Environment-target dynamic adaptive stereoscopic component laser detection system and method

Abstract

The present invention discloses an environment-target dynamic adaptive three-dimensional component laser detection system and method, belonging to the field of laser detection technology. It includes: supercontinuum pulse + single wavelength pulse combined laser emission, beam adaptive focusing scanning, wide-band reflection and plasma emission spectrum fusion reception, pulse waveform and spectrum distribution composite data intelligent processing and target feature extraction. Based on the double-pulse induced breakdown effect, the present invention realizes the integrated reception of reflected laser signals and induced breakdown plasma radiation signals on the basis of enhancing the emission spectrum intensity, and realizes the environment-target dynamic adaptive target three-dimensional component information acquisition through adaptive beam control and data intelligent processing.

Description

Technical Field

[0001] The present invention belongs to the field of laser detection technology, and in particular relates to an environment-target dynamic adaptive stereoscopic component laser detection system and method. Background Art

[0002] As an advanced non-contact active detection method, laser detection has 1) its relatively short wavelength determines its high-precision information collection ability; 2) the narrow beam, short pulse and active controllable mode of laser detection make it have good anti-interference and environmental adaptability, and can obtain high-precision information in complex scene environments; 3) the multivariate composition of the laser active emission signal enables it to accept multiple effects of objects, thereby obtaining multi-dimensional complete characteristics of objects, which can not only measure the movement of objects, but also measure their three-dimensional shape, and better perceive the material composition. Therefore, laser detection has the advantage of obtaining more element information with high precision without interference, making it the most potential to solve the problem of accurate and complete physical property data acquisition needs. For example, pulsed laser radar can obtain the three-dimensional structure of centimeter-level targets at a distance of tens of kilometers, and multispectral laser remote sensing can obtain target spectral information with a spectral resolution of 5nm in tens of wavelength channels in the visible and near-infrared bands; in the non-contact material composition analysis, such as metal smelting and coal analysis, only laser can obtain the chemical element composition and content information of the sample to be tested by analyzing the plasma spectrum generated by the induced breakdown of the object. Therefore, laser detection as an effective method to simultaneously obtain accurate and complete data on the target's three-dimensional structure and composition information deserves in-depth study.

[0003] Laser detection has a certain research foundation in high precision and anti-interference, but the current detection methods have large differences in data accuracy when operating in dynamic changing scenes at far and near, and cannot achieve the high precision of "environmental adaptation"; at the same time, conventional laser remote sensing mainly obtains physical properties such as the three-dimensional structure of the target, and conventional laser induced breakdown mainly measures chemical properties such as composition, which cannot achieve the simultaneous acquisition of the three-dimensional structure and composition distribution of the target. Specifically, multi- / high-spectral full-waveform laser remote sensing based on the laser elastic scattering mechanism has appropriate detection accuracy over a large range, can obtain structural, dynamic and material reflection spectrum information, but cannot obtain the composition content of the object, and the near-field effect is not enough; laser induced breakdown and Raman scattering based on inelastic scattering or stimulated radiation can currently obtain the composition and content of the object, but cannot obtain physical properties such as the external morphology and dynamic characteristics of the object, and have high requirements for the operating environment. In short, both of the above detection mechanisms have the problem of simultaneous detection of physical and chemical properties with high detection accuracy requirements and completeness of environmental adaptability. Therefore, while fully leveraging its advantages of high spatiotemporal resolution and active controllability, laser detection urgently needs to focus on solving the problem of "dynamically adaptive three-dimensional component distribution detection of environment-target".

[0004] The advancement and application of laser technology has led to the development of laser detection technology. After decades of development, the remote sensing measurement and spectral detection theory and technology of laser detection have achieved target ranging based on laser echo flight time measurement, target velocity measurement based on laser echo Doppler frequency shift measurement, target structure three-dimensional imaging based on laser traversal precision scanning and ranging, target surface material identification based on laser echo spectrum detection, target element composition content detection based on laser induced breakdown (LIBS) and Raman spectroscopy measurement, and other target multi-dimensional information acquisition technologies.

[0005] (1) Dynamic ranging and speed measurement of targets and three-dimensional structural imaging are developing towards multi-spectral / hyperspectral detection based on laser remote sensing technology.

[0006] In response to the need to obtain more dimensional information about the target, researchers are also looking for the development direction of laser detection technology. On the one hand, the target dynamics and structure can be realized by measuring the single-wavelength laser echo. On the other hand, the target properties closely related to the spectrum, such as the target material type, require spectral detection of light signals of different wavelengths. At present, the multi-spectral detection and laser radar ranging, speed measurement, and imaging developed by combining multi-spectral detection are effective means for laser detection technology to achieve multi-dimensional information acquisition of the target. This technology mainly uses a supercontinuum light source or a combination of multi-wavelength light sources as the transmitting end, combined with multi-channel spectroscopic gratings or liquid crystal tunable filters (LCTF), acousto-optic tunable filters (AOTF) and other filtering elements to realize the acquisition of multi-wavelength full-waveform laser echo data, and has the ability to obtain multi-domain signal characteristics such as spectrum, waveform, polarization, and frequency shift, and is mainly used for target dynamics, morphological measurement and surface material classification.

