Abstract:Segmented Planar Imaging Detector for Electro-optical Reconnaissance (SPIDER) is a new type of electro-optical imaging technology based on the principle of interferometric imaging. First, it utilizes the optical system with a waveguide structure to coherently sample the object in a spatial frequency domain. Then, it uses an image-processing module to restore the object image through inverse Fourier transform in the frequency domain. Its imaging principle, implementation and the influence of structure parameters on imaging are deeply studied. On this basis, the simulation verification of imaging in the wavelength range from 1.5 μm to 1.9 μm is completed. The results show that the segmented planar imaging detector for electro-optical reconnaissance has a good imaging effect.

Abstract:As a new hot spot in the field of polarization detection technology in recent years, the focal plane segmented polarization detector has obvious effect both in the enhancement of the contrast between target and background and in the recognition of camouflaged targets. On the basis of a linear near-infrared micro-polarization detector, an imaging system which uses a Jetson TX1 as the core is designed. The imaging system mainly includes a control module and a processing module. The control module which consists of a detector driver board and a FPGA control board is used to implement the AD conversion of detector signals and the usb transmission of image signals. The processing module which uses the Jetson TX1 platform is used to implement the usb reception of image signals, the adjustment of integral time, the correction of non-uniformity, the calculation of polarization degree and the parallel algorithm by GPU. Infrared and polarization imaging experiment is carried out on target scenes and the infrared and infrared polarization images with the resolution of 500×268 pixels are obtained.

Abstract:A high-speed infrared video acquisition system is proposed. The design and implementation of a PCI-E acquisition board which is the key module of the acquisition system is mainly presented. The acquisition board is mainly based on the hardware structure of Kintex7 Field-Programmable Gate Array (FPFA). In the acquisition system, the high-speed fiber transmission and PCI-E technology are used successfully by utilizing Aurora core and Xilinx 7 series PCI-E IP core. The experimental results show that this high-speed infrared video acquisition system can operate stably and reliably for a long time. It is very suitable for the real-time acquisition, display and storage of high frame rate and large array infrared video.

Abstract:A design scheme of the calibration system for a neutral density filter type near-infrared multichannel variable optical attenuator is proposed. To verify the scheme, a calibration system is built. The results show that compared with the traditional calibration scheme, this scheme not only can shorten the calibration time for the multi-channel variable optical attenuator, but also can improve the calibration accuracy of the instrument.

Abstract:A high sensitivity near infrared fiber-optical power meter is designed on the basis of InGaAs photodiodes. Because the preamplifier circuit of this optical power meter is designed with a multilevel amplification range and a shielding box is added in, the measurement dynamic range is improved. In the power meter, the system common-mode noise is suppressed by a differential amplifier and a differential A/D conversion chip; the ambient temperature is measured by a temperature sensor; and the preamplifier circuit is adjusted to zero automatically by a D/A conversion chip. Thus, the stray light noise of the system, the dark current of the photoelectric detector and the temperature drift of the photoelectric detector are suppressed. The test results show that the light power measurement sensitivity of the fiber-optical power meter is up to -90 dBm in the fiber-optical communication band. Its measurement dynamic range, system signal-to-noise ratio and linearity are improved notably.

Abstract:The use of analog 4-binning mode in Mars mineral spectrum analyzers enables the instruments to deal with the complex environment in deep space and hence to meet the requirements of high signal-to-noise ratio and large dynamic range. A new on-chip 2×2 pixel binning driving method is proposed and the rationality of its output is verified. For an e2v CCD 47-20 chip, the CCD structure is studied and the 2-binning methods both in horizontal direction and in vertical direction are designed so that the driving wave can be controlled by FPGA. On the basis of the structure of a CCD on-chip amplifier, a four-stage model is obtained by simulating the 2×1-horizontal output analog signal model. The result obtained in the uniform light illumination experiment shows that the CCD output signal in the 4-binning mode is the same as the simulation result. This confirms the CCD amplification theory. The output voltage presents a linear trend when the value is than 600 mV, which proves that the combined results of the scheme are accurate and reliable.

Abstract:A semantic segmentation network based on deep learning is proposed. The network designs a spatial pyramid module which can extract multi-scale information from images through Atrous convolution. It also explores the influence of Atrous convolution sampling rate and multi-scale branches on the performance of network through extensive experiment. the impact of hyperparameters on network performance during training is discussed. The test results on the SUN RGB-D dataset show that compared with other state-of-the-art semantic segmentation networks, the performance of the network we proposed is outstanding. Finally, the semantic segmentation based on infrared images is explored preliminarily.

Abstract:Rotary Fourier transform spectrometers have many advantages such as high reliability, high spectral resolution and fast test speed etc. But when it scans optical path difference by uniform rotation, there is a nonlinear problem for the optical path. The nonlinear optical path difference scanning of a transmission Fourier spectrometer is analyzed. A mathematical model of nonlinear optical path difference is proposed. The optical path difference equation is established and the MATLAB simulation is carried out. When a rotating mirror is controlled to rotate at a constant speed by a speed feedback motor, the inhomogeneity of time can be obtained according to the established nonlinear equation. So the non-uniform sampling of time can be realized and accurate spectra can be obtained. Instead of using a stabilized He-Ne laser to trigger sampling, it can save cost and reduce the volume and weight of the spectrometer.

Abstract:Rotary kilns are important calcining equipment in cement clinker production. In the production process of cement clinker, the surface temperature of a rotary kiln should be measured so as to monitor the internal production state of the kiln. Traditional methods usually use single-band infrared temperature measurement technology to monitor the surface temperature of rotary kilns. However, their temperature measurement accuracy are not high enough to obtain accurate temperature measurement data. To improve the measurement accuracy of surface temperature of a rotary kiln, a dual-band infrared temperature measurement method is proposed. First, the traditional rotary kiln temperature measurement method and several main factors which affect temperature measurement are analyzed theoretically. Then, the principle of dual-band temperature measurement is presented and a dual-band temperature measurement model based on the rotary kiln is deduced. Finally, the model is verified experimentally. The experimental result shows that this dual-band infrared temperature measurement method has higher temperature measurement accuracy compared with the traditional single-band temperature measurement method in temperature measurement.