Omnidirectional bearing-only see-and-avoid for small aerial robots

Motivated by the growing interest in unmanned aerial system’s applications in indoor and outdoor settings and the standardisation of visual sensors as vehicle payload. This work presents a collision avoidance approach based on omnidirectional cameras that does not require the estimation of range between two platforms to resolve a collision encounter. It will achieve a minimum separation between the two vehicles involved by maximising the view-angle given by the omnidirectional sensor. Only visual information is used to achieve avoidance under a bearing-only visual servoing approach. We provide theoretical problem formulation, as well as results from real flight using small quadrotors.

Practical improvements to simultaneous computation of multi-view geometry and radial lens distortion

This paper discusses practical issues related to the use of the division model for lens distortion in multi-view geometry computation. A data normalisation strategy is presented, which has been absent from previous discussions on the topic. The convergence properties of the Rectangular Quadric Eigenvalue Problem solution for computing division model distortion are examined. It is shown that the existing method can require more than 1000 iterations when dealing with severe distortion. A method is presented for accelerating convergence to less than 10 iterations for any amount of distortion. The new method is shown to produce equivalent or better results than the existing method with up to two orders of magnitude reduction in iterations. Through detailed simulation it is found that the number of data points used to compute geometry and lens distortion has a strong influence on convergence speed and solution accuracy. It is recommended that more than the minimal number of data points be used when computing geometry using a robust estimator such as RANSAC. Adding two to four extra samples improves the convergence rate and accuracy sufficiently to compensate for the increased number of samples required by the RANSAC process.

Negative determinant of Hessian features

Local image feature extractors that select local maxima of the determinant of Hessian function have been shown to perform well and are widely used. This paper introduces the negative local minima of the determinant of Hessian function for local feature extraction. The properties and scale-space behaviour of these features are examined and found to be desirable for feature extraction. It is shown how this new feature type can be implemented along with the existing local maxima approach at negligible extra processing cost. Applications to affine covariant feature extraction and sub-pixel precise corner extraction are demonstrated. Experimental results indicate that the new corner detector is more robust to image blur and noise than existing methods. It is also accurate for a broader range of corner geometries. An affine covariant feature extractor is implemented by combining the minima of the determinant of Hessian with existing scale and shape adaptation methods. This extractor can be implemented along side the existing Hessian maxima extractor simply by finding both minima and maxima during the initial extraction stage. The minima features increase the number of correspondences by two to four fold. The additional minima features are very distinct from the maxima features in descriptor space and do not make the matching process more ambiguous.

Scene invariant crowd counting and crowd occupancy analysis

In public places, crowd size may be an indicator of congestion, delay, instability, or of abnormal events, such as a fight, riot or emergency. Crowd related information can also provide important business intelligence such as the distribution of people throughout spaces, throughput rates, and local densities. A major drawback of many crowd counting approaches is their reliance on large numbers of holistic features, training data requirements of hundreds or thousands of frames per camera, and that each camera must be trained separately. This makes deployment in large multi-camera environments such as shopping centres very costly and difficult. In this chapter, we present a novel scene-invariant crowd counting algorithm that uses local features to monitor crowd size. The use of local features allows the proposed algorithm to calculate local occupancy statistics, scale to conditions which are unseen in the training data, and be trained on significantly less data. Scene invariance is achieved through the use of camera calibration, allowing the system to be trained on one or more viewpoints and then deployed on any number of new cameras for testing without further training. A pre-trained system could then be used as a ‘turn-key’ solution for crowd counting across a wide range of environments, eliminating many of the costly barriers to deployment which currently exist.

Multi-sensor data fusion for UAV navigation during landing operations

This paper presents a practical framework to synthesize multi-sensor navigation information for localization of a rotary-wing unmanned aerial vehicle (RUAV) and estimation of unknown ship positions when the RUAV approaches the landing deck. The estimation performance of the visual tracking sensor can also be improved through integrated navigation. Three different sensors (inertial navigation, Global Positioning System, and visual tracking sensor) are utilized complementarily to perform the navigation tasks for the purpose of an automatic landing. An extended Kalman filter (EKF) is developed to fuse data from various navigation sensors to provide the reliable navigation information. The performance of the fusion algorithm has been evaluated using real ship motion data. Simulation results suggest that the proposed method can be used to construct a practical navigation system for a UAV-ship landing system.

