No abstract

Video cameras are widely used in underwater robotics to construct 3D coordinates of the workspace. However, the accuracy evaluating that takes into account the properties of the underwater environment, and technical implementation in uncontrolled conditions is still a difficult task. This assessment is especially important for robots that are focused on performing operations with items. In this paper, we propose a novel technique for accuracy analysis and demonstrate its possibilities on real data. It is based on a statistical approach that allows estimating the influence of all sources of perturbations using only experimental data obtained in an underwater environment.

Finite impulse response (FIR) state estimation algorithms have been much discussed in literature lately. It is well known that they allow overcoming the Kalman filter divergence caused by modeling uncertainties. In this paper, new receding horizon unbiased FIR filters ignoring noise statistics for time-varying discrete state-space models are proposed. They have the following advantages. First, the proposed filters use only known means of state vector components at starting points of sliding windows. This allows us to take into account priory statistical information (on average) about specified movements of the system. Second, the iterative version of the filter has a Kalman-like form. Besides, its initialization does not include a training cycle in a batch form. Such filters may have a wide range of applications. In this paper, position and speed estimation of sea targets using angle measurements in azimuth and elevation is considered as an example.

Distortion of underwater images can impair both the accuracy and robustness of 3D scene reconstruction algorithms. The problems that arise are related to the lack of robustness of these methods to changes in the underwater environment and features of transmitting and receiving signals under water, including, in particular, uneven illumination of the underwater environment, rapid attenuation, scattering and refraction of light when passing through an inhomogeneous medium of air-water-glass, limiting the frequency spectrum of passing light, which leads to the absorption of low-frequency components to a greater extent than light of higher frequencies. All this seriously complicates the ability to extract information about the scene as a whole and objects of interest located in the underwater environment, limits the use of standard image processing algorithms and requires their significant improvement. This article offers a new approach to analyzing the accuracy of constructing 3D coordinates of the working space of an underwater robot. The approach is based on underwater camera calibration, assessment of camera image centers taking into account the waterproof shell. We use statistical analysis that allows us to evaluate the impact of all sources of disturbances (both hardware and software) based only on experimental data. In particular, it shows how to get the error distribution using the measured values of the calibration sample and obtained by triangulation under underwater conditions. This makes it possible to simultaneously evaluate the systematic error and the distribution characteristics of the random component of the error in restoring 3D coordinates of the workspace. An important feature of the proposed approach is the ability to assess the impact of all sources of disturbances in the aggregate, including the design of a waterproof shell, based only on experimental data obtained in the underwater environment. In addition, the same approach can also provide estimates of the position of camera image centers, allowing for the presence of a waterproof shell to improve the accuracy of image processing algorithms. The proposed approach was tested on real data.

No abstract