[0007] In addition to the Doppler shift and micro-Doppler effect caused by motion, this detection method also means that the laser echo information reflected from the target is not modulated by the components inside the target, and the composition content of the object being measured cannot be determined, which limits its application in more precise quantitative remote sensing detection and industrial process target property measurement. In addition, the planar resolution of laser remote sensing is closely related to its scanning interval and spot size. Currently, almost all laser remote sensing emits lasers at a certain divergence angle. For extremely small targets in industrial processes, the resolution of scanning laser imaging and measurement is obviously insufficient. Especially when short-range operations require extremely high resolution, new emission and reception methods need to be sought.

[0008] (2) LIBS technology for measuring target component content improves the accuracy of component analysis by combining it with other spectral data while also hoping to obtain object morphology information.

[0009] Although the detection method of multispectral remote sensing has been proven to have excellent ability to obtain multi-dimensional information of the target at the same time, the reflection spectrum of remote sensing cannot reflect the chemical properties such as the precise composition content of the target. In LIBS technology, in order to solve the problem of laser focusing on complex target surfaces and determine their volume characteristics, it is often necessary to obtain target morphological information. In the current research, CCD photography is combined with LIBS technology in scrap steel sorting to obtain target color images and three-dimensional structures at the same time, realizing scrap steel classification and LIBS automatic focusing. At the same time, it can be seen that the use of traditional CCD camera visualization optimizes ablation focus and sample positioning by capturing optical images of the sample surface.

[0010] In order to simultaneously obtain the component content information of different positions of the target, people use lasers to explore the sample surface in a scanning configuration mode, obtain a spectrum for each spatial position of the predefined grid at a time, and generate an element distribution image superimposed on the target image, forming a laser-induced breakdown spectroscopy imaging (LIBS imaging) technology, which is the simultaneous acquisition of target morphological information and component content information. However, most LIBS imaging instruments rely on an XY stage that moves the sample surface instead of the laser beam, and in online control application scenarios, it is necessary to scan a large sample surface at a high frequency, and the measurement needs to be performed by moving the laser beam. Therefore, the current LIBS imaging technology is still mainly based on close-range microstructures and is not suitable for acquiring long-distance target properties.

[0011] In order to ensure accurate acquisition of composition information, people combine LIBS with other spectral data based on signal enhancement and improved inversion algorithms. For example, the remote LIBS-Raman combined active spectrometer analyzes information in atoms and molecules in materials; the Raman, LIBS, and visible-infrared reflectance spectroscopy multispectral combination commonly used in Mars exploration, etc., have been proven to be able to improve the classification accuracy of several chemometric methods.

[0012] Among the above methods, the simultaneous acquisition of composition and morphological information using LIBS imaging as the main method has not yet been applied to long-range target detection and changing operation scenarios, and the imaging speed is slow and the amount of data is huge. At the same time, in the fusion with other spectral data, these spectra of the target objects are relatively independently acquired, and are obtained in time-sharing by different light sources and detection systems. After acquiring the spectral data, various data fusion algorithms are used to perform component analysis. This data fusion technology still belongs to the data-end fusion after multi-system detection, and the fusion effect is limited by the performance of each detection system.

[0013] In order to obtain the three-dimensional distribution of components, it is necessary to combine laser remote sensing and laser induced breakdown spectroscopy (LIBS) into one. This faces the problem of integrating the two detection mechanisms. In the process of achieving organic integration, it is inevitable to encounter objective realities such as difficulty in describing the physical process, harsh beam control conditions, multiple coupling of signals in time, space and frequency domains, difficulty in processing high-dimensional data, and difficulty in experimental verification. Therefore, there are the following technical bottlenecks that need to be overcome:

[0014] (1) Laser remote sensing and induced breakdown integrated multi-spectral combined laser source signal and optical system design technology that meets the needs of three-dimensional dynamic measurement of component distribution;

[0015] (2) High-precision control technology for accurate target positioning and precise focusing of scanning laser beams applicable to dynamic scenes;

[0016] (3) Highly sensitive and accurate detection technology for the composite spectral signal of backscattering and excited plasma emission of laser signals modulated by multiple targets;

[0017] (4) Scene decoupling signal processing and accurate inversion technology of target object structure and component properties in dynamic scenes.

[0018] To summarize, the above-mentioned technical bottlenecks require new solutions from the aspects of light source, optical system, laser control, echo detection and processing inversion, and a new technical system needs to be proposed. Within this system framework, a laser beam of specific composition is selected to act on the target so that effects such as scattering and breakdown can multiply modulate the laser. This is achieved through signal-level fusion detection and intelligent data processing, starting from a unified physical process. Summary of the invention

[0019] In response to the above technical problems, the present invention proposes a laser detection solution based on environment-target dynamic adaptive stereo components.

