systems (UASs) have the potential to provide multi-view data, but the approaches used to extract the multi-view data from UAS and investigation of their use in image classification are currently unavailable in publications to our best knowledge. This study presents a method that combines collinearity equations and a two-phase optimization procedure to automatically project a point from real world coordinate system of an orthoimage to UAS image coordinate system (row and column numbers) to be used in multi-view data extraction. The results show average errors for the computed UAS column and row numbers were 1.6 and 1.8 pixels respectively evaluated with leave-one-out method. Based on this algorithm, it’s also for the first time that object-based multi-view data were extracted and presented, and the potential of using the multi-view data to aid Geographic Object-Based Image Analysis(GEOBIA) through bidirectional reflectance distribution function (BRDF) modelling was evaluated with two representatives of BRDFs, the Rahman-Pinty-Verstraete(RPV) and Ross-Thick-LiSparse (RTLS). Our results indicate the RPV model tends to overestimate the bidirectional reflectance for land cover types with high reflectance, while perform well for those with relatively low reflectance in our study area. To test the impact of using multi-view data on image classification, we extracted parameters from BRDF models and used these parameters as object features for object-based classification. The 10-fold cross validation results show that the 3-parameter RTLS significantly improved overall accuracy compared to the classifications relying only on the orthoimage features, while other BRDF models did not show significant improvements, raising the needs to develop new methods to better utilize the multi-view information in GEOIBA in the future.

Georeferencing of mobile ground-based hyperspectral digital single-lens reflex imagery

Abstract. The georeferencing accuracy of a ground-based mobile mapping system designated for agricultural applications is tested. The system integrates a hyperspectral sensor, digital camera, global navigation satellite system receivers, and an inertial navigation system. Acquired imagery was synchronized with GPS time using custom hardware and software solutions developed in-house. The imaging platform was mounted on a forklift and used to conduct three imaging missions along a paved road segment and agricultural beds. Sixteen ground control points were established in each site and used to calibrate the system and test the positional accuracy. The control point coordinates were determined using GNSS and total station observations independent from the imaging data. The navigation data were postprocessed to extract sensor positions and attitude along the imaging trajectories. The pushbroom hyperspectral images were georeferenced using ReSe Parge software, while the digital camera images were analyzed using Agisoft PhotoScan software. Control point coordinates extracted from the georeferenced imagery were compared to corresponding ground-surveyed coordinates. The maximum root mean square errors obtained for the hyperspectral images in all experiments were 2.4 and 3.1 cm in the easting and northing directions, respectively. These results were achieved using only two control points at both ends of the scan line to estimate the boresight offsets. The RMSE values of the orthorectified image constructed using the digital camera images and two control points at each end of the agricultural site were 1.6 and 2.6 cm in the easting and northing directions.

Virtual reality using remote sensing

Virtual reality is becoming increasingly important as a tool to provide cost-effective alternatives for training and to provide enhanced capabilities for activities, such as mission preview, planning and rehearsal. The ability to generate virtual reality utilizing a photo database or remote sensed satellite imagery is particularly of interest. The key to ensure the success of remote sensing-based virtual reality is a system that is able to quickly reconstruct a 3D scene in object space with a realistic appearance. This paper proposes a system to accomplish this task. Main issues of the system include: (1) image registration, (2) feature correspondence and extrusion, and (3) realistic 3D feature rendering. The image registration is achieved by employing a novel method based on the higher-dimension concept. To obtain a high speed, the feature correspondence is implemented using a mathematically well-defined, edge-based method in a multiresolution scheme. Realistic 3D feature rendering creates a photo realistic scene. To further accelerate the processing speed, the system is to be implemented on a parallel computer nCUBE 2 with a Silicon Graphics workstation as a host machine. An example is presented in this paper to demonstrate the capability of the system.

Dual-Antenna Terrestrial Laser Scanner Georeferencing Using Auxiliary Photogrammetric Observations

Terrestrial laser scanning typically requires the use of artificial targets for registration and georeferencing the data. This equipment can be burdensome to transport and set up, representing expense in both time and labor. Environmental factors such as terrain can sometimes make target placement dangerous or impossible, or lead to weak network geometry and therefore degraded product accuracy. The use of additional sensors can help reduce the required number of artificial targets and, in some cases, eliminate the need for them altogether. The research presented here extends methods for direct georeferencing of terrestrial laser scanner data using a dual GNSS antenna apparatus with additional photogrammetric observations from a scanner-mounted camera. Novel combinations of observations and processing methods were tested on data collected at two disparate sites in order to find the best method in terms of processing efficiency and product quality. In addition, a general model for the scanner and auxiliary data is given which can be used for least-squares adjustment and uncertainty estimation in similar systems with varied and diverse configurations. We found that the dual-antenna system resulted in cm-level accuracy practical for many applications and superior to conventional one-antenna systems, and that auxiliary photogrammetric observation significantly increased accuracy of the dual-antenna solution.