A suspicious behaviour detection using a context space model for smart surveillance systems

Video surveillance systems using Closed Circuit Television (CCTV) cameras, is one of the fastest growing areas in the field of security technologies. However, the existing video surveillance systems are still not at a stage where they can be used for crime prevention. The systems rely heavily on human observers and are therefore limited by factors such as fatigue and monitoring capabilities over long periods of time. This work attempts to address these problems by proposing an automatic suspicious behaviour detection which utilises contextual information. The utilisation of contextual information is done via three main components: a context space model, a data stream clustering algorithm, and an inference algorithm. The utilisation of contextual information is still limited in the domain of suspicious behaviour detection. Furthermore, it is nearly impossible to correctly understand human behaviour without considering the context where it is observed. This work presents experiments using video feeds taken from CAVIAR dataset and a camera mounted on one of the buildings Z-Block) at the Queensland University of Technology, Australia. From these experiments, it is shown that by exploiting contextual information, the proposed system is able to make more accurate detections, especially of those behaviours which are only suspicious in some contexts while being normal in the others. Moreover, this information gives critical feedback to the system designers to refine the system.

Towards automatic object segmentation with sequential multiple views

Object segmentation is one of the fundamental steps for a number of robotic applications such as manipulation, object detection, and obstacle avoidance. This paper proposes a visual method for incorporating colour and depth information from sequential multiview stereo images to segment objects of interest from complex and cluttered environments. Rather than segmenting objects using information from a single frame in the sequence, we incorporate information from neighbouring views to increase the reliability of the information and improve the overall segmentation result. Specifically, dense depth information of a scene is computed using multiple view stereo. Depths from neighbouring views are reprojected into the reference frame to be segmented compensating for imperfect depth computations for individual frames. The multiple depth layers are then combined with color information from the reference frame to create a Markov random field to model the segmentation problem. Finally, graphcut optimisation is employed to infer pixels belonging to the object to be segmented. The segmentation accuracy is evaluated over images from an outdoor video sequence demonstrating the viability for automatic object segmentation for mobile robots using monocular cameras as a primary sensor.

Loop closure detection on a suburban road network using a continuous appearance-based trajectory

This paper presents a novel technique for performing SLAM along a continuous trajectory of appearance. Derived from components of FastSLAM and FAB-MAP, the new system dubbed Continuous Appearance-based Trajectory SLAM (CAT-SLAM) augments appearancebased place recognition with particle-filter based ‘pose filtering’ within a probabilistic framework, without calculating global feature geometry or performing 3D map construction. For loop closure detection CAT-SLAM updates in constant time regardless of map size. We evaluate the effectiveness of CAT-SLAM on a 16km outdoor road network and determine its loop closure performance relative to FAB-MAP. CAT-SLAM recognizes 3 times the number of loop closures for the case where no false positives occur, demonstrating its potential use for robust loop closure detection in large environments.

Improving recall in appearance-based visual SLAM using visual expectation

In this paper, we present a new algorithm for boosting visual template recall performance through a process of visual expectation. Visual expectation dynamically modifies the recognition thresholds of learnt visual templates based on recently matched templates, improving the recall of sequences of familiar places while keeping precision high, without any feedback from a mapping backend. We demonstrate the performance benefits of visual expectation using two 17 kilometer datasets gathered in an outdoor environment at two times separated by three weeks. The visual expectation algorithm provides up to a 100% improvement in recall. We also combine the visual expectation algorithm with the RatSLAM SLAM system and show how the algorithm enables successful mapping

An exploration of feature detector performance in the thermal-infrared modality

Thermal-infrared images have superior statistical properties compared with visible-spectrum images in many low-light or no-light scenarios. However, a detailed understanding of feature detector performance in the thermal modality lags behind that of the visible modality. To address this, the first comprehensive study on feature detector performance on thermal-infrared images is conducted. A dataset is presented which explores a total of ten different environments with a range of statistical properties. An investigation is conducted into the effects of several digital and physical image transformations on detector repeatability in these environments. The effect of non-uniformity noise, unique to the thermal modality, is analyzed. The accumulation of sensor non-uniformities beyond the minimum possible level was found to have only a small negative effect. A limiting of feature counts was found to improve the repeatability performance of several detectors. Most other image transformations had predictable effects on feature stability. The best-performing detector varied considerably depending on the nature of the scene and the test.

