Self-calibration ultrashort pulse laser shape measuring device and method

Abstract

The invention discloses a self-calibration ultrashort pulse laser shape measuring device and method, and belongs to the technical field of ultrashort pulse laser. The device comprises a laser, a beam splitter, a first plane mirror, a time domain scanning module, a second plane mirror, a third plane mirror, a fourth plane mirror, a dispersion scanning module, a fifth plane mirror, a sixth plane mirror, an off-axis parabolic mirror, a nonlinear optical medium, a spectrum detection module and a computer terminal. According to the method, a neural network is optimized by using a particle swarm algorithm, so that laser pulse shape reconstruction is realized, and to-be-measured laser pulse shape and width parameters are accurate. According to the invention, the accuracy and stability of ultrashort pulse laser time domain width and shape measurement can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of ultrashort pulse lasers, and in particular relates to a self-calibration ultrashort pulse laser shape measurement device and method. Background Art

[0002] Ultrashort laser pulses, which refer to lasers with durations on the picosecond or femtosecond scale, have become a core tool for exploring scientific frontiers and advancing technological innovation, finding widespread application in fields such as physics, chemistry, and precision machining. Measuring the time-domain characteristics of lasers directly impacts the applicability of femtosecond laser pulses in these fields, providing critical input parameters for laser technology research and application. Currently, commonly used methods for measuring the time-domain parameters of ultrashort laser pulses include autocorrelation, frequency-resolved optical switching, and spectral coherence electric field reconstruction. While the autocorrelation method is relatively simple to implement, it does not capture the full range of pulse information. Spectral phase interferometry direct electric field reconstruction utilizes interference between the spectrum of a known pulse and its frequency-shifted replica, then records the resulting spectral interference pattern. This allows for rapid measurement of pulse width and phase information, but the setup is complex, difficult to adjust, and expensive. Compared to the first two methods, the frequency-resolved optical shutter method decomposes the spectrum of the autocorrelation signal to generate a trajectory pattern. Using an integration algorithm, the full characteristics of the measured pulse can be obtained. However, this method requires an assumption about the shape of the measured pulse, requiring numerical calculations to determine the pulse shape under the assumed waveform. This leads to large initial pulse estimation errors and susceptibility to noise. Summary of the Invention

[0003] In response to the problems of low accuracy, complex optical path, and poor signal-to-noise ratio performance in the existing technology, the present invention proposes a self-calibrating ultrashort pulse laser shape measurement device and method. This method combines the two basic technologies of frequency domain interferometry technology and dispersion scanning technology. By measuring the second harmonic signal generated by two beams of light on the nonlinear optical medium under different delay scanning and dispersion scanning, the two measurement results can be verified with each other, which can improve the accuracy and stability of ultrashort pulse laser time domain width and shape measurement. It has the advantages of high precision and strong anti-interference ability, and provides technical support for diagnostic research and engineering applications of ultrashort pulse laser technology.

[0004] The technical solutions of the present invention are as follows:

[0005] A self-calibrating ultrashort pulse laser shape measurement device comprises a laser, a beam splitter, a first plane reflector, a time domain scanning module, a second plane reflector, a third plane reflector, a fourth plane reflector, a dispersion scanning module, a fifth plane reflector, a sixth plane reflector, an off-axis parabolic reflector, a nonlinear optical medium, a spectrum detection module and a computer terminal;

[0006] The laser emits a femtosecond laser beam to be measured, which generates two laser pulses with the same amplitude and orthogonal directions after passing through a beam splitter. One of the laser pulses passes through a first plane reflector, a time domain scanning module, and a second plane reflector in sequence to generate a laser pulse with a time delay, and the other laser pulse passes through a third plane reflector, a fourth plane reflector, a dispersion scanning module, a fifth plane reflector, and a sixth plane reflector in sequence to generate a laser pulse with a dispersion bias. The two laser pulses pass through the beam splitter to form two parallel lights, which are focused on the center of a nonlinear optical medium through an off-axis parabolic reflector. The second harmonic signals generated by the two laser pulses in the nonlinear optical medium are collected and measured by a spectral detection module, and a computer terminal processes the collected data to obtain the shape and phase information of the femtosecond laser pulse to be measured.