[0020] The first aspect of the present invention discloses an environment-target dynamic adaptive three-dimensional component laser detection system, the system comprising: a high-energy single-wavelength pulse transmitter, a supercontinuum pulse transmitter, an optical system, a transmission optical fiber, a spectroscopic and detection system, and a data processing system; wherein:

[0021] The high-energy single-wavelength pulse laser emitted by the high-energy single-wavelength pulse transmitter irradiates the target to stimulate and generate a plasma cloud; at the same time, the supercontinuum pulse laser emitted by the supercontinuum pulse transmitter irradiates the target to form an optical signal;

[0022] Wherein, the optical signal includes a laser signal scattered by a target, a laser signal scattered by a plasma, and a signal emitted by the plasma cloud itself;

[0023] After the optical signal is received by the optical system, it is transmitted to the spectroscopic and detection system through the transmission optical fiber for fusion detection, and the data processing system processes the detection result and outputs the three-dimensional distribution result of the target component.

[0024] According to the system of the first aspect of the present invention, the combined light source uses a high-energy single-wavelength pulse laser and a supercontinuum pulse laser to achieve dual-pulse laser induced breakdown, and the absorption and scattering effects of the two laser beams on the target surface and in the plasma are reflected in the optical signal.

[0025] According to the system of the first aspect of the present invention, the target's motion state, structural morphology, material type, and component content are used as target attribute features to be acquired, and the target attribute features are modulated in the laser signal or plasma spectrum through the interaction between the laser and the target; specifically, the system includes:

[0026] High-energy single-wavelength pulse laser and supercontinuum pulse laser are used as incident lasers, which induce breakdown and scattering on the target surface. The plasma is heated, expanded, and cooled by the laser and radiates spectral line signals reflecting the element content information. The scattered light of the target passes through the plasma and forms a composite scattering signal with the incident laser scattered by the plasma itself.

[0027] According to the system of the first aspect of the present invention, the optical system adopts an off-axis free-form surface optical system, including an aspherical small secondary mirror and a free-form surface primary mirror; wherein:

[0028] After being collimated, the laser light source is emitted into the aspheric small secondary mirror. By adjusting the angle of the aspheric small secondary mirror, the laser is reflected by the aspheric small secondary mirror and illuminates different areas of the free-form primary mirror. The surface parameters of the primary mirror in different areas are adjusted to achieve theoretical diffraction-limited focusing of several points within the target range by cooperating with the aspheric small secondary mirror. By adjusting the change in the small mirror spacing along the optical axis of the aspheric small secondary mirror, a diffraction-limited convergence light spot is achieved between adjacent convergence points on the aspheric surface.

[0029] According to the system of the first aspect of the present invention, the optical system, based on continuous zoom, has a plane scanning optical mechanical structure that adopts a galvanometer scanning system to achieve plane two-dimensional scanning. The size of the reflector in the galvanometer scanning system will completely cover the light beam incident on the surface of the reflector, and achieve plane two-dimensional scanning on the scanning plane without changing the focusing condition of the light beam; the optical system is a Gregorian optical system, presenting a positive image at the focus; the transmitting scanning galvanometer and the receiving scanning secondary mirror in the galvanometer scanning system are rigidly connected, and spatial scanning of the transmitting light beam and the receiving field of view is achieved through synchronous rotation, and the receiving field of view is overlapped with the transmitting light beam.

[0030] A second aspect of the present invention discloses an environment-target dynamic adaptive stereoscopic component laser detection method, the method comprising:

[0031] The high-energy single-wavelength pulse laser emitted by the high-energy single-wavelength pulse transmitter irradiates the target to stimulate and generate a plasma cloud; at the same time, the supercontinuum pulse laser emitted by the supercontinuum pulse transmitter irradiates the target to form an optical signal;

[0032] Wherein, the optical signal includes a laser signal scattered by a target, a laser signal scattered by a plasma, and a signal emitted by the plasma cloud itself;

[0033] After the optical signal is received by the optical system, it is transmitted to the spectroscopic and detection system through the transmission optical fiber for fusion detection, and the data processing system processes the detection result and outputs the three-dimensional distribution result of the target component.

[0034] According to the method of the second aspect of the present invention, the combined light source uses a high-energy single-wavelength pulse laser and a supercontinuum pulse laser to achieve dual-pulse laser induced breakdown, and the absorption and scattering effects of the two laser beams on the target surface and in the plasma are reflected in the optical signal.

[0035] According to the method of the second aspect of the present invention, the target's motion state, structural morphology, material type, and component content are used as target attribute features to be acquired, and the target attribute features are modulated in the laser signal or plasma spectrum through the interaction between the laser and the target; specifically, the method includes:

[0036] High-energy single-wavelength pulse laser and supercontinuum pulse laser are used as incident lasers, which induce breakdown and scattering on the target surface. The plasma is heated, expanded, and cooled by the laser and radiates spectral line signals reflecting the element content information. The scattered light of the target passes through the plasma and forms a composite scattering signal with the incident laser scattered by the plasma itself.

