Stereo vision is one of the most active research areas in computer vision. While hundreds of stereo reconstruction algorithms have been developed, little work has been done on the evaluation of such algorithms and almost none on evaluation on Near-Infrared (NIR) images. Of almost a hundred examined, we selected a set of 15 stereo algorithms, mostly with real-time performance, which were then categorized and evaluated on several NIR image datasets, including single stereo pair and stream datasets. The accuracy and run time of each algorithm are measured and compared, giving an insight into which categories of algorithms perform best on NIR images and which algorithms may be candidates for real-time applications. Our comparison indicates that adaptive support-weight and belief propagation algorithms have the highest accuracy of all fast methods, but also longer run times (2-3 seconds). On the other hand, faster algorithms (that achieve 30 or more fps on a single thread) usually perform an order of magnitude worse when measuring the per-centage of incorrectly computed pixels.

Vehicular automation is a rapidly advancing field within robotics. These autonomous machines have the potential to perform burdensome and dangerous tasks that historically have been executed by humans which has been a long-time goal for the industry. This thesis aims to develop a computer vision system to enable volume estimation and material classification of the material inside the bucket of an autonomous wheel loader. This information is crucial for autonomous wheel loaders to make decisions. The system is intended to be self-calibrating to ensure future adaptability to different bucket sizes. A Convolutional Neural Network (CNN) based edge detecting network referred to as Dense Extreme Inception Network for Edge Detection (DexiNed) is proposed to both remove redundant information and enhance desired information. By combining the depth perception from a stereo camera and the information extracted from the DexiNed a proposed solution to estimate the volume is presented. A Simple Linear Iterative Clustering (SLIC) approach is applied to extract the material to enable classification of the material. The estimated volume is compared to an annotated true baseline for validation of the system. The thesis presents the precision of the volume estimation and showcases the result of material extraction using three different segment sizes with the SLIC. Additionally, the thesis presents issues concerning material classification.

This thesis presents a novel profile shape matching stereo vision algorithm. This algorithm is able to obtain 3D information in real time from a pair of stereo images. This algorithm produces the 3D information by matching the profile intensity shapes on the same row of the two images from a stereo image pair. The advantage of this profile shape matching algorithm is that the detection of correspondences relies on intensity profile shape not on intensity values, which subject to lighting variations. The user can choose an interval of disparity, and then an object in a desired distance range can be segmented out from the background. In other words, the algorithm detects the object according to its distance to the cameras. Based on the resulting 3D information, the movement and gesture of the control agents, in our test cases the human body and fingers, in space in a desired distance range can be determined. The body movement and gestures can then be analyzed for human-computer interface purposes. In this thesis, the algorithm was applied for human pose and hand gesture estimation. To demonstrate its performance the estimation results were interpreted as inputs and sent to a smart phone to control its functions. While this algorithm does have a trade-off between accuracy and processing speed, we found a balance that can produce the result in real time, and the result has sufficient accuracy for practical use of recognizing human poses and hand gesture. The experimental result shows that the proposed algorithm has higher accuracy and is 1.14× faster than the original version on tested stereo image pairs.

The aim of this project is to examine the feasibility andreliability of the use of a low cost computer vision system totrack and intercept a thrown object. A stereo vision systemtracks the object using color recognition and then guides amobile wheeled robot towards an interception point in orderto capture it. Two different trajectory prediction modelsare compared. One model fits a second degree polynomialto the collected positional measurements of the object andthe other uses the Forward Euler Method to construct theobjects flight path.To accurately guide the robot, the angular position of therobot must also be measured. Two different methods ofmeasuring the angular position are presented and their respectivereliability are measured. A calibrated magnetometeris used as one method while pure computer vision isimplemented as the alternative method.A functional object tracking and interception system thatwas able to intercept the thrown object was constructed usingboth the polynomial fitting trajectory prediction modelas well as the one based on the Forward Euler Method.The magnetometer and pure computer vision are both viablemethods of determining the angular position of therobot with an error of less than 1.5°. / I detta projekt behandlas konstruktionen av och pålitligheteni en bollfånganderobot och dess bakomliggande lågbudgetkamerasystem.För att fungera i tre dimensioner användsen stereokameramodul som spårar bollen med hjälpav färgigenkänning och beräknar bollbanan samt förutspårnedslaget för att ge god tid till roboten att genskjuta bollen.Två olika bollbanemodeller testas, där den ena tar hänsyntill luftmotståndet och nedslaget beräknas numeriskt ochden andra anpassar en andragradspolynom till de observeradedatapunkterna.För att styra roboten till den tänkta uppfångningspunktenbehövs både robotens position, vilket bestäms med kameramodulen,och robotens riktning. Riktningen bestäms medbåde en magnetometer och med kameramodulen, för attundersöka vilken metod som passar bäst.Den förslagna konstruktionen för roboten och kamerasystemetkan spåra och fånga objekt med bådadera de testademodellerna för att beräkna bollbana, dock så är tillförlitligheteni den numeriska metoden betydligt känsligare fördåliga mätvärden. Det är även möjligt att använda sig avbåde magnetometern eller endast kameramodulen för attbestämma robotens riktning då båda ger ett fel under 1.5°.