[0007] Furthermore, the time domain scanning module includes a precision translation stage and two plane reflectors, and the two plane reflectors are symmetrically placed and fixed on the precision translation stage.

[0008] Furthermore, the dispersion scanning module includes a precision translation stage and two wedge prisms, the two wedge prisms are fixed on both sides of the optical axis, the light beam is incident on the surface of the first wedge prism in sequence along the optical axis, and is emitted from the surface of the second wedge prism after refraction; the two wedge prisms are fixed on the precision translation stage, and the precision translation stage drives the two wedge prisms to move bidirectionally and synchronously.

[0009] Furthermore, the nonlinear optical medium is made of barium metaborate crystal material.

[0010] Furthermore, the spectrum detection module includes a fiber coupler, a single-mode fiber, and a spectrometer. The second harmonic signal is coupled into the single-mode fiber by the fiber coupler and finally introduced into the spectrometer to detect the interference signal.

[0011] Furthermore, the computer terminal is used for system control and data processing, including controlling the movement of the precision translation stage, executing the laser pulse shape reconstruction model to obtain the shape and phase information of the femtosecond laser pulse to be measured, and outputting and displaying the calculation results.

[0012] A self-calibration ultrashort pulse laser shape measurement method, using the self-calibration ultrashort pulse laser shape measurement device as described above, the method comprises the following steps:

[0013] Step 1: The femtosecond laser pulse to be measured is passed through a beam splitter to generate two sub-beams with the same amplitude. One beam passes through a time domain scanning module, and the other beam passes through a dispersion scanning module. The two beams form an angle between them and are incident on a nonlinear optical medium to generate a second harmonic signal.

[0014] Step 2: Control the glass medium thickness of the dispersion scanning module to zero, control the time domain scanning module to generate laser pulses with different time intervals, and use the spectrum detection module to collect the second harmonic signal of each delayed scan to obtain the initial spectrum trace diagram;

[0015] Step 3: Control the glass medium thickness of the dispersion scanning module to increase, control the time domain scanning module to generate laser pulses with different time intervals, and obtain a spectrum trace diagram with dispersion bias;

[0016] Step 4: Repeat step 3, controlling the glass medium thickness of the dispersion scanning module to increase at equal intervals, and controlling the time domain scanning module to perform time delay scanning of the laser pulse under different glass medium thickness conditions to obtain a series of spectral traces with different dispersion biases;

[0017] Step 5: Build a laser pulse shape reconstruction model and perform training optimization;

[0018] Step 6: Input the dataset consisting of the initial spectral trace diagram and the dispersion-biased spectral trace diagram into the trained laser pulse shape reconstruction model for prediction, and finally reconstruct the shape and width parameters of the laser pulse to be measured.

[0019] Furthermore, in step 5, the specific training process of the laser pulse shape reconstruction model is as follows:

[0020] Step 5.1, initialize the various parameters of the laser pulse shape reconstruction model, including BP neural network structure initialization, particle swarm initialization, and encode the weight vector and bias vector of the neural network into the particle position vector;

[0021] Step 5.2: Use the known femtosecond laser spectral trace dataset as the input of the neural network and calculate the initial fitness value of each particle through the objective function;

[0022] Step 5.3, update the velocity and position of the particle;

[0023] Step 5.4: Determine the optimal state of the current particle by comparing it with the current particle state, and adjust the particle position and velocity based on this result to obtain the current optimal state; if this state is better than the current global extreme value, then it is considered the current optimal state;

[0024] Step 5.5: Determine whether the number of iterations has reached the maximum number of iterations. If so, output the optimal solution for this round; otherwise, return to step 5.3.

[0025] Step 5.6: Decode and convert the optimized particle swarm optimal solution into the weight value and bias value of the BP neural network to achieve accurate prediction of the laser pulse shape.

[0026] The beneficial technical effects brought about by the present invention are:

[0027] (1) A self-calibrated ultrashort pulse laser shape measurement device and method are proposed. The dispersion scanning technology and frequency domain interferometry technology are combined. The known spectral phase is applied to the pulse to be measured, and then the influence of different dispersion bias amounts on the second harmonic signal is measured. The measurement results can be cross-validated and calibrated. The measurement results are highly accurate and can accurately detect the time domain shape and width of the femtosecond laser. It has the advantages of high measurement accuracy and strong anti-interference ability.