[0037] According to the method of the second aspect of the present invention, the optical system adopts an off-axis free-form surface optical system, including an aspherical small secondary mirror and a free-form surface primary mirror; wherein:

[0038] After being collimated, the laser light source is emitted into the aspheric small secondary mirror. By adjusting the angle of the aspheric small secondary mirror, the laser is reflected by the aspheric small secondary mirror and illuminates different areas of the free-form primary mirror. The surface parameters of the primary mirror in different areas are adjusted to achieve theoretical diffraction-limited focusing of several points within the target range by cooperating with the aspheric small secondary mirror. By adjusting the change in the small mirror spacing along the optical axis of the aspheric small secondary mirror, a diffraction-limited convergence light spot is achieved between adjacent convergence points on the aspheric surface.

[0039] According to the method of the second aspect of the present invention, the optical system, based on continuous zoom, has a plane scanning optical mechanical structure that adopts a galvanometer scanning system to achieve plane two-dimensional scanning. The size of the reflector in the galvanometer scanning system will completely cover the light beam incident on the surface of the reflector, and achieve plane two-dimensional scanning on the scanning plane without changing the focusing condition of the light beam; the optical system is a Gregorian optical system, presenting a positive image at the focus; the transmitting scanning galvanometer and the receiving scanning secondary mirror in the galvanometer scanning system are rigidly connected, and the spatial scanning of the transmitting light beam and the receiving field of view is achieved through synchronous rotation, and the receiving field of view is overlapped with the transmitting light beam.

[0040] In summary, the present invention adopts the technical system of "supercontinuum pulse + single wavelength pulse combined laser emission, beam adaptive focusing scanning, wide spectrum reflection and plasma emission spectrum fusion reception, intelligent processing of pulse waveform and spectral distribution composite data and target feature extraction". Based on the double-pulse induced breakdown effect, on the basis of enhancing the emission spectrum intensity, the integrated reception of reflected laser signals and induced breakdown plasma radiation signals is realized, and through adaptive beam control and data intelligent processing, the environment-target dynamic adaptive target three-dimensional component information acquisition is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 is a schematic diagram of an environment-target dynamic adaptive target three-dimensional component detection system according to an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of the physical process relationship between the laser and the target multiple actions according to an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of a remote continuous focusing optical system based on an off-axis free reflection surface according to an embodiment of the present invention;

[0045] Figure 4 is a schematic diagram of an optical system based on a Gregorian structure according to an embodiment of the present invention;

[0046] Figure 5 A schematic diagram of a target attribute scene decoupling algorithm according to an embodiment of the present invention;

[0047] Figure 6Schematic diagram of a hierarchical data fusion processing method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] The overall technical scheme of the present invention is: adopting the technical system of "supercontinuum pulse + single wavelength pulse combined laser emission, beam adaptive focusing scanning, wide spectrum reflection and plasma emission spectrum fusion reception, intelligent processing of pulse waveform and spectral distribution composite data and target feature extraction", based on the double pulse induced breakdown effect, on the basis of enhancing the emission spectrum intensity, to achieve integrated reception of reflected laser signals and induced breakdown plasma radiation signals, through adaptive beam control and data intelligent processing, to achieve environment-target dynamic adaptive target three-dimensional component information acquisition.

[0050] The detection scheme proposed by the present invention is implemented as follows (eg Figure 1 As shown in the figure): A high-energy single-wavelength pulse laser first irradiates the target to stimulate the formation of a plasma cloud. At the same time, a supercontinuum pulse irradiates the target. The optical signal formed includes the laser signal scattered by the target, the laser signal scattered by the plasma, and the emission signal of the plasma cloud itself. The above signals are collected by the receiving optical system and transmitted to the spectroscopic and detection system through the optical fiber for fusion detection. The obtained series of data are processed by the data processing system to output the three-dimensional distribution results of the target components.

[0051] 1. Physical description of the action process

[0052] In the present invention, the combined light source adopts single-wavelength high-energy pulse and supercontinuum pulse laser to realize double-pulse laser induced breakdown, and the absorption and scattering effects of the two laser beams on the target surface and in the plasma are reflected in the return signal.

[0053] The target motion state, structural morphology, material type, and component content are the target attribute characteristics that need to be obtained. Through the interaction process between laser and matter, the above information is modulated in the laser signal or plasma spectrum. In this physical process, single-pulse laser and supercontinuum laser are used as incident lasers to induce breakdown and scattering on the surface of the object. The plasma is heated, expanded, and cooled by the laser and radiates spectral line signals reflecting the element content information. After the target scattered light passes through the plasma, it forms a composite scattering signal with the incident laser scattered by the plasma itself. The physical process relationship is shown in the figure below. Figure 2 shown.

