In view of the problem about the loss of detail and color distortion in multi-exposure image fusion, this paper proposed a multi-exposure image fusion method based on tensor decomposition and convolution sparse representation. Tensor decomposition, as an approach of low-rank approximation for high-dimensional data, has great potential in feature extraction of multi-exposure images. Convolution sparse representation is a sparse optimization method for the whole image, which can preserve the detail information of the image to the greatest extent. At the same time, in order to avoid color distortion in the fused image, this paper adopted the method of separately fusing luminance and chrominance. Firstly, the core tensor of the source image was obtained by tensor decomposition. Besides, edge features were extracted on the first sub-band which contains the most information. Then the edge feature map was sparsely decomposed to obtain the activity level of each pixel by using L1 norm of the decomposition coefficient. Finally, take "winner-take-all" strategy to generate weight map so as to obtain the fused luminance components. Unlike the process of luminance fusion, chrominance components were fused by simple Gaussian weighting method, which solves the color distortion problem for the fused image to a certain extent. The experimental results show that the proposed method has great detail preservation ability.

Traditional pinhole spherical wave digital in-line holography has proved to be powerful imaging tools. Image quality is affected by uncertain round of pinhole. Here, we propose a well-distributed sphere wave generation method and it demonstrates wide field of view and high resolution microscopy. The laser focuses into an infinitesimal spot through laser beam expander and microscope objective. Pinhole permutation with different sizes is utilized to match the focal point, and emerges an ideal spherical wave. Interference fringes pattern, formed by reference sphere wave and scattered sphere wave of object, is collected by large area image sensor. The influence of dirty in image sensor and parasitic light is eliminated through subtraction with and without object. Fresnel inverse transformation reconstruction algorithm presents the object information. Biology microscopy experiments demonstrate that the proposed techniques increase the flexibility in producing well-distributed point light source and improve the image quality. Field of view is 3.22 mm×3.22 mm and resolution is 5.09 μm. Furthermore, adjustable field of view with magnification, fast, no-contact make it to be a promising tool in optical element measurement, material identification, biology and medicine.

The evaluation for the matching level between the optical resolution and the pixels number on the full field of view of the capsule endoscope, and the evaluation methods for the resolution effectiveness of pixels number were established. This method selects the horizontal optical axis section as the meridian plane to analyze. The section passing through the planar array sensor is row or column scanned. The line resolution elements number, the line pixel efficiency, the center field matching rate, and the full field maximum matching rate were derived by the analysis unit using the optical resolution angle and the pixel element projection angle. The simple conversion method of resolution length and resolution angle on the spherical field of view can simplify the measurement. These parameters not only constitute the evaluation basis for matching ratio and resolution validity of the pixel number of the capsule endoscope, but also offer a reference for product design, analysis and modification.

Endoscopic image quality plays an important role in the diagnosis of early lesions and dysplasia. Therefore, a blood vessel enhancement algorithm based on spectral absorption characteristics of blood vessels is proposed in this paper. First of all, RGB channels are obtained from the color image and divided into the brightness layer with the high dynamic range and the detail layer with the detail image information through the guided filter. Then, the detail layer of each channel is adaptively enhanced based on SNR (signal noise ratio), and the brightness layer is stretched to improve the GB channel information and to reduce R channel information. Finally each channel is merged to generate an enhanced image. In this article, a large number of endoscopic images is applied to this algorithm and compared with Karl Stroz's Spectra B enhancement technology. This method performs better in image enhancements while using the Detail Variance-Background Variance index and the Weber contrast index to evaluate.

A novel efficient method based on the ultrasound radio frequency (RF) signals is proposed to distinguish the breast tumors grades. First, we utilize the multi-scale geometric characteristic of Shearlet transformation to extract the multi-scale and multi-directional features of ultrasound RF signal, and then reduce the high-dimensional Shearlet features by multi-scale directional binary pattern which can effectively preserve the sufficient discriminated information. At last, we draw on the feature difference between different grades of breast tumors to design a cascade binary tree SVM classifier which not only overcome the problem of sample quantity disequilibrium but also conform to the subjective diagnosis rule of sonographer. Extensive experiments on 928 breast ultrasound RF signals collected from the hospital demonstrate the effectiveness of the new proposed method and its precision, sensitivity, specificity, PPV, NPV and MCC are 89.29%, 75.62%, 94.54%, 97%, 98.3% and 81.01%, respectively.