[0028] (2) A laser pulse shape reconstruction method based on an improved neural network algorithm is proposed. In order to solve the problems of slow speed and large error in the process of reconstructing the pulse shape using the traditional iterative approximation algorithm, the particle swarm algorithm is used to optimize the selection of the initial parameters and weights of the neural network, and the structural parameters of the neural network model are reasonably selected to make its performance parameters closer to the set target values. The algorithm has the advantages of high reconstruction accuracy and strong robustness, and can accurately invert the shape and width information of the pulse. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a structural diagram of the self-calibration ultrashort pulse laser shape measurement device of the present invention;

[0030] Among them, 1-laser; 2-beam splitter; 3-first plane mirror; 4-time domain scanning module; 5-second plane mirror; 6-third plane mirror; 7-fourth plane mirror; 8-dispersion scanning module; 9-fifth plane mirror; 10-sixth plane mirror; 11-off-axis parabolic mirror; 12-nonlinear optical medium; 13-spectral detection module; 14-computer terminal. DETAILED DESCRIPTION

[0031] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0032] The present invention provides a self-calibrating ultrashort pulse laser shape measurement device and method. This method combines frequency-domain interferometry and dispersion scanning techniques. The laser beam to be measured is divided into two parts. One beam passes through a delay scanning module to form laser pulses with different time intervals, while the other beam passes through a dispersion scanning module to form different degrees of spectral broadening. The device then measures the second harmonic signals generated by the two pulsed lasers in a nonlinear optical medium under different delay and dispersion bias conditions. The spectral interference fringe data is input into a neural network for training, and the shape information of the measured pulse is reconstructed. This method offers the advantages of high measurement accuracy and strong anti-interference capabilities. It can improve the accuracy and stability of ultrashort pulse laser time-domain width and shape measurements, providing technical support for diagnostic research and engineering applications of ultrashort pulse laser technology.

[0033] The present invention provides a self-calibrating ultrashort pulse laser shape measurement device and method. By measuring the second harmonic signal generated by two beams of light on a nonlinear optical medium under different delay scans and dispersion scans, the amplitude and phase data are calculated using a BP neural network. This solves the problems of low measurement accuracy and poor repeatability of traditional autocorrelation interferometers. The device and method have the characteristics of compact structure, simple optical path, stability and easy adjustment.

[0034] In a first aspect, the present invention provides a self-calibrating ultrashort pulse laser shape measurement device, such as Figure 1 As shown, the device includes a laser 1, a beam splitter 2, a first plane reflector 3, a time domain scanning module 4, a second plane reflector 5, a third plane reflector 6, a fourth plane reflector 7, a dispersion scanning module 8, a fifth plane reflector 9, a sixth plane reflector 10, an off-axis parabolic reflector 11, a nonlinear optical medium 12, a spectrum detection module 13 and a computer terminal 14; wherein the laser 1 generates an ultrashort pulse laser beam of the femtosecond order, and the laser beam is divided into two beams of equal proportion after passing through the beam splitter 2, one of which can form laser pulses with different time intervals through the time domain scanning module 4, and the other beam can form different degrees of spectral broadening through the dispersion scanning module 8, and the two beams interfere on the nonlinear optical medium 12 to generate a second harmonic signal, and the spectrum detection module 13 measures the second harmonic signal under different time domain scanning and dispersion bias conditions, and finally the computer terminal 14 performs data processing to calculate the ultrashort pulse laser shape and width parameters. Figure 1 The direction indicated by the arrow is the direction of light propagation.

[0035] The time domain scanning module 4 is composed of a plane mirror and a precision displacement stage. The plane mirror is fixed on the precision displacement stage to realize the light beam returning along the incident direction. By controlling the equal step displacement of the precision displacement stage, the pulse time delay can be realized.

[0036] The dispersion scanning module 8 is composed of a wedge prism and a precision translation stage. The thickness of the glass medium can be adjusted by lateral displacement of the wedge prism, thereby introducing continuously variable dispersion.

