Study On Broadband Signal Processing And High-speed Imaging Based On Photonic Technology

Information field has always been a global science and technology industry,which is also listed as one of the current innovation development strategy objectives in the 2020 national "14th Five-Year Plan".However,as the performance requirements of information systems continue to improve,communication rates and bandwidth are becoming bottlenecks.In order to solve this problem,microwave photonics is proposed as a key technology.It integrates photonic technology with microwave radio frequency technology,which has the advantages of large bandwidth,high speed,low loss and antielectromagnetic interference,and is widely used in communication,sensing,aerospace,military defense,biomedical and other fields.Photonic technology has not only made significant progress in broadband signal processing,but is also prospective to play an important role in high-speed imaging.Under the supports of the National Natural Science Foundation of China,this dissertation has thoroughly carried out a series of theoretical,simulated,and experimental study on the broadband signal processing and high-speed imaging based on photonic technology,including microwave photonic generation,instantaneous microwave frequency measurement and time stretch imaging system.The main innovations of the work are summarized as follows:1.A scheme for the generation of frequency-doubling microwave photonic signal with tunable phase shift is proposed and simulated analyzed.By dual-parallel polarization,two sidebands with suppressed carrier along the orthogonal polarization directions are obtained,and then the phase shift is introduced by a phase modulator.After beating frequency,a frequency-doubling microwave signal with 0?360°full range phase shift is achieved,which has a range of 22?40GHz and a nearly flat phase response.The scheme avoids the use of filters and wavelength related components,so it owns good frequency tunability.2.A scheme for the generation of frequency-doubling triangular-shaped pulse signal based on spectrum manipulation is proposed and experimentally studied.With intensity modulation of an 10 GHz input sinusoidal signal,five even-order sidebands in the spectrum are achieved.Then a spectrum shaper performs manipulation on these sidebands.It changes the time domain characteristic of the modulated signal by affecting its spectrum,and finally achieves a frequency-doubling triangular-shaped pulse signal of 20 GHz.The scheme can use a flexible modulation index,so it owns a good tunability of the repetition frequency.3.The mechanism of tunable instantaneous frequency measurement(IFM)system based on polarization modulation is studied.The principle of polarization modulation and polarization beam splitting is used to load and separate microwave signals in different optical polarization dimensions,and the radio frequency power fading effect induced by dispersion is used to obtain amplitude comparison function(ACF)for frequency estimation.Two control approaches by adjusting polarization angle or direct-current bias voltage can be used to achieve a tunable measurement range of 2?17.3GHz and resolution of <±0.15 GHz.Then tunable IFM systems with single ACF and multiple ACFs are proposed and studied,which improve both measurement range and resolution.The maximum measurement range is 2?20GHz and the highest resolution is around 100 MHz.4.A time stretch high-speed imaging system based on deep learning is proposed and experimentally studied.Quantitative phase imaging is applied to achieve label-free high throughput cell imaging,and then a deep convolutional neural network is used in data processing to effectively reduce processing time to as fast as milliseconds.The system has a high accuracy with balance accuracy of 95.74% and F1 score of 95.71%while accelerating cell classification,making it possible to perform cell sorting in flow cytometry.