To realize a simplified keratometer, we proposed a design proposal based on corneal reflex imaging. Six pointolites which arranged in a regular hexagon were used to emit parallel light to the surface of a cornea and were reflected by the cornea. Then the image was captured by a telecentric optical system in the object space to a CMOS camera. In order to obtain the corneal curvature, the distance between two pointolites located on the regular hexagonal diagonal in the corneal reflection images were calculated by using the center of gravity algorithm. The imaging quality, measuring range and measuring precision of the system were theoretically analyzed, and the Ziess’s model eye and human eyes were used to conduct experiments to verify the theoretical analysis results. The experiment results have shown that the precision of the measurement error is ±0.02 mm and the measurement range is from 5.5 mm to 11.6 mm (30 m-1 to 60 m-1 in diopter of cornea). The research will provide technical supports for the design and optimization of automatic keratometer.

In order to reduce the alignment deviation of the imaging keratometer along the optical axis and improve the measurement accuracy of corneal diopter, a high precision imaging keratometer optical system was designed. The optical system includes imaging system and low coherence interferometry system. The imaging system consists of imaging objective, cornea, and measurement target ring, wherein the imaging objective lens adopts a double telocentric optical path design. The low-coherence interferometry system uses the grating scale to measure the displacement of the scanning mirror, and then locates the vertices of the cornea and the measuring target ring by low-coherence interference signals, achieving accurate measurement between the apex of the cornea and the distance of the measuring target ring. The imaging objective has a modulation transfer function greater than 0.4 at a maximum spatial frequency of 70 lp/mm, the distortion is less than 0.05%. The simulation results show that the system has a compact structure, good imaging quality and simple operation. It meets the demand for high precision measurement of corneal refractive power by an imaging keratometer.

In this paper the principle of image rotating mechanism based on prism is introduced and the image rotating mechanism using Pechan prism matching the high speed rotating mirror camera is designed. The designed mechanism can be used in FJZ-250 or SJZ-15 type rotating mirror camera as a fixed part. Equipped with the designed mechanism, the rotating mirror cameras can rotate the image of the object by any angle in the range of 0°~360° before recording it. As a result, the measurement problem of different research directions of detonation test is solved when multiple cameras are used synchronously, which plays an important role in acquisition of experimental data and debugging of outdoors targets, thus, it is a great convenience for the camera. The results of the image quality specification experiment indicate that the equipment of the designed image rotating mechanism based on Pechan prism induced no degradation to the image quality and even slightly improved it.

Aiming at the application of natural objects false target in laser decoy jamming, its deceiving ability characterized by decoy airspace was calculated and analyzed. On the premise of full considering the reflection characteristics of target, the intensity of jamming-indicating laser, atmospheric laser attenuation coefficient, and other factors, the analysis model of decoy jamming ability for false target was established under the condition of suppression coefficient K (≥1). Furthermore, the effectual decoy airspace of common natural objects was analyzed under typical condition. The results show that the effective cheating airspace of false target is closely related to the type of targets, and the deceiving ability of vegetation, gravel and diffuse reflection objects is enhanced in turn under the same conditions. There exists a mirror reflection component that is much smaller than the diffuse reflection component in the decoy airspace of vegetation and gravel natural objects, and the variation of the incident angle of jamming laser has obvious influence on the distribution of effectual decoy airspace. However, there is no obvious trend change rule for the decoy airspace of vegetation targets with the incidence angle. The research results are instructive for the reasonable application of typical natural objects used for false target.

For ship angular flexure measurement based on the ring laser gyro units, a simplified attitude matching method has been proposed, where the Kalman filter observation provides direct measurement of the desired ship angular flexure plus the ‘relative attitude’ term. The ‘relative attitude’, insensitive to the gyro biases of each LGU, arises from the gyro bias difference and initial ship angular flexure. Additionally, considering its slow-varying characteristics, the angular rate of the quasi-static angular flexure should be modeled as random walks. Numerical simulations validate that the simplified attitude matching method can track both the slow-varying angular flexure caused by sunshine heating and the short-time large-magnitude angular flexure caused by factors such as helm’s operation. According to full-scale experiments in several actual ships, the proposed method can reach an accuracy of 20″.