[0037] The nonlinear optical medium 12 is made of barium borate (BBO) crystal material, and has a very wide light transmission range, a large phase matching angle, an excellent light loss resistance threshold, broadband temperature matching, and excellent optical uniformity.

[0038] The spectrum detection module 13 includes a fiber coupler, a single-mode fiber, and a spectrometer. The second harmonic beam is coupled into the single-mode fiber by the fiber coupler and finally introduced into the spectrometer, so that the interference signal can be accurately detected.

[0039] The computer terminal 14 is used for system control and data processing, including controlling the movement of the precision translation stage, executing the laser pulse shape reconstruction algorithm, and outputting and displaying calculation results.

[0040] In a second aspect, the present invention provides a self-calibration ultrashort pulse laser shape measurement method, comprising the following steps:

[0041] Step 1: The femtosecond laser pulse to be measured passes through a beam splitter 2 to generate two sub-beams with the same amplitude. One beam passes through a time-domain scanning module 4, and the other passes through a dispersion scanning module 8. The two beams are incident on a frequency-doubling crystal at a specific angle to generate a second harmonic signal. The frequency-doubling crystal here refers to a nonlinear optical medium 12.

[0042] Step 2: Control the glass medium thickness of the dispersion scanning module 8 to zero, control the time domain scanning module 4 to generate laser pulses at different time intervals, and use the spectrum detection module 13 to collect the second harmonic signal of each delayed scan to obtain an initial spectrum trace diagram;

[0043] Step 3: Control the glass medium thickness of the dispersion scanning module 8 to increase, control the time domain scanning module 4 to generate laser pulses with different time intervals, and obtain a spectrum trace diagram with a certain dispersion bias;

[0044] Step 4: Repeat step 3, controlling the glass medium thickness of the dispersion scanning module 8 to increase at equal intervals, and controlling the time domain scanning module 4 to perform time delay scanning of the laser pulse under different glass medium thickness conditions to obtain a series of spectral traces with different dispersion biases;

[0045] Step 5: Construct a laser pulse shape reconstruction model and perform training optimization; the laser pulse shape reconstruction model is a BP neural network model optimized based on a particle swarm algorithm, and is an improved neural network model.

[0046] The laser pulse shape reconstruction model uses an improved neural network algorithm to predict the input scanning spectrum trace diagram and uses a particle swarm algorithm to optimize the weight and bias parameters of the BP neural network, thereby improving the accuracy of the training results. It can solve the problems of slow speed and large errors in the pulse shape reconstruction process of traditional iterative approximation algorithms. The specific training process of the laser pulse shape reconstruction model is as follows:

[0047] Step 5.1, initialize the various parameters of the laser pulse shape reconstruction model, including BP neural network structure initialization, particle swarm initialization, and encode the weight vector and bias vector of the neural network into a particle position vector; the dimension of the particle position vector needs to be the same as the number of variables in the problem to be optimized, so the particle position vector of the present invention is a two-dimensional vector, corresponding to the weight vector and bias vector that need to be optimized.

[0048] Step 5.2: Use the known femtosecond laser spectral trace dataset as the input to the neural network. Calculate the initial fitness value of each particle using the objective function, and use this as a reference to determine the optimal global extremum. The objective function can be, for example, a standard deviation.

[0049] Step 5.3: Update the particle's velocity and position, and ensure that the updated velocity and position are within the specified range.

[0050] Step 5.4: Determine the optimal state of the current particle by comparing it with the current particle state, and adjust the particle position and velocity based on this result to obtain the current optimal state. If this state is better than the current global extreme value, then it can be regarded as the current optimal state.

[0051] Step 5.5: Determine whether the number of iterations has reached the maximum number of iterations. If so, output the optimal solution for this round; otherwise, return to step 5.3.

[0052] Step 5.6: Decode the optimized particle position vectors in sequence to obtain the optimal weight and bias parameters of the BP neural network to achieve accurate prediction of the laser pulse shape.

[0053] Step 6: Input the dataset consisting of the initial spectral trace diagram and the dispersion-biased spectral trace diagram into the trained laser pulse shape reconstruction model for prediction, and finally reconstruct the shape and width parameters of the laser pulse to be measured.

