A CNN-Based High-Accuracy Registration for Remote Sensing Images

In this paper, a convolutional neural network-based registration framework is proposed for remote sensing to improve the registration accuracy between two remote-sensed images acquired from different times and viewpoints. The proposed framework consists of four stages. In the first stage, key-points are extracted from two input images—a reference and a sensed image. Then, a patch is constructed at each key-point. The second stage consists of three processes for patch matching—candidate patch pair list generation, one-to-one matched label selection, and geometric distortion compensation. One-to-one matched patch pairs between two images are found, and the exact matching is found by compensating for geometric distortions in the matched patch pairs. A global geometric affine parameter set is computed using the random sample consensus algorithm (RANSAC) algorithm in the third stage. Finally, a registered image is generated after warping the input sensed image using the affine parameter set. The proposed high-accuracy registration framework is evaluated using the KOMPSAT-3 dataset by comparing the conventional frameworks based on machine learning and deep-learning-based frameworks. The proposed framework obtains the least root mean square error value of 34.922 based on all control points and achieves a 68.4% increase in the matching accuracy compared with the conventional registration framework.

Quasi-geodesics in relativistic gravity

A four-force parallel to the trajectory of a massive particle can always be eliminated by going to an affine parametrization, but the affine parameter is different from the proper time. The main application is to cosmology, in which elements of the cosmic fluid are subject to a pressure gradient parallel to their four-velocities. Natural implementations of parallel four-forces occur when the particle mass changes, in scalar–tensor cosmology, and in cosmic antifriction due to particle production.

Control of a small helicopter with linear matrix inequality-based design assuring stability and performance

This article focuses on linear matrix inequality-based controller designs that can achieve stabilization and reference tracking for a small unmanned helicopter at various flight conditions. A nonlinear mathematical model of a small-scale helicopter is constructed. Then trim conditions are found and linearized around different equilibrium points. Local [Formula: see text] controllers are designed at trim conditions based on the local linear models. The pointwise controllers achieve local stability and performance, but fail at stabilization and tracking over the full envelope. A scheduling controller is built by blending the local controller outputs. In addition, grid-based [Formula: see text] controllers are designed at each operating point with common Lyapunov function. This allows controller scheduling between the adjacent design points with guaranteed stability and performance across the design envelope. Based on the family of linear systems which are obtained from the nonlinear model, an affine parameter-dependent model is built to exploit the approximate linear parameter dependency. Then, a parameter-dependent linear parameter varying controller is synthesized for the affine parameter-dependent model. Although local performance is satisfactory for all given design methods, local [Formula: see text] controllers and affine parameter-dependent controller cannot yield satisfactory performance over the full flight envelope apart from the grid-based controller with common Lyapunov function approach.

Homogeneous geodesics in pseudo-Riemannian nilmanifolds

We study the geodesic orbit property for nilpotent Lie groups N endowed with a pseudo-Riemannian left-invariant metric. We consider this property with respect to different groups acting by isometries. When N acts on itself by left-translations we show that it is a geodesic orbit space if and only if the metric is bi-invariant. Assuming N is 2-step nilpotent and with non-degenerate center we give algebraic conditions on the Lie algebra n of N which imply that every geodesic is the orbit of a one-parameter subgroup of N.Auto(N). In addition we present an example of an almost g.o. space such that for null homogeneous geodesics, the natural parameter of the orbit is not always the affine parameter of the geodesic.

An Affine Parameter Dependent Controller of an Helicopter for Various Forward Velocity Conditions

In this paper, we address the design of a controller that accomplishes stabilization and reference tracking at various flight conditions by using linear helicopter models. The suggested controller is in the form of an H-infinity gain-scheduler, and is used for stabilization and reference tracking for the 4 axis (heave, pitch, roll and yaw) autopilot. Based on the linear models given for the Puma helicopter, an approximate affine parameter dependent model has been built. Then, a linear parameter dependent controller is synthesized which stabilizes the affine parameter dependent helicopter model. By doing so, a single controller achieves stabilization and reference tracking of a family of linear models by scheduling the controller gains based on the online measurement of the scheduling parameter, which is the forward velocity. Moreover, the affine parameter dependent controller is fitted into the linear models. It is observed that this single parameter dependent controller successfully stabilizes a family of linear helicopter model at different forward flight conditions.