[0054] 2. Large dynamic adaptive focusing scanning beam control based on a new combined optical system

[0055] In order to adapt to the large dynamic changes in the relative position of the target and the detection system in practical applications, to adapt to the needs of three-dimensional distribution perception of components of targets with different structures, and to solve the problem of "environment-target adaptation", the present invention adopts a longitudinal and lateral adaptive control method for the emission light beam. First, the detection process of the scene where the system works is characterized based on the physical process model to provide a basic background environment basis for beam control; the spatiotemporal distribution parameters of the dual-pulse laser are determined based on the scanning requirements of laser remote sensing and the requirements for beam energy focusing in induced breakdown; the emission light beam is controlled using a coarse / fine adjustment two-level beam control strategy, and the optical system is specifically designed using two methods: a remote continuous focusing optical system of an off-axis free reflection surface and a transceiver integrated galvanometer precise scanning, so as to achieve longitudinal precise adaptive focusing and lateral precise positioning and flexible scanning, and correct the scanning distortion based on beam position feedback.

[0056] Considering the feasibility and novelty of the experimental system, by optimizing the design of the optical parameters of the traditional reflective optical system, the continuous spectrum light source can be continuously adjusted in a small range (the focusing ability is close to the wide spectrum diffraction limit in a small range), but it is difficult to achieve continuous and effective focusing in a large range. To this end, the present invention adopts the following off-axis free-form surface optical system. Figure 3 As shown in the figure, the system is mainly composed of an aspherical secondary mirror and a free-form primary mirror. The working principle of the system is as follows: the laser light source is collimated and then incident on the secondary mirror. By precisely adjusting the angle of the secondary mirror, the laser beam is reflected by the secondary mirror and illuminates different areas of the primary mirror. The surface parameters of the primary mirror in different areas are combined with the secondary mirror to achieve theoretical diffraction-limited focusing of several points n within the range L. By finely adjusting the change in the micro-mirror spacing of the secondary mirror along the optical axis, a convergent light spot close to the diffraction limit can be achieved between the convergence point of the primary mirror and the next convergence point.

[0057] On the basis of the above-mentioned continuous zoom system, the plane scanning optical mechanical structure of the present invention intends to use a galvanometer scanning system to realize two-dimensional scanning of the plane. As long as the size of the reflector in the galvanometer system is guaranteed to cover all the light beams incident on the reflector surface, the laser two-dimensional scanning on the scanning plane can be realized without changing the focusing condition of the light beam. The receiving optical system used in the present invention is a Gregorian optical system, such as Figure 4 As shown, a positive image is presented at the focus. The transmitting scanning galvanometer and the receiving scanning secondary mirror are rigidly connected, and the spatial scanning of the transmitting beam and the receiving field of view is realized through synchronous rotation, and the receiving field of view and the transmitting beam are strictly matched.

[0058] 3. Multi-domain information fusion detection of composite spectral signals based on time-space frequency domain matching

[0059] The fusion detection of scattered light signals of the target and plasma and the induced breakdown emission signals is a multi-wavelength, multi-pulse complex optical signal detection problem. The present invention is based on the idea of ​​fusion reception of composite spectral signals with time-space frequency domain matching. The corresponding light filtering spectroscopic array detector is selected in the frequency band, the time delay matching is controlled in time to receive laser scattered pulses and plasma discrete spectral line signals, and the field of view is controlled in space to align with the focused light spot. After the signal detection and reception, the coupling matching relationship between the scattered signal and the plasma radiation signal is effectively utilized, the multi-domain information of the composite signal is used for detection, and the matching relationship is used for signal analysis correction and noise reduction processing.

[0060] Based on the pulse accumulation, the signal multi-domain information fusion detection is carried out. Since the plasma radiation spectrum line signal is relatively weak compared with the scattered signal, but the scattered signal needs to record relatively more information dimensions, the present invention uses multi-pulse accumulation to complete the detection of the signal multi-domain information. For the scattered signal, the polarizer direction is adjusted between pulses to realize the detection of signals in different polarization states, and waveform acquisition is completed by multi-pulse accumulation under the same polarization state, and scattered spectrum detection is realized by waveform peak value; for the plasma scattered signal, the spectrum line enhancement is realized by multi-pulse accumulation.

[0061] The waveform, energy spectrum distribution, polarization and other signal characteristics of the detected signal are detected. Based on the time-space-frequency correlation between the scattered signal and the plasma emission signal, the target attribute is preliminarily judged by the laser echo that first reaches the detector, and the possible delay interval of the plasma signal is obtained. When the spectrometer performs plasma spectrum reception delay control, the signal-to-noise ratio of the element radiation spectrum line signal is improved. At the same time, the shielding effect of the broken plasma on the laser signal is used to invert the plasma spectral absorption effect through the obtained plasma radiation spectrum, and the possible spectral distortion of the scattered signal is corrected.

[0062] 4. Intelligent detection of multi-dimensional features of targets based on multi-layer autoencoding and scene decoupling

[0063] In order to give full play to the mutual confirmation relationship between multi-dimensional information and improve the accuracy of simultaneous acquisition of multi-dimensional attributes of the target, the present invention adopts a multi-layer autoencoder deep learning algorithm, combines the physical process model with the machine learning method, and forms a mapping relationship library between target attributes and signal characteristics based on model simulation and detection signal sample data, and establishes a multi-dimensional target attribute precise inversion algorithm based on hierarchical fusion.