[0054] An embodiment of the present invention provides a self-calibrating ultrashort pulse laser shape measurement device. The devices used in the embodiment device include a laser 1, a beam splitter 2, a first plane reflector 3 (silver-coated, with a diameter of approximately 25 mm), a time domain scanning module 4, a second plane reflector 5 (silver-coated, with a diameter of approximately 25 mm), a third plane reflector 6 (silver-coated, with a diameter of approximately 25 mm), a fourth plane reflector 7 (silver-coated, with a diameter of approximately 25 mm), a dispersion scanning module 8, a fifth plane reflector 9 (silver-coated, with a diameter of approximately 25 mm), a sixth plane reflector 10 (silver-coated, with a diameter of approximately 25 mm), an off-axis parabolic reflector 11, a nonlinear optical medium 12, a spectrum detection module 13 (including a fiber coupler, a single-mode optical fiber and a spectrometer) and a computer terminal 14.

[0055] The laser 1 emits a femtosecond laser beam to be measured, and after passing through the beam splitter 2, two laser pulses with the same amplitude and orthogonal directions are generated. One sub-beam passes through the first plane reflector 3, the time domain scanning module 4, and the second plane reflector 5 in sequence to generate a laser pulse with a time delay, and the other sub-beam passes through the third plane reflector 6, the fourth plane reflector 7, the dispersion scanning module 8, the fifth plane reflector 9, and the sixth plane reflector 10 in sequence to generate a laser pulse with a dispersion bias. The two sub-beams pass through the beam splitter 2 to form two parallel lights, which are focused at the center of the nonlinear optical medium 12 through the off-axis parabolic reflector 11. The second harmonic signals generated by the two laser beams in the nonlinear optical medium are measured by the spectrum detection module 13. The computer terminal 14 is used to control the movement of the precision translation stage and at the same time control the spectrum detection module to perform data acquisition to obtain a spectrum trace diagram data set. After processing, the shape information of the femtosecond laser pulse to be measured is obtained.

[0056] The laser 1 emits femtosecond laser pulses with a central wavelength of 780 nm and a frequency of 80 MHz.

[0057] The beam splitter 2 uses an ultrafast broadband beam splitter with a splitting ratio of 50:50, which can reduce the influence of the lens thickness on the splitting, generate two pulses to be tested with an intensity of 50% of the original pulse, and provide equal group delay dispersion in the transmission arm and the reflection arm.

[0058] The time domain scanning module 4 consists of a precision translation stage and two plane mirrors. The two plane mirrors are symmetrically placed, which can make the incident light return along the original direction and at a certain distance in space to prevent the ultrashort pulse laser from returning to the resonant cavity of the laser 1. The precision moving platform is used to carry the movement of the two plane mirrors to produce precise optical delay.

[0059] The dispersion scanning module 8 comprises two wedge prisms and a precision translation stage. The two wedge prisms are fixed on either side of the optical axis. A light beam is incident on the surface of the first wedge prism along the optical axis and then refracted and emitted from the surface of the second wedge prism, returning to its original direction. The two wedge prisms are fixed to the precision translation stage, which drives the two wedge prisms to move synchronously in both directions, ensuring that the distance relative to the optical axis of the light beam entering each wedge prism is the same.

[0060] The nonlinear optical medium 12 is a barium borate (BBO) crystal, and the cutting angle of the nonlinear optical medium 12 is selected to be the main cross-section collinear phase matching frequency doubling angle, which is calculated to be 29.3°. In order to control the angle at which the pulse to be measured is incident on the nonlinear optical medium 12, the crystal is fixed on a rotatable bracket with a scale and can be rotated 360° as needed. There is a non-collinear angle between the incident reflected light and the transmitted light in the relevant crystal, which causes a spatial delay between the reflected light and the transmitted light at the intersection.

[0061] The spectrometer detection module 13 is fixed to a hand-built two-dimensional displacement platform. The probe center position can be freely adjusted according to the probe size to ensure that the second harmonic is always incident on the probe center. The entire experimental environment is minimized from stray light interference, maximizing the purity of the second harmonic input to the spectrometer probe, thereby improving the signal-to-noise ratio of the generated scanning spectrum trace and ensuring the accuracy of the reconstruction.

