The imaging quality of the aerodynamically heated optical dome was evaluated under the comprehensive influence of aero-optical transmission effect and aero-thermal radiation effect. The ray propagating algorithm based on the fourth order Runge-Kutta method was used to trace the target ray and the thermal radiation ray of the optical dome. Three imaging quality evaluation parameters were proposed to evaluate aero-optical effect: Modulation transfer function (MTF), irradiance, peak signal-to-noise ratio (PSNR) of distorted images. The results show that: as the flight speed increased, the MTF decreased observably compared with the diffraction-limit MTF, the irradiance on the photosensitive surface of the detector increased gradually, and the distorted imaging quality under the influence of the comprehensive aero-optical effect gradually deteriorated. However, as the thickness of the optical dome increased, the MTF decreased sharply and the irradiance decreased gradually, that indicated the aero-optical transmission effect was reinforced and the aero-thermal radiation effect was weakened. The imaging quality improved with thickness increasing. The influence of aero-thermal radiation effect on the PSNR of the image was more serious than that of the aero-optical transmission effect.

Carbon-nanotube (CNT) is a promising material owing to its compelling mechanical, thermal and electrical properties and has been applied in a broad variety of fields such as composite, fiber, film and microelectronic. Although the introductions of CNT have brought huge improvement for many applications, these properties of macrostructures prepared by CNTs still cannot meet those of individual CNT. Disordered alignment of CNTs in the matrix results in degradation of performance and hinders further application. Nowadays, quantities of methods are being researched to realize alignments of CNTs. In this paper, we introduce the application of CNTs and review some typical pathways for vertical and horizontal alignment, including chemical vapor disposition, vertical self-assembly, external force, film assisted, electric field, magnetic field and printing. Besides that, advantages and disadvantages of specific methods are also discussed. We believe that these efforts will contribute to further understanding the nature of aligned CNT and generating more effective ideas to the relevant workers.

Traditional optical domes are spherical, which introduces constant aberrations with look angle. However, spherical domes are not optimum for reducing aerodynamic drag. Conformal domes deviate from spherical to reduce drag but they generate dynamic aberrations varying significantly with look angle in the field of regard. Thus, conformal domes require unique challenges for aberration correction. This paper presents a method to reduce the dynamic aberrations through designing the inner surface of conformal domes. This method follows the principle that the optical axis ray of the imaging system maintains the same direction after refraction through the conformal dome for different look angles. Based on this principle, equations that the inner surface should satisfy are established and a numerical solution method is introduced. Eventually, Zernike polynomial coefficients of Z4, Z5, Z8, and Z9, which represent defocus, astigmatism, coma, and spherical aberration, respectively, are analyzed for quadratic domes with different inner surfaces. Compared with domes with traditional inner surfaces, quadratic domes with the inner surfaces calculated by this method have smaller Zernike aberrations. In conclusion, this design method for the inner surface can effectively reduce dynamic aberrations.

Accurately sensing the surrounding 3D scene is indispensable for drones or robots to execute path planning and navigation. In this paper, a novel monocular depth estimation method was proposed that primarily utilizes a lighter-weight Convolutional Neural Network (CNN) structure for coarse depth prediction and then refines the coarse depth images by combining surface normal guidance. Specifically, the coarse depth prediction network is designed as pre-trained encoder–decoder architecture for describing the 3D structure. When it comes to surface normal estimation, the deep learning network was designed as a two-stream encoder–decoder structure, which hierarchically merges red-green-blue-depth (RGB-D) images for capturing more accurate geometric boundaries. Relying on fewer network parameters and simpler learning structure, better detailed depth maps are produced than the existing states. Moreover, 3D point cloud maps reconstructed from depth prediction images confirm that our framework can be conveniently adopted as components of a monocular simultaneous localization and mapping (SLAM) paradigm.

Solving the problem that features prediction of self-supervised monocular depth estimation is still ambiguous in low-texture regions and boundaries. We proposed an innovative self-supervised monocular depth estimation method, monocular depth self-supervised network, which integrates three effective strategies to construct an innovative self-supervised monocular depth estimation framework: (1) the attention mechanism and feature fusion module are adopted to enhance the semantic and spatial information of feature images, (2) the threshold segmentation mask is utilized to solve object motion and low-texture regions to increase image details, and (3) the residual pose module and deep reconstruction loss are used to enhance the feature extraction capability of the model to improve the accuracy of depth and pose estimation. Comprehensive experiments and visual analysis results demonstrate the effectiveness of each component in isolation. Compared to existing self-supervised methods, our model not only achieves outstanding results on KITTI and NYU Depth V2 datasets but also can be suitable to different environments.