Visual odometry (VO) is a key technology for estimating camera motion from captured images. In this paper, we propose a novel RGB-D visual odometry by constructing and matching features at the superpixel level that represents better adaptability in different environments than state-of-the-art solutions. Superpixels are content-sensitive and perform well in information aggregation. They could thus characterize the complexity of the environment. Firstly, we designed the superpixel-based feature SegPatch and its corresponding 3D representation MapPatch. By using the neighboring information, SegPatch robustly represents its distinctiveness in various environments with different texture densities. Due to the inclusion of depth measurement, the MapPatch constructs the scene structurally. Then, the distance between SegPatches is defined to characterize the regional similarity. We use the graph search method in scale space for searching and matching. As a result, the accuracy and efficiency of matching process are improved. Additionally, we minimize the reprojection error between the matched SegPatches and estimate camera poses through all these correspondences. Our proposed VO is evaluated on the TUM dataset both quantitatively and qualitatively, showing good balance to adapt to the environment under different realistic conditions.

A novel ostraciiform swimming, vision-based autonomous robotic fish is developed in this paper. Its feasibility and capability are shown by implementing a dynamic target following task in a swimming pool. Inspired by boxfish that is highly stable and fairly maneuverable, the robotic fish is designed and constructed by locating multiple propulsors peripherally around a rigid body. Swimming locomotion of the robotic fish is achieved through harmonic oscillations of the tail and pectoral fins. The forces and moments acting on the fins and body are analyzed and the governing motion equations are derived. Through coordinating the movements of the propulsors, several typical swimming patters are empirical designed and realized. A digital camera is integrated in the robotic fish, and the visual information is processed with the embedded microcontroller. To treat the degradation of underwater image, a continuously adaptive mean shift (Camshift) algorithm is modified to keep visual lock on the moving target. A fuzzy logic controller is designed for motion regulation of a hybrid swimming pattern, which employs synchronized pectoral fins for thrust generation and tail fin for steering. A simple target following task is designed via an autonomous robotic fish swimming after a manually controlled robotic fish with fixed distance. The swimming performance of the robotic fish is tested and the effectiveness of the proposed target following method is verified experimentally.

We present a visual feedback method for closed loop control of automated microassembly. A CAD model based multi-camera visual tracking system that is well suited for flexible automation and assembly of complex 3D geometries was developed. The system is capable of providing six degree-of-freedom pose feedback on the observed micro-components in real-time (30 Hz). Using CAD models of the observed objects, a complete description of the observed scene, including the effect of occlusions, is available and dependence on distinctive visual features such as fudicial marks is avoided.

In this paper an algorithm for real-time estimation of the position and orientation of a moving object using a video camera is presented. The algorithm is based on the extended Kalman filter which iteratively computes the object pose from the position measured in the image plane of a set of feature points of the object. A new technique is proposed for the selection of the optimal feature points based on the representation of the object geometry by means of a Binary Space Partitioning (BSP) tree. At each sample time, a visit algorithm of the tree allows pre-selecting all the feature points of the object that are visible from the camera in the pose predicted by the Kalman filter. A further selection is performed to find the optimal set of visible points to be used for image feature extraction. Experimental results are presented which confirm the feasibility and effectiveness of the proposed technique.