Abstract:An passive infrared or millimeter wave imaging system based on the theory of compressed sensing is designed. Compressed sensing is a technology using sparse or compressed priority information to acquire and reconstruct signals. It can sample signals at a rate much lower than the Nyquist sampling rate and implement high precision reconstruction of the signals. The above designed system has overcome the shortcomings of the traditional systems and can conduct its imaging according to the theory of compressed sensing. The simulation result shows that the system has good imaging performance, its peak signal to noise ratio and resolution are improved and the data required to be stored are reduced greatly.

Abstract:Compared with a cooled infrared detector, an uncooled infrared detector has the advangtages of small size, low weight and low power consumption because it has not any cooler. To improve the encapsulation process of an uncooled infrared detector after exhaustion and achieve its configuration optimization and reliability enhancement, a technical scheme to use laser welding to encapsulate the uncooled infrared detector in high vacuum is proposed. The test result shows that the uncooled infrared detector assembly has not apparent variation in performance and can meet the requirements for actual application.

Abstract:To investigate the relation between the optical property and the crystal size of a nano structure, a plane wave approximate method based on the density functional theory of first principle is used to calculate the different nanocrystal models in SiO$_2$. The band structure, state density and optical property of the nano structure embedded in the SiO2 are analyzed. The result shows that for Ge and Si structures, their intermediate levels are introduced at about 3.3eV and 4.3eV above the Fermi level respectively. With the increase of the crystal size, the light absorption edge firstly moves to the red side and then moves to the blue side. It is considered that the visible light emission in the Ge and Si nanocrystal structures comes from the defects at their interfaces. In the Ge and Si nanocrystals embedded in the SiO2, there exist minimum size limits suitable for quantum-confined models. These conclusions give us a certain props both for improvement of the optical property of materials and for further study of the luminescence mechanism of nano structures.

Abstract:Since the traditional neural network non-uniformity correction algorithm has lower convergence speed and lower correction precision, it is more difficult to obtain satisfactory correction result when high background noise is present. For this reason, a neural network method for correcting the non-uniformity of an infrared focal plane array (IRFPA) based on median filtering is proposed. In the method, the median filtering is first used to preprocess the high noise and then an improved neural network algorithm is used to correct the non-uniformity of the IRFPA adaptively. The experimental result shows that the algorithm has higher convergence speed and higher correction precision than the traditional method and can make the peak signal to noise ratio be improved by 10dB at least.

Abstract:Using spectral analysis techniques, this research aims to estimate kalium content of apple flowers. Firstly, the original spectral reflectance (Ri) of apple fresh flowers at full-bloom stage of Red Fuji and flower kalium content (KC) were measured, and correlations between kalium content and Ri and eight kinds of spectral transformation data (Ri/、Ri//、1/Ri、(1/Ri))/、lg(1/Ri)、(lg(1/Ri))/、lgRi、(lgRi)/) were analyzed. Sectral reflectance variant with the highest correlation coefficient was discovered. Secondly, stepwise regression analysis of kalium content and the corresponding spectral reflectance variant with higher correlation coefficients was carried out to determine sensitive characteristic wavelength. Estimation models for kalium content of apple flowers were established with sensitive wavelength. The best estimation model was determined by model optimization and validation. The results showed that the highest correlation coefficient appeared between kalium content of apple flower and the first derivative of spectral reflectance Ri/ and the second highest relevance was for 1/Ri, lg (1/Ri) and lgRi. Their absolute correlation coefficient peaks (valleys) appeared at wavelength of 669 nm, 952 nm, 1164 nm, 1442 nm, wavelength of 351 nm, 352 nm, 362 nm, 366 nm, wavelength of 351 nm, 366 nm, and wavelength of 351 nm, 366 nm. The sensitive wavelength were 669 nm, 1442 nm, 352nm, 351nm, 351nm, which were defined by stepwise regression analysis. The established regression models had better linear trend, and the estimation model with the first derivative of spectral reflectance variable had the maximal correlation coefficient up to 0.6113. The estimation model was tested by the test data samples, with its fitting equation coefficients of determination (R2) of 0.6955, root mean square error (RMSE) of 2.7, and the relative error (RE) of 4.9%. It was indicated that the estimation model of kalium content of apple flower with the first derivative of spectral reflectance had the highest accuracy, which was the best estimation model with a certain degree of stability and applicability. Our results provided theoretical basis and reference value for the rapid estimation of apple blossom nutrient content and real-time nutritional diagnosis.

Abstract:Atmospheric turbulance is one of the main factors which have influence on the performance of a laser communication system. When other noises in a laser communication system are ignored and only the bit error rate (BER) resulted from the atmospheric turbulance is considered, the relation among BER, signal-to-noise ratio (SNR) and log-amplitude variance of a laser communication system when its laser beam is propagating in a slant atmospheric turbulence path is studied. According to the ITU-R turbulance structure constant model, the SNR and BER varied with the altitude at different wavelengths and different zenith angles are simulated numerically with FORTRAN. The result shows that for the same altitude, if the long wavelength laser beam is propagating slantly at a small zenith angle, the BER of the laser communication system can be reduced, the SNR can be increased and the communication quality can be improved.

Abstract:In order to segment an infrared image with a complex harbor background effectively, rapidly and accurately, an infrared image segmentation method based on fractal theory is proposed. Firstly, the segmentation method uses the Otsu to segment the raw infrared image coarsely on the basis of the fractal processing. Then, it uses the Garbor transform and Garbor multichannel filters to extract the harbor contour in the image. Finally, it uses the median filter to filter the noise and hence obtain the true harbor contour image. The experimental result shows that the proposed method is feasible and effective.

Abstract:Because the angle of field of view of an infrared thermal imager is limited and higher resolution is required for the obtained thermal wave image, the thermal wave nondestructive detection of a large object can be implemented only when several images are jointed. For the joint of images, the premise is that the images should be registered precisely. According to the concept of template matching, the Genetic Algorithm (GA) is incorporated into the infrared thermal wave image registration. Through the modification of selection strategy, an adaptive crossover and mutation probability is designed to realize precise registration of thermal images. The experimental result shows that the method can realize the precise registration of thermal images and its operation speed and stability are improved.

Abstract:A method for calculating the ratio of the radiation rays falling on a detector chip from the inner wall of a circular cold shield by using the Monte Carlo method and MATLAB is presented. It is based on the directional number concept in three-dimensional space. First, a random point on the surface of the inner wall in the cold shield is generated with MATLAB. Then, a group of random directional numbers is associated to that point to form an equation for a random straight line and wether it is projected to the chip plane directly or to the chip plane after reflection is judged. In the case of reflection, the radiation rays reflected from the inner wall of the cold shield are generated according to the property of specular reflection and wether they fall on the chip region is judged. After statistical analysis, the ratio of the number of the radiation rays falling on the detector chip to the total number of the radiation rays emitted from the surface of the inner wall of the cold shield can be calculated. This method has the features of easy programming and simple operation.