Turbine blades, as one of the core components of aero-engines, operate in high-temperature and high-pressure environments for a long time. Radiation thermometry has been widely applied in temperature monitoring of turbine blades in high-temperature environments due to its advantages of not affecting the temperature field distribution of the blades, high sensitivity and fast response speed. To address the limitations of the existing reflection radiation error correction methods in radiation thermometry, a multiple-reflection radiation error correction method for radiation thermometry of aero-engine turbine blades is proposed. Combining the two-color pyrometry techniques, a multiple-reflection radiation model is established that simultaneously accounts for reflections from the front stage guide vane, adjacent rotating blades and the combustion chamber walls. Based on the model, the simulation analysis and experiment of reflection radiation error correction are performed. Experimental results demonstrate that after correcting the reflection radiation error within the measured temperature range, the temperature measurement error is reduced from 2.3%~3.3% to ±0.78%, effectively achieving the correction of reflection radiation error.

With the continuous improvement and development of optical communication technology, equipment manufacturers are demanding increasingly stringent accuracy requirements for the performance indicators of high-speed optical communication devices. Scattering parameters (S parameters) can accurately characterize the transmission characteristics of optical communication devices. In traditional separate and integrated measurement instruments, random errors generated during individual measurement exhibit change irregularly over time. They cannot be eliminated through calibration, but the stochastic impact on measurement results can be reduced through averaging. In view of this, an automatic measurement system for high-speed broadband optical network parameters is proposed, which can perform multiple continuous measurements of optical communication devices after one connection and calibration under the high-speed broadband conditions, thereby avoiding the drift errors caused by repeated disconnection, connection and calibration of devices between different measurements and reducing the impact of random errors within the same connection on the measurement results. Theoretical analysis and experimental results show that this design has better measurement repeatability, higher accuracy and better reliability.

Heavy metal contamination in water solution is increasing, requiring a simple and fast detection method. Laser-induced breakdown spectroscopy (LIBS) technology meets these needs. Applying LIBS to detect heavy metals in water solution faces problems such as the contamination of optical devices for liquid splashing and plasma quenching. Without preconcentration, in the experiment, the lead (Pb) detection sensitivity in water solution is improved by using thin laminar flowing, spatial confinement and adjusting water flowing speed, which has the characteristics of simple device and fast detection. LIBS technology is used to detect heavy metals in water solution, by optimizing experimental parameters such as characteristic spectral line, laser energy and delay time, the spectral line intensity is eventually increased by four times, which has a high application value.

A new tunable molecular filter is developed based on gas molecular filtering technology, through deep investigation of the infrared spectrum characteristics of carbon monoxide (CO) gas. The filter combined with differential absorption spectroscopy technology, through two independent spectral channels, the background noise is effectively eliminated, thereby the detection accuracy of CO gas is enhanced. To verify the performances of the system, a series of experiments are designed such as concentration and gas chamber length verification experiments. The precise control for the infrared transmittance of CO gas is realized through changing the gas concentration in gas chamber and the length of the gas chamber. The strong absorption peak of CO at a wavelength of 4.6 m is particularly focused on. Using this tunable molecular filter, it is possible to achieve a transmittance over 80% for CO at the 4.6 m peak wavelength, which provides favorable conditions for subsequent imaging application.

Based on a 640×512 uncooled detector with a pixel size of 12 m×12 m, a high performance long wave infrared continuous zoom system is designed by the continuous zoom optical system design theory, with a focal length of 30~180 mm and a telephoto end F number of 1.2. Infrared materials such as germanium and zinc selenide are used in the system, and the off-axis aberration of the system is corrected by aspheric surface. The image quality of the system is evaluated from the aspects of modulation transfer function (MTF), spot diagram, distortion, etc. After the cam curve is optimized, the pressure rise angle of the zoom group in the short focal length is reduced, and the zooming process is smooth. Experimental results show that the system has the advantages of good imaging quality, smooth cam curve and compact structure, which is suitable for infrared monitoring field.

Aming at the problem that the detection accuracy of infrared small target is not high in changeable and complex background environment, the advantages and disadvantages of typical algorithms are analyzed, the characteristics of infrared small target images are studied, and an infrared small target detection algorithm based on feature enhancement and multi-scale feature fusion is designed. A feature enhancement module based on multi-scale feature fusion is proposed to enhance the expression of target features. A multi-scale feature information fusion module enhances the target boundary and theme information. A feature fusion decoding module realizes scene and target segmentation. Experimental results show that the infrared small target detection algorithm based on feature enhancement and multi-feature fusion can segment infrared small targets in complex background. In IRSTD-1K dataset, the intersection over union is 60.25%, false alarm rate is 19.6×10-6 and probability of detection is 85.71%. In NUDT-SIRST dataset, the intersection over union is 87.13%, false alarm rate is 9.14×10-6 and probability of detection is 96.08%. Compared with other existing methods, it has certain advantages and strong segmentation performance.

