Robust aerial infrared target recognition with multi-scale and multi-angle characteristics is a key technique in infrared systems. However, traditional algorithms often fail to achieve a high accuracy and robustness due to simple features and classifiers. Moreover, deep learning algorithms mainly focus on improving accuracy with the price of high complexity. To address above issues, we propose a two-stage lightweight aerial infrared target recognition based on convolutional neural networks(CNN). We propose the coarse region extraction based on the local contrast in the first stage, which combines infrared image characteristics properly. In the second stage, we propose the find target recognition, which constructs lightweight CNN by changing network layers and convolution kernels. Experimental results demonstrate the algorithm proposed can achieve recognition for six kinds of aerial infrared target. Compared with other algorithms, our algorithm obtains higher accuracy and robustness.
The photopolymer materials are good media to record thick hologram gratings, because photopolymer materials have high resolution, low cost, simple process technology and so on. According to coupled wave theory for thick hologram gratings, we know that the same object beam can be reconstructed if the same reference beam is used to retrieve a thick hologram grating. However, the shrinkage always occurs in the photopolymer materials because of environment temperature, humidity, vibration etc. For instance, the same object beam cannot be reconstructed even the same reference beam to be used. In this paper, we will analysis the shrinkage influence of photopolymer materials for thick hologram gratings. We divide the photopolymer materials into several geometry layers, and analysis the reconstructed characteristics separately basing on coupled wave theory of Kogelnik. Through gradually continuous changing the angle between gratings and the border (we call it slant angle), we can build the geometry model of gratings bending caused by shrinkage of materials. We calculate wave complex amplitude diffracted from every layer, and superpose them to compute the total diffraction efficiency. We simulate above methods to obtain the curve of diffraction efficiency with reconstruction wavelength by using Matlab software. Comparing the simulated results with the experiments results, we can deduce the probable situation of thick hologram gratings bending after photopolymer materials shrink.
Waveguide display systems are widely used in various display fields, especially in head mounted display. Comparing
with the traditional head mounted display system, this device dramatically reduce the size and mass. However, there are
still several fatal problems such as high scatting, the cumbersome design and chromatic aberration that should be solved.
We designed and fabricated a monochromatic portable eyewear display system consist of a comfortable eyewear device
and waveguide system with two holographic gratings located on the substrate symmetrically. We record the gratings on
the photopolymer medium with high efficiency and wavelength sensitivity. The light emitting from the micro-display is
diffracted by the grating and trapped in the glass substrate by total internal reflection. The relationship between the
diffraction efficiency and exposure value is studied and analyzed, and we fabricated the gratings with appropriate
diffraction efficiency in a optimization condition. To avoid the disturbance of the stray light, we optimize the waveguide
system numerically and perform the optical experiments. With this system, people can both see through the waveguide to
obtain the information outside and catch the information from the micro display. After considering the human body
engineering and industrial production, we design the structure in a compact and portable way. It has the advantage of
small-type configuration and economic acceptable. It is believe that this kind of planar waveguide system is a potentially
replaceable choice for the portable devices in future mobile communications.