Face recognition across pose on video using eigen light-fields

We propose an approach to employ eigen light-fields for face recognition across pose on video. Faces of a subject are collected from video frames and combined based on the pose to obtain a set of probe light-fields. These probe data are then projected to the principal subspace of the eigen light-fields within which the classification takes place. We modify the original light-field projection and found that it is more robust in the proposed system. Evaluation on VidTIMIT dataset has demonstrated that the eigen light-fields method is able to take advantage of multiple observations contained in the video.

A semi-local method for iterative depth-map refinement

We present an iterative hierarchical algorithm for multi-view stereo. The algorithm attempts to utilise as much contextual information as is available to compute highly accurate and robust depth maps. There are three novel aspects to the approach: 1) firstly we incrementally improve the depth fidelity as the algorithm progresses through the image pyramid; 2) secondly we show how to incorporate visual hull information (when available) to constrain depth searches; and 3) we show how to simultaneously enforce the consistency of the depth-map by continual comparison with neighbouring depth-maps. We show that this approach produces highly accurate depth-maps and, since it is essentially a local method, is both extremely fast and simple to implement.

Feature-domain super-resolution framework for gabor-based face and iris recognition

The low resolution of images has been one of the major limitations in recognising humans from a distance using their biometric traits, such as face and iris. Superresolution has been employed to improve the resolution and the recognition performance simultaneously, however the majority of techniques employed operate in the pixel domain, such that the biometric feature vectors are extracted from a super-resolved input image. Feature-domain superresolution has been proposed for face and iris, and is shown to further improve recognition performance by capitalising on direct super-resolving the features which are used for recognition. However, current feature-domain superresolution approaches are limited to simple linear features such as Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), which are not the most discriminant features for biometrics. Gabor-based features have been shown to be one of the most discriminant features for biometrics including face and iris. This paper proposes a framework to conduct super-resolution in the non-linear Gabor feature domain to further improve the recognition performance of biometric systems. Experiments have confirmed the validity of the proposed approach, demonstrating superior performance to existing linear approaches for both face and iris biometrics.

A mask-based approach for the geometric calibration of thermal-infrared cameras

Accurate and efficient thermal-infrared (IR) camera calibration is important for advancing computer vision research within the thermal modality. This paper presents an approach for geometrically calibrating individual and multiple cameras in both the thermal and visible modalities. The proposed technique can be used to correct for lens distortion and to simultaneously reference both visible and thermal-IR cameras to a single coordinate frame. The most popular existing approach for the geometric calibration of thermal cameras uses a printed chessboard heated by a flood lamp and is comparatively inaccurate and difficult to execute. Additionally, software toolkits provided for calibration either are unsuitable for this task or require substantial manual intervention. A new geometric mask with high thermal contrast and not requiring a flood lamp is presented as an alternative calibration pattern. Calibration points on the pattern are then accurately located using a clustering-based algorithm which utilizes the maximally stable extremal region detector. This algorithm is integrated into an automatic end-to-end system for calibrating single or multiple cameras. The evaluation shows that using the proposed mask achieves a mean reprojection error up to 78% lower than that using a heated chessboard. The effectiveness of the approach is further demonstrated by using it to calibrate two multiple-camera multiple-modality setups. Source code and binaries for the developed software are provided on the project Web site.

Improved facial expression recognition via uni-hyperplane classification

Large margin learning approaches, such as support vector machines (SVM), have been successfully applied to numerous classification tasks, especially for automatic facial expression recognition. The risk of such approaches however, is their sensitivity to large margin losses due to the influence from noisy training examples and outliers which is a common problem in the area of affective computing (i.e., manual coding at the frame level is tedious so coarse labels are normally assigned). In this paper, we leverage the relaxation of the parallel-hyperplanes constraint and propose the use of modified correlation filters (MCF). The MCF is similar in spirit to SVMs and correlation filters, but with the key difference of optimizing only a single hyperplane. We demonstrate the superiority of MCF over current techniques on a battery of experiments.