Target identification method based on nonlinear induction characteristic spectrum drift mechanism

Abstract

The invention discloses a target identification method based on a nonlinear induction characteristic frequency spectrum drift mechanism, and the method mainly aims at the interference of the homogeneous and different-spectrum, heterogeneous and same-spectrum and other characteristics of heterogeneous targets on the target classification and identification in the target classification and identification process. Through combination of characteristic spectrum measurement of a target and characteristic spectrum drift characteristic measurement under active induction and fusion of dynamic image information of target operation, space-time fine characteristics such as a target characteristic spectrum, an evolution spectrum, a dynamic spectrum and an image are obtained. The method can support the realization of spectrum analysis of a target under homogeneous and heterogeneous spectrums and spectrum analysis of heterogeneous and same spectrums, and solves the identification problems caused by homogeneous and heterogeneous spectrums and heterogeneous and same spectrums.

Description

Technical Field

[0001] The invention relates to the field of target recognition information technology, in particular to a target recognition method based on a nonlinear induced characteristic spectrum drift mechanism. Background Art

[0002] Based on the development and improvement of the fusion processing of spectral information and the target classification and recognition theory, combined with active detection information, the application in the fields of target feature information mining, feature analysis and recognition, and automatic processing has become increasingly significant, and has continuously developed into one of the important basic means of target recognition. However, with the widespread application of various new materials, new structures, coating treatment and other technologies on various targets, the characteristic intensity of the target has been greatly reduced, and the characteristic types have been continuously reduced, which has seriously affected the target detection and recognition capabilities of the equipment instruments. In addition, the application of new technologies has introduced "invisible phenomena" such as homogeneous and heterogeneous spectra and heterogeneous and homogeneous spectra into the characteristic spectrum of the target, which seriously affects the detection and recognition of the target, bringing huge challenges to the application of photoelectric detection and recognition technology. In order to meet the detection and classification and recognition needs of complex targets, the continuous improvement of the detection system sensitivity, spectral resolution, spatial resolution and other indicators are greatly restricted in terms of hardware systems and technical applications. At the same time, it is still difficult to solve the application problems of target classification and recognition caused by homogeneous and heterogeneous spectra and heterogeneous and homogeneous spectra, which will still be a major problem facing future detection and recognition systems.

[0003] Laser detection is a typical detection method that can obtain three-dimensional characteristic information of the target. When combined with spectral modulation, it forms a multi-(high) spectrum laser detection technology, which can further collect the target spectral characteristic information and further enhance the information dimension of target detection. Even so, the "stealth" characteristics of various new types of target detection and identification are still difficult to deal with. At this stage, by fusing with infrared images, the information dimension of target detection is further increased in the time dimension, forming a spectrum recognition technology that integrates two-dimensional information, point cloud information, elevation information, and spectral information in the time space, which has shown great application value in the field of target detection and identification. However, the phenomenon of isomorphous and heterogeneous spectra and heterogeneous and homogeneous spectra in the characteristic spectra of new targets is still one of the important problems that plague target identification.

[0004] Using laser-induced target radiation characteristic spectrum drift method to increase the information dimension of target characteristic change will be an effective way to improve the probability and recognition ability of target recognition. Using multi-spectral laser to induce target radiation spectrum change, through spectrum drift comparison analysis, combined with infrared dynamic imaging, dynamic tracking and recognition of targets can be achieved, which is expected to solve an extremely important problem in this field. Summary of the invention

[0005] The present invention provides a target recognition method based on the mechanism of nonlinear induced characteristic spectral drift, which is characterized in that: by utilizing the change of the radiation characteristic spectrum on the surface of the target induced by laser, on the basis of the original radiation characteristic spectrum of the target, further collect the data of the spectral drift of the radiation characteristic of the target induced by laser, so as to solve the problem of target recognition caused by the phenomena of same-spectrum with different substances and different-spectrum with the same substance in the target characteristic spectrum. A typical feature is to introduce a method for collecting and fusing target characteristic information integrating laser detection, laser induction and dynamic imaging, and construct the basic information dimension of target recognition with the expansion of the spectral drift characteristic dimension. The nonlinear induced characteristic spectral drift is achieved by inducing the change of the target characteristic spectrum by multi-spectrum lasers, and the core lies in the induction and matching recognition of spectral drift.

[0006] The target recognition of the nonlinear induced characteristic spectral drift mechanism involved in the present invention involves data dimension expansion, data fusion and feature matching, etc. Among them, data dimension expansion mainly refers to: using laser to induce the change of the radiation characteristic spectrum of the target, with the emphasis on introducing the dimension information of the spectral drift of the target characteristic spectrum under the active induction of laser on the basis of the original radiation characteristic spectrum information of the target, and mainly solving the interference problem of the phenomena of same-spectrum with different substances and different-spectrum with the same substance shown by new materials in target recognition. Further utilize the spatio-temporal characteristic information of the image to realize the dynamic recognition of the target.

[0007] The specific implementation method steps are as follows:

[0008] S1. Collect, gather and sort out the spectral distribution information of the target constituent materials for subsequent spectral matching;

[0009] S2. Construct a laser detection system, which consists of a multi-spectrum laser system, a laser output module, an echo receiving system, an echo fine spectrum screening module, etc., and has the functions of adjustable power, wavelength and frequency at the same time, so as to effectively form rich spectral drift excitation conditions;

[0010] S3. Construct an image acquisition module for obtaining the process of target characteristic spectral perturbation, mainly composed of an infrared camera and a signal acquisition module, responsible for acquiring the whole process image information of the spectral change of the target induced by laser;

[0011] S4. Construct a target simulation module, with the emphasis on the output control of the simulation parameters of the target radiation characteristics;

[0012] S5. Turn on the image acquisition subsystem and acquire the dynamic image of the target radiation;

[0013] S6. Turn on the laser detection system, use low-power laser, and initially obtain the relationship between the target characteristic radiation spectrum and the laser detection spectral band by adjusting the laser output wavelength;

[0014] S7. Adjust the spectral distribution of the target radiation characteristics, and further obtain the variation relationship between the target characteristic radiation spectrum and the laser detection spectral band;

[0015] S8. Continuously increase the laser output power, and further obtain the target laser detection spectra under different induced wavelengths and different target spectra. By comparing and analyzing with the low-power laser detection spectra, obtain the characteristic spectral drift characteristic distribution of the target;

[0016] S9. Continuously repeat the above S8 step. In the obtained target images, obtain the target image change characteristics under the induced fine spectral changes;

[0017] S10. Compare with the characteristic information in S1 step to obtain the fine characteristics of the characteristic spectral difference;

[0018] S11. Refer to the fine characteristic information of the spectral difference in S10 step, and fuse the dynamic characteristic information obtained in S9 step to form multi-source data of dynamic characteristics, spectral characteristics, and spectral perturbations;

[0019] S12. Analyze the multi-source data in S11 step, continuously improve the data characteristics, and promote the target classification and recognition;

[0020] In the above S1 step, collect the existing target characteristic spectra and screen the spectral data of the specific target composition materials;

[0021] In the above S2 step, the multi-spectral laser system and