Robust sequential view planning for object recognition using multiple cameras

Abstract

In this thesis the problem of object recognition/pose estimation using active sensing is investigated. It is assumed that multiple cameras acquire images from different view angles of an object belonging to a set of a priori known objects. The eigenspace method is used to process the sensory observations and produce an abstract measurement vector. This step is necessary to avoid the manipulation of the original sensor data, i.e. large images, that can render the sensor modelling and matching process practically infeasible.The eigenspace representation is known to have shortcomings in dealing with structured noise such as occlusion. To overcome this problem, models of occlusions and sensor noise have been incorporated into the probabilistic model of sensor/object to increase robustness with respect to such uncertainties. The active recognition algorithm has also been modified to consider the possibility of occlusion, as well as variation in the occlusion levels due to camera movements.A recursive Bayesian state estimation problem is formulated to model the observation uncertainties through a probabilistic scheme. This enables us to identify the object and estimate its pose by fusing the information obtained from individual cameras. To this end, an extensive training step is performed, providing the system with the sensor model required for the Bayesian estimation. In order to enhance the quality A multi-camera active vision system has been developed in order to implement the ideas proposed in this thesis. Comparative Monte Carlo experiments conducted with the two-camera system demonstrate the effectiveness of the proposed methods in object classification/pose estimation in the presence of structured noise. Different concepts introduced in this work, i.e., the multi-camera data fusion, the occlusion modelling, and the active camera movement, all improve the recognition process significantly. Specifically, these approaches all increase the recognition rate, decrease the number of steps taken before recognition is completed, and enhance robustness with respect to partial occlusion considerably.iv

An optimal path planning strategy for multiple target search by a mobile robot

“…Figures 15(a)–(c) are the maps of Environment I, Environment II, and Environment III, respectively. Many approaches have been developed to recognize and track targets 26,27. In this paper, the CamShift algorithm [28], which uses histogram to track targets with a known hue, is employed.…”

Stereo‐imaging network design for precise and dense 3d reconstruction

“…View planning has been used for a variety of systems based on both active and passive sensors (Farshidi et al., ). While many researchers have worked on VP of active range‐finding technology or triangulation‐based active sensors (Scott et al., ; Scott, ), and several research projects have been carried out with networks of passive imaging sensors (Mason, ; Saadatseresht et al., ; Olague and Dunn, ; Saadatseresht and Varshosaz, ; Saadatseresht and Samadzadegan, ; Dunn and Frahm, ; Alsadik et al., ), there are few methods presented in the literature of sensor placement of passive stereo vision range‐finding technology (Chen and Li, ).…”

Unposed Object Recognition using an Active Approach

“…The strategy of improving object recognition through multiple viewpoints is referred to as active object recognition (D. Wilkes, 1992). Several ((Denzler and Brown, 2002;Farshidi et al, 2009;LaPorte and Arbel, 2006)) have proposed probabilistic frameworks for active object recognition. These frameworks serve to both incorporate multiple viewpoints as well as incorporating prior probability.…”