[0064] Based on the scene representation and target prior information, the multi-layer self-encoding deep learning algorithm is used to map the physical model of multiple interactions between laser and matter to the machine learning model, and a multi-layer self-encoding deep learning representation mapping relationship solution algorithm is established. Figure 5 As shown in Figure 1, based on the signal representation of the same target under different scene features, the mapping relationship between the target attribute vector, scene feature matrix, and signal representation matrix under multiple scenes is obtained. Since the scene changes but the target attribute does not change, the target attribute decoupled from the scene feature can be obtained by performing an inverse operation on the signal representation matrix on the scene feature.

[0065] In terms of information fusion method, the present invention adopts a detection algorithm of multi-dimensional features with high efficiency fusion of layered fusion, such as Figure 6 As shown. The first layer of data processing uses an artificial neural network feature solution method to extract its signal features in time-sharing to form key feature vectors at different times. The signal vectors at different times are combined to obtain the key feature matrix of the second layer. Among them, the induced breakdown signal is a time-resolved spectrum, and its key spectral feature parameters are obtained by spectral principal component analysis; the reflected echo signal is a multi-wavelength echo waveform, and its key feature parameters most related to the target characteristics (such as multi-polarization echo intensity ratio, waveform characteristics, pulse flight time, etc.) are obtained after being processed by a 2D convolutional neural network; beam control parameters include beam position, system focusing distance, light source beam quality information, etc. The key feature matrix is ​​fused to obtain the multi-dimensional feature fusion matrix of the third layer. Combined with the multiple modulation mechanism model, the principal component analysis of the fusion matrix of the multi-dimensional features is performed, and the fully connected feedforward network is used to realize the target information inversion, and finally obtain the target high-precision multi-dimensional information data.

[0066] V. Typical Scenario Description

[0067] 1. According to the typical application environment and working conditions, determine the variation range of basic detection parameters such as detection distance, spatial scanning angle resolution, radial resolution, and spectral resolution of the experimental system to provide a basic basis for the construction of the experimental system.

[0068] 2. Determine the beam control strategy based on the variation range of the basic detection parameters, combine the design results of the detection system, select and develop key components of the experimental system, and build the experimental system.

[0069] (1) Using the laser radar equation and the typical LIBS excitation threshold, combined with the optical system parameters, the energy, pulse width and wavelength of the LIBS laser are determined to ensure that the experimental system can stably excite the plasma; combined with the spectral response curve, detection bandwidth and optical system loss parameters of the typical detector, the power, pulse width and band range of the supercontinuum laser are determined. That is, through theoretical calculations, the selection basis of the two types of lasers is obtained.

[0070] (2) Based on the aforementioned plasma excitation process model, the emission interval and detection delay of the LIBS laser and supercontinuum laser are determined to ensure high signal-to-noise ratio detection of the plasma radiation spectrum.

[0071] (3) Based on the basic detection parameters determined above, the development of the receiving / transmitting optical system is carried out. In terms of remote focusing optical systems, combined with the radial resolution index, the design and manufacture of high-precision secondary mirrors and off-axis free-form surface reflecting primary mirrors are carried out, focusing on image quality evaluation from aspects such as wavefront and diffuse spots, so that the image quality of each functional optical path meets the working requirements; in terms of receiving optical systems, combined with parameters such as fiber core diameter and NA value, the design and manufacture of high coupling efficiency reflective optical systems are carried out. Based on the finite element principle, with reference to the system working environment, the structural integration design of the experimental system is carried out from the perspective of force and heat; with the help of detection instruments and equipment such as laser interferometers, optical detection scheme research is carried out to complete the precise adjustment of the transmitting and receiving optical system and the overall integration of the system; system calibration research is carried out under typical detection conditions, and the detection results are compared with those of standard instruments to determine the uncertainty of system detection.

[0072] (4) In combination with the aforementioned working mode and spectral resolution parameters, develop high-precision spectroscopic systems and detection systems. After shaping and collimating the optical fiber light, it enters the two detection channels of LIBS spectrum and supercontinuum spectrum. In the LIBS spectrum detection channel, use filters and time delay control to allow the plasma radiation signal to pass through with low loss, and determine the appropriate ICCD and other photoelectric detector models. In the supercontinuum echo spectrum detection channel, use high-resolution diffraction gratings to divide the incident weak echo into multiple bands in space according to the spectrum. Each band is equipped with a focusing optical system and a photoelectric detector to achieve accurate detection of multi-spectral echo waveforms. In the photoelectric detection part, it is necessary to calibrate the response of each spectral detection unit to reduce system errors.

[0073] 3. According to the target information detection needs of actual application scenarios such as scrap steel sorting and metal smelting, combined with the actual working environment and target attributes, through comparative analysis of detection data under complex working conditions, form an applicability evaluation standard based on the integrated detection system, including ranging accuracy, imaging resolution, material recognition accuracy, key component content measurement accuracy, data update speed, etc., and analyze their mutual constraints.

