Offline or deferred solutions are frequently employed for high quality and reliable results in current video stabilization. However, neither of these solutions can be used for strict real-time applications. In this paper, we propose a practical and robust algorithm for real-time video stabilization. To achieve this, a novel and efficient motion model based on inter-frame homography estimation is proposed to represent the video motion. An important feature of the proposed motion model is that it updates at each frame input to reduce the accumulation errors caused by parallax or scene changes. We also propose a novel Kalman filter for the motion smoothing and a unique mosaic algorithm for the video completion. The proposed Kalman filter and mosaic algorithm enable the development of a practical real-time video stabilizer that not only produces steady video but also retains the full resolution of the original video. We verify the proposed algorithm through a broad range of video sequences that demonstrate that the proposed algorithm is computationally efficient while being able to robustly stabilize videos with various challenges.

High-precision navigation of the underground excavating machine is the core of realizing safe excavating and autonomous control, but it is difficult to realize high-precision navigation for the underground excavating machine. The integration of GNSS and inertial navigation system are generally adopted for the high-precision navigation of the ground equipment, which can provide high-precision navigation for a long time. However, the GNSS signals cannot be used in underground operation, so the navigation scheme applied to the ground cannot be used underground. At the same time, the velocity of the underground excavation machine is very slow and the operation time is very long, which makes it impossible to use pure inertial navigation scheme for navigation. Because the inertial navigation error accumulates with time, if the operation time is too long, the navigation accuracy cannot be guaranteed. The integration of inertial navigation and odometer is considered as a reliable scheme, but simulation and experimental analysis show that the traditional integration of inertial navigation and odometer cannot meet the operation requirements of the underground excavating machine in the presence of serious slipping and measurement noises. This paper proposes an integration of fully damped inertial navigation/odometer to provide long-term high-precision orientation navigation information for mining equipment. The advantage of the fully damped inertial navigation/odometer integration approach is that it can maintain high accuracy of orientation measurement in one week or longer under the condition of sensor drift error and odometer measurement error. Simulation and field tests are carried out to validate the performance of the proposed approach.

The inertial navigation system (INS) is widely used in spacecraft, missiles, and airplanes for its full autonomy and complete information. Skewed inertial measurement unit (IMU) is a typical device for INS. This IMU is of great significance for key projects such as manned spacecraft or heavy-lift rockets. Calibration of the IMU parameters has always been the key technology for INS. However, the calibration for skewed IMU is difficult to conduct for the skewed placement of the sensors and the complicated model. An all-parameter calibration for six-axis skewed IMU is proposed in this article. The error model is established. The measurement equations for Kalman filter are deduced. The calculation can be accomplished in two steps: The calibration for three gyros+ three accelerometers and the parameter compensation and iteration calculation. Simulations and experiments results indicate that: The complexity of calibration data and procedure is the same level as that of traditional IMU; all the parameters can be precisely estimated with the designed 12 positions sequence.

Video stabilization is a critical step for improving the quality of videos captured by unmanned aerial vehicles. However, the complicated scenarios in the video and the need for instantaneously presenting a stabilized image posed significant challenges to the existing methods. In this work, an instantaneous video stabilization method for unmanned aerial vehicles is proposed. This new approach serves several purposes: smoothing the video motion in both two-dimensional and three-dimensional (3-D) scenes, decreasing the lags in response, and instantaneously providing the stabilized image to users. For each input frame, our approach regenerates four short motion trajectories by applying interframe transformations to the four corners of the image rectangle. An adaptive filter is then performed to smooth motion trajectories and suppress the lags in response simultaneously. Finally, at the stage of image composition, the quality of image is considered for selecting a visually plausible stabilized video. Experiments show that our approach can stabilize various videos without the need for user interaction or costly 3-D reconstruction, and it works as an instant-process for videos from an online source.

A unified model is developed in this paper to characterize multiple stochastic errors in fiber optic gyro (FOG). Differential equations to characterize different stochastic errors in FOG are derived and an equivalent relationship is proved. The unified model of the multiple stochastic errors in FOG is derived by using the differential equations of individual stochastic errors and the equivalent relationship. The derived stochastic error model is used in the initial alignment of a self-designed FOG inertial navigation system (INS). The experimental test results are compared with those of using the normal White-Noise model and Gauss-Markov model, which has shown the better performances of the proposed modeling approach.

Reliable and robust matching of multimodal images is a challenging task because of the nonlinear grayscale distortion caused by radiometric differences and the geometric deformations caused by viewpoint changes between multimodal images. To address these problems, we propose a dense descriptor that uses multiscale structure principal direction to capture the structural features of the image, as well as a novel dissimilarity measurement. The proposed descriptor adapts to nonlinear gray distortion effectively and is robust to image noise. The dissimilarity measurement addresses the reversal of direction caused by intensity inversion. In addition, we propose an improved pixel selection method to speed up the algorithm. Experimental results show that the proposed matching algorithm is offers a superior performance relative to the current state-of-the-art methods.