The thesis deals with incorporating the shape-from-shading technique into the multi-view stereo (MVS) reconstruction framework using the Oren-Nayar reflectance model for rough natural materials. Two methods for enhancing the MVS algorithm with new photo-consistency measure are proposed. Experiments with the laboratory images as well as with images of Mars's surface were conducted, proving that the proposed plane-sweeping method using shading information suitable for combining with MVS can nd the correct position of surface in 3D scene. The experiments also showed, that the Oren-Nayar reflectance model is very accurate for some real-world materials and it can be succesfuly used in the plane-sweeping method to accomplish better results than the Lambert's reflectance model. With precisely estimated material parameters and the light source and camera directions, it is possible to achieve the accuracy of few centimeters in estimating the position of real surface in the scene. / <p>Validerat; 20110825 (anonymous)</p>

The need for accurate 3D reconstructions of complex and large environments or structures has risen dramatically in recent years. In this context, devices known as portable mobile mapping systems have lately emerged as fast and accurate reconstruction solutions. While most of the research and commercial works have relied so far on laser scanners, solutions solely based on cameras and photogrammetry are attracting an increasing interest for the minor costs, size and power consumption of cameras. This thesis presents a novel handheld mobile mapping system based on stereo vision and image-based 3D reconstruction techniques. The main novelty of the system is that it leverages Visual Simultaneous Localization And Mapping (V-SLAM) technology to support and control the acquisition of the images. The real-time estimates of the system trajectory and 3D structure of the scene are used not only to enable a live feedback of the mapped area, but also to optimize the saving of the images, provide geometric and radiometric quality measures of the imagery, and robustly control the acquisition parameters of the cameras. To the best of authors’ knowledge, the proposed system is the first handheld mobile mapping system to offer these features during the acquisition of the images, and the results support its advantages in enabling accurate and controlled visual mapping experiences even in complex and challenging scenarios.

No description available.

This thesis focuses on the use of scaled physical experiments to better understand the development and long-term evolution of fault systems that are otherwise impossible to observe directly. The document is divided into three chapters. The first chapter documents the implementation of an inexpensive stereo vision method for acquiring high resolution three-dimensional strain data for table-top experiments. The second chapter applies the stereo vision method to a tectonic problem—the development of slip partitioning in obliquely loaded crustal systems. Slip partitioned fault systems accommodate oblique convergence with different slip rake on two or more faults and are well documented in the crust. In this chapter, we simulate oblique convergence using blocks with 30° dipping contacts under wet kaolin clay. The experiments reveal three styles of slip partitioning development—contingent upon convergence angle and the presence or absence of a pre-existing vertical fault. Across all experiments, the slip rates along slip-partitioned faults vary temporally suggesting that the faults continuously adjust to conditions produced by the other fault. The lack of steady state in the experiments suggests that slip-partitioned crustal systems may also evolve with oscillating behavior rather than developing a single efficient active fault structure to accommodate oblique convergence. The third chapter documents rheological tests of wet kaolin for applications to crustal deformation experiments. This chapter investigates thixotropy in the clay as well as the role of grain size distribution and water content on its shear strength.

Stereopsis or Stereo vision is a technique that has been extensively used in computer vision these days helps to percept the 3D structure and distance of a scene from two images taken at different viewpoints, precisely the same way a human being visualizes anything using both eyes. The research involves object matching by extracting features from images and includes some preliminary tasks like camera calibration, correspondence and reconstruction of images taken by a stereo vision unit and 3D construction of an object. The main goal of this research work is to estimate the position and the orientation of a front loader bucket of an autonomous mobile robot configured in a work machine name 'Avant', which consists a stereo vision unit and several other sensors and is designed for outdoor operations like excavation. Several image features finding algorithms, including the most prominent two, SIFT and SURF has been considered for the image matching and object recognition. Both algorithms find interest points in an image in different ways which apparently accelerates the feature extraction procedure, but still the time requires for matching in both cases is left as an important issue to be resolved. As the machine requires to do some loading and unloading tasks, dust and other particles could be a major obstacle for recognizing the bucket at workspace, also it has been observed that the hydraulic arm and other equipment comes inside the FOV of the cameras which also makes the task much challenging. The concept of using markers has been considered as a solution to these problems. Moreover, the outdoor environment is very different from indoor environment and object matching is far more challenging due to some factors like light, shadows, environment, etc. that change the features inside a scene very rapidly. Although the work focuses on position and orientation estimation, optimum utilization of stereo vision like environment perception or ground modeling can be an interesting avenue of future research / <p>Validerat; 20101230 (ysko)</p>

A system is described for identifying obstacle-free landing sites for a vertical-takeoff-and-landing (VTOL) semi-autonomous unmanned aerial vehicle (UAV) from point cloud data obtained from a stereo vision system. The relatively inexpensive, commercially available Bumblebee stereo vision camera was selected for this study. A "point cloud viewer" computer program was written to analyze point cloud data obtained from 2D images transmitted from the UAV to a remote ground station. The program divides the point cloud data into segments, identifies the best-fit plane through the data for each segment, and performs an independent analysis on each segment to assess the feasibility of landing in that area. The program also rapidly presents the stereo vision information and analysis to the remote mission supervisor who can make quick, reliable decisions about where to safely land the UAV. The features of the program and the methods used to identify suitable landing sites are presented in this thesis. Also presented are the results of a user study that compares the abilities of humans and computer-supported point cloud analysis in certain aspects of landing site assessment. The study demonstrates that the computer-supported evaluation of potential landing sites provides an immense benefit to the UAV supervisor. / Master of Science