[0074] 4. Based on the hierarchical analysis method, weights are designed for different evaluation criteria, and a dynamic applicability multi-objective function is constructed. The system control strategy and data processing algorithm are optimized based on the evaluation results.

[0075] 5. Based on the experimental results, the applicability characteristics of the technology are clearly revealed. According to the optimization results of the system control strategy and algorithm, the system design and intelligent control solutions for typical application scenarios are proposed, and a database is established.

[0076] In summary, the laser emitted by the present invention is a combination of supercontinuum + single-wavelength pulsed laser, and both remote sensing and induced breakdown use adaptive precise laser focusing. The focused laser performs adaptive traversal scanning on the object, and the induced breakdown plasma is heated by the supercontinuum laser; a new combination of transceiver optical system is designed with a new idea, and the composite spectrum signal adopts the time-space frequency domain matching detection method; the multi-dimensional physical and chemical properties of the target are accurately inverted and solved through intelligent processing methods, and a signal-level fusion and multi-dimensional feature intelligent detection and processing method based on the time-space frequency domain matching physical model of scattered signals and plasma emission signals in the process of laser and matter interaction is established. This detection method will provide new technical means and solutions for the simultaneous measurement of physical and chemical properties of targets under different needs in social and economic activities, and has important research significance and application value.

[0077] For example, in the application of scrap steel sorting, the method proposed in the present invention can quickly obtain the spatial geometric information of scrap steel samples in the target area and the component content information at the corresponding position, and quickly realize the quantity and quality evaluation of scrap steel samples. This makes the scrap steel sorting process more concise, with higher sorting accuracy and faster speed.

[0078] Compared with the prior art, the advantages and positive effects of the present invention are: integrated detection of remote sensing information and laser-induced breakdown spectroscopy information, superimposing accurate chemical composition information on the premise of obtaining target physical location information. Thanks to the integrated detection design of the present invention, the information fusion barrier of non-homologous spectra of the existing means is broken, signal-level fusion is achieved, the target information obtained is generated simultaneously and in the same position, and the signal fusion error is eliminated in the technical system, so higher precision and accuracy can be achieved.

[0079] The dynamic scanning and focused imaging methods proposed in the present invention have not been reported in the current laser induced breakdown technology. This method of three-dimensional distribution measurement of target object composition information by a laser radar scanning method is the first time it has appeared in remote LIBS technology.

[0080] Please note that the technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above-mentioned embodiments only express several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present invention, several variations and improvements can be made, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the attached claims.

Claims

1. An environment-target dynamic adaptive stereoscopic component laser detection system, characterized in that: The system comprises: a high-energy single-wavelength pulse transmitter, a supercontinuum pulse transmitter, an optical system, a transmission optical fiber, a spectroscopic and detection system, and a data processing system; wherein: The high-energy single-wavelength pulse laser emitted by the high-energy single-wavelength pulse transmitter irradiates the target to stimulate and generate a plasma cloud; at the same time, the supercontinuum pulse laser emitted by the supercontinuum pulse transmitter irradiates the target to form an optical signal; Wherein, the optical signal includes a laser signal scattered by a target, a laser signal scattered by a plasma, and a signal emitted by the plasma cloud itself; After the optical signal is received by the optical system, it is transmitted to the spectroscopic and detection system through the transmission optical fiber for detection of multi-domain information fusion of composite spectral signals based on time-space frequency domain matching, and the data processing system processes the detection results and outputs the three-dimensional distribution results of the target components. Based on the multi-domain information fusion of composite spectral signals based on time-space frequency domain matching, the corresponding light filtering spectroscopic array detector is selected in the frequency band, and the time delay matching is controlled in time to receive the laser scattering pulse and the plasma discrete spectral line signal, and the field of view is controlled in space to align with the focused light spot; after the signal detection is received, the coupling matching relationship between the scattering signal and the plasma radiation signal is used to detect the multi-domain information of the composite signal, and the matching relationship is used to perform signal analysis correction and noise reduction processing; The combined light source uses high-energy single-wavelength pulsed laser and supercontinuum pulsed laser to achieve dual-pulse laser-induced breakdown. At the same time, the absorption and scattering effects of the two laser beams on the target surface and in the plasma are reflected in the optical signal. Among them, the target's motion state, structural morphology, material type, and component content are the target attribute characteristics that need to be obtained. Through the interaction between the laser and the target, the target attribute characteristics are modulated in the laser signal or plasma spectrum; specifically including: High-energy single-wavelength pulse laser and supercontinuum pulse laser are used as incident lasers to induce breakdown and scattering on the target surface. The plasma is heated, expanded, cooled by the laser and radiates spectral line signals reflecting the element content information. The scattered light of the target passes through the plasma and forms a composite scattering signal with the incident laser scattered by the plasma itself. Based on the signal representation of the same target under different scene characteristics, the mapping relationship between the target attribute vector, scene feature matrix, and signal representation matrix under multiple scenes is obtained. The scene characteristics are inversely operated through the signal representation matrix to obtain the target attributes decoupled from the scene characteristics. The optical system adopts an off-axis free-form surface optical system, including an aspherical secondary mirror and a free-form surface primary mirror; wherein: After being collimated, the laser light source is incident on the aspheric minor mirror. By adjusting the angle of the aspheric minor mirror, the laser is reflected by the aspheric minor mirror and illuminates different areas of the free-form primary mirror. The surface parameters of the primary mirror in different areas are adjusted to achieve theoretical diffraction-limited focusing of several points within the target range by cooperating with the aspheric minor mirror. By adjusting the distance between the minor mirrors of the aspheric minor mirror along the optical axis, a diffraction-limited convergence spot is achieved between adjacent convergence points on the aspheric surface. Among them, on the basis of continuous zoom, the optical system adopts a galvanometer scanning system in its plane scanning optical mechanical structure to realize plane two-dimensional scanning. The size of the reflector in the galvanometer scanning system will completely cover the light beam incident on the surface of the reflector, and realize the plane two-dimensional scanning on the scanning plane without changing the focusing condition of the light beam; the optical system is a Gregorian optical system, which presents a positive image at the focus; the transmitting scanning galvanometer and the receiving scanning secondary mirror in the galvanometer scanning system are rigidly connected, and the spatial scanning of the transmitting light beam and the receiving field of view is realized through synchronous rotation, and the receiving field of view is overlapped with the transmitting light beam.