[0062] The computer terminal 14 is also used to control the precision translation stage of the dispersion scanning module 8 to generate glass media of different thicknesses to change the phase of the femtosecond laser beam.

[0063] The computer terminal 14 is used to control the precision translation stage of the time domain scanning module 4 to generate different time delays, and at the same time control the spectrometer to continuously scan the frequency-doubled pulse spectrum to generate a delayed scanning spectrum trace diagram.

[0064] The computer terminal 14 controls the spectrum detection module 13 to collect delayed scanning spectrum trace diagrams with different dispersion biases, and inputs the spectrum trace diagrams into the laser pulse shape reconstruction model, obtains the shape and width information of the laser pulse to be measured by calculation, and outputs the results.

[0065] This embodiment provides a self-calibrated ultrashort pulse laser shape measurement method, which includes the following steps:

[0066] First, a femtosecond laser outputs a laser pulse with a central wavelength of 780nm. After passing through a beam splitter, it enters the time domain scanning module 4 and the dispersion scanning module 8 respectively. The two beams are incident on the frequency doubling crystal at a certain angle to generate a second harmonic signal with a central wavelength of 390nm. The second harmonic signal is measured using the spectrum detection module 13, which collects p sampling points within the spectral range.

[0067] Then, the computer terminal 14 controls the optical thickness of the glass medium of the dispersion scanning module 8 to be zero, controls the time domain scanning module 4 to generate n delayed continuous scanning frequency-doubled pulse spectra, and controls the spectrometer to collect spectrum traces of n delayed sampling points;

[0068] Then, the computer terminal 14 controls the optical thickness of the glass medium of the dispersion scanning module 8 to increase from 1 mm to 10 mm, including m optical thicknesses, controls the time domain scanning module 4 to generate a delayed continuous scanning doubled frequency pulse spectrum, and controls the spectrometer to collect each delayed scanning spectrum trace diagram to generate an m×n×p data matrix.

[0069] Finally, the scanning spectrum trace data set of the laser pulse to be measured is input into the trained laser pulse shape reconstruction model to calculate the intensity and shape information of the laser pulse to be measured;

[0070] In this embodiment, the laser pulse shape reconstruction model is a BP neural network model optimized based on the particle swarm algorithm. The weight parameters and bias parameters of the BP neural network are sequentially encoded as the initial position of the particle swarm algorithm. The fitness of each particle in the population is evaluated by the relative standard deviation. By continuously adjusting the speed and position of the particles, the individual extreme values ​​and the overall optimal value are updated until the maximum number of iterations or the specified error range is reached, and the optimal particle position is obtained. The corresponding value after decoding the optimal particle position is the optimal weight and bias of the BP neural network, thereby reconstructing the shape information of the laser pulse to be measured. The model has high accuracy and reliability.

[0071] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the essential scope of the present invention should also fall within the scope of protection of the present invention.

Claims

1. A self-calibrating ultrashort pulse laser shape measurement device, characterized in that: It includes a laser, a beam splitter, a first plane reflector, a time domain scanning module, a second plane reflector, a third plane reflector, a fourth plane reflector, a dispersion scanning module, a fifth plane reflector, a sixth plane reflector, an off-axis parabolic reflector, a nonlinear optical medium, a spectrum detection module and a computer terminal; The laser emits a femtosecond laser beam to be measured, which generates two laser pulses with the same amplitude and orthogonal directions after passing through a beam splitter. One of the laser pulses passes through a first plane reflector, a time domain scanning module, and a second plane reflector in sequence to generate a laser pulse with a time delay. The other laser pulse passes through a third plane reflector, a fourth plane reflector, a dispersion scanning module, a fifth plane reflector, and a sixth plane reflector in sequence to generate a laser pulse with a dispersion bias. Two laser pulses pass through a beam splitter to form two parallel light beams, which are focused at the center of the nonlinear optical medium by an off-axis parabolic reflector. The second harmonic signals generated by the two laser pulses in the nonlinear optical medium are collected and measured by a spectral detection module. The computer terminal processes the collected data to obtain the shape and phase information of the femtosecond laser pulse to be measured.