A non-uniformity correction method for infrared images based on calibration combined scene is proposed to address the problems in the use of calibration and scene based methods. The blackbodies at different temperatures in the laboratory are used to construct second-order calibration parameter of infrared detection systems. Infrared images are performed non-uniformity correction by neural network algorithms to adjust the correction coefficients dynamically. Algorithm verification is performed through constant temperature blackbody and real external scenes, it is proved that the non-uniformity of the image can be reduced effectively, which is an effective method for infrared image non-uniformity correction.

Aiming at the problems of low contrast, blurred edge and easy to be disturbed by noise of infrared targets in complex background, an infrared target detection algorithm based on deep learning is proposed. At first, the image is preprocessed by image smoothing and Gamma transform. And then, the local patch network (LPNet) with global concern is used to detect the small target by considering the global and local characteristics of the infrared small target image. From a global perspective, it is a supervised attention module trained by a small target diffusion graph to suppress most background pixels unrelated to small target features. From a local perspective, local patches are separated from global features and share the same convolutional weights with each other in the patch network. By utilizing both global and local characteristics, the data-driven framework of the proposed algorithm incorporated with multi-scale features is used to small target detection. The experimental results on SIRST dataset show that the proposed algorithm can effectively realize infrared target detection under complex background, which verifies the effectiveness and robustness of the proposed algorithm.

With the development of infrared imaging technology, the resolution and frame rate of infrared imaging detection equipments are getting higher and higher, which leads to image processing systems shall have more powerful data processing capability. A multi-channel infrared image processing system is designed based on FPGA, DSP and CPU. And the acquisition, transmission, processing and mass storage of five-channel infrared images is realized, which meet the image processing requirements of multi-channel infrared detection equipments. It is verified by practical application that the system has stable operation and good operation effect.

As the core actuator of high-precision optical systems, the control strategy of fast steering mirror directly affects the performances of the system. Traditional control strategy of the fast steering mirror relies on independent control boxes and linear power amplifier modules, resulting in large volume, high heat and limited application of fast steering mirror control systems. However, the fast steering mirror driven by metal-oxide-semiconductor field-effect transistor (MOSFET) switch values is difficult to achieve accuracy requirements in control. A control strategy combining open-loop zero pole configuration and proportional integral derivative (PID) through modeling and frequency sweep analysis of fast steering mirror is proposed, which is simulated on the hardware of the fast steering mirror. Experimental results show that this control strategy can ensure the high-precision requirements of fast steering mirror control without relying on additional controllers and driving modules.

With the continuous development of the manufacturing industry, aluminum alloy box parts have been widely used in fields such as electronic information, instruments and meters, aerospace. The complex structure and high performance requirement have brought challenges to the manufacturing process. Based on the design for manufacturing (DFM) technology, an in-depth analysis of the manufacturability of aluminum alloy box parts is conducted. By expounding the principle and method of DFM, combining with the characteristics of aluminum alloy box parts, the model based definition technology is used to represent processing and manufacturing information. According to the own manufacturing technology and technology level of enterprises, a process rule library is constructed, and the manufacturability analysis of the three-dimensional product models can be realized. Through the inspection and verification for the manufacturability of the design models, unnecessary design and process changes during the manufacturing stage is avoided, the design quality is enhanced, the design cost is reduced, and the product development cycle is shortened.

A device for testing recoil force on a plane is designed, which is used to analyze the force-exerting situation of the supporting parts of equipment a planar contact mounting. The core of the device is a piezoelectric sensor, by taking advantage of its high sensitivity and fast response characteristics, accurate capture of recoil force signals is achieved. A mathematical model of the recoil force is established through relevant mechanical principles to describe the quantitative relationships between the recoil force and various influencing factors. With the help of dynamic analysis, the dynamic behavior of the device during the testing process is deeply analyzed to clarify the force-bearing and motion states of each component, the influence of various parameters on the testing results is analyzed, and the structural design is optimized. Experimental verification is carried out. By comparing the experimental data with the theoretical analysis data, the rationality and accuracy of the device design are verified.