2. An environment-target dynamic adaptive stereoscopic component laser detection method, characterized in that: The method comprises: The high-energy single-wavelength pulse laser emitted by the high-energy single-wavelength pulse transmitter irradiates the target to stimulate and generate a plasma cloud; at the same time, the supercontinuum pulse laser emitted by the supercontinuum pulse transmitter irradiates the target to form an optical signal; Wherein, the optical signal includes a laser signal scattered by a target, a laser signal scattered by a plasma, and a signal emitted by the plasma cloud itself; After the optical signal is received by the optical system, it is transmitted to the spectroscopic and detection system through the transmission optical fiber for detection of multi-domain information fusion of composite spectral signals based on time-space frequency domain matching, and the data processing system processes the detection results and outputs the three-dimensional distribution results of the target components. Based on the multi-domain information fusion of composite spectral signals based on time-space frequency domain matching, the corresponding light filtering spectroscopic array detector is selected in the frequency band, and the time delay matching is controlled in time to receive the laser scattering pulse and the plasma discrete spectral line signal, and the field of view is controlled in space to align with the focused light spot; after the signal detection is received, the coupling matching relationship between the scattering signal and the plasma radiation signal is used to detect the multi-domain information of the composite signal, and the matching relationship is used to perform signal analysis correction and noise reduction processing; The combined light source uses high-energy single-wavelength pulsed laser and supercontinuum pulsed laser to achieve dual-pulse laser-induced breakdown. At the same time, the absorption and scattering effects of the two laser beams on the target surface and in the plasma are reflected in the optical signal. Among them, the target's motion state, structural morphology, material type, and component content are the target attribute characteristics that need to be obtained. Through the interaction between the laser and the target, the target attribute characteristics are modulated in the laser signal or plasma spectrum; specifically including: High-energy single-wavelength pulse laser and supercontinuum pulse laser are used as incident lasers to induce breakdown and scattering on the target surface. The plasma is heated, expanded, cooled by the laser and radiates spectral line signals reflecting the element content information. The scattered light of the target passes through the plasma and forms a composite scattering signal with the incident laser scattered by the plasma itself. Based on the signal representation of the same target under different scene characteristics, the mapping relationship between the target attribute vector, scene feature matrix, and signal representation matrix under multiple scenes is obtained. The scene characteristics are inversely operated through the signal representation matrix to obtain the target attributes decoupled from the scene characteristics. The optical system adopts an off-axis free-form surface optical system, including an aspherical secondary mirror and a free-form surface primary mirror; wherein: After being collimated, the laser light source is incident on the aspheric minor mirror. By adjusting the angle of the aspheric minor mirror, the laser is reflected by the aspheric minor mirror and illuminates different areas of the free-form primary mirror. The surface parameters of the primary mirror in different areas are adjusted to achieve theoretical diffraction-limited focusing of several points within the target range by cooperating with the aspheric minor mirror. By adjusting the distance between the minor mirrors of the aspheric minor mirror along the optical axis, a diffraction-limited convergence spot is achieved between adjacent convergence points on the aspheric surface. Among them, on the basis of continuous zoom, the optical system adopts a galvanometer scanning system in its plane scanning optical mechanical structure to realize plane two-dimensional scanning. The size of the reflector in the galvanometer scanning system will completely cover the light beam incident on the surface of the reflector, and realize the plane two-dimensional scanning on the scanning plane without changing the focusing condition of the light beam; the optical system is a Gregorian optical system, which presents a positive image at the focus; the transmitting scanning galvanometer and the receiving scanning secondary mirror in the galvanometer scanning system are rigidly connected, and the spatial scanning of the transmitting light beam and the receiving field of view is realized through synchronous rotation, and the receiving field of view is overlapped with the transmitting light beam.

Images