2. The self-calibrating ultrashort pulse laser shape measurement device according to claim 1, characterized in that: The time domain scanning module includes a precision translation stage and two plane reflectors, and the two plane reflectors are symmetrically placed and fixed on the precision translation stage.

3. The self-calibrating ultrashort pulse laser shape measurement device according to claim 2, characterized in that: The dispersion scanning module includes a precision translation stage and two wedge prisms. The two wedge prisms are fixed on both sides of the optical axis. The light beam is incident on the surface of the first wedge prism in sequence along the optical axis, and is emitted from the surface of the second wedge prism after refraction. The two wedge prisms are fixed on the precision translation stage, and the precision translation stage drives the two wedge prisms to move bidirectionally and synchronously.

4. The self-calibrating ultrashort pulse laser shape measurement device according to claim 3, characterized in that: The nonlinear optical medium is made of barium metaborate crystal material.

5. The self-calibrating ultrashort pulse laser shape measurement device according to claim 4, characterized in that: The spectrum detection module includes a fiber coupler, a single-mode fiber, and a spectrometer. The second harmonic signal is coupled into the single-mode fiber by the fiber coupler and finally introduced into the spectrometer to detect the interference signal.

6. The self-calibrating ultrashort pulse laser shape measurement device according to claim 5, characterized in that: The computer terminal is used for system control and data processing, including controlling the movement of the precision translation stage, executing the laser pulse shape reconstruction model to obtain the shape and phase information of the femtosecond laser pulse to be measured, and outputting and displaying the calculation results.

7. A self-calibration ultrashort pulse laser shape measurement method, characterized in that: Using the self-calibrating ultrashort pulse laser shape measurement device according to any one of claims 1 to 6, the method comprises the following steps: Step 1: The femtosecond laser pulse to be measured is passed through a beam splitter to generate two sub-beams with the same amplitude. One beam passes through a time domain scanning module, and the other beam passes through a dispersion scanning module. The two beams form an angle between them and are incident on a nonlinear optical medium to generate a second harmonic signal. Step 2: Control the glass medium thickness of the dispersion scanning module to zero, control the time domain scanning module to generate laser pulses with different time intervals, and use the spectrum detection module to collect the second harmonic signal of each delayed scan to obtain the initial spectrum trace diagram; Step 3: Control the glass medium thickness of the dispersion scanning module to increase, control the time domain scanning module to generate laser pulses with different time intervals, and obtain a spectrum trace diagram with dispersion bias; Step 4: Repeat step 3, controlling the glass medium thickness of the dispersion scanning module to increase at equal intervals, and controlling the time domain scanning module to perform time delay scanning of the laser pulse under different glass medium thickness conditions to obtain a series of spectral traces with different dispersion biases; Step 5: Build a laser pulse shape reconstruction model and perform training optimization; Step 6: Input the dataset consisting of the initial spectral trace diagram and the dispersion-biased spectral trace diagram into the trained laser pulse shape reconstruction model for prediction, and finally reconstruct the shape and width parameters of the laser pulse to be measured.

8. The self-calibration ultrashort pulse laser shape measurement method according to claim 7, characterized in that: In step 5, the specific training process of the laser pulse shape reconstruction model is as follows: Step 5.1, initialize the various parameters of the laser pulse shape reconstruction model, including BP neural network structure initialization, particle swarm initialization, and encode the weight vector and bias vector of the neural network into the particle position vector; Step 5.2: Use the known femtosecond laser spectral trace dataset as the input of the neural network and calculate the initial fitness value of each particle through the objective function; Step 5.3, update the velocity and position of the particle; Step 5.4: Determine the optimal state of the current particle by comparing it with the current particle state, and adjust the particle position and velocity based on this result to obtain the current optimal state; if this state is better than the current global extreme value, then it is considered the current optimal state; Step 5.5: Determine whether the number of iterations has reached the maximum number of iterations. If so, output the optimal solution for this round. Otherwise, return to step 5.3; Step 5.6: Decode and convert the optimized particle swarm optimal solution into the weight value and bias value of the BP neural network to achieve accurate prediction of the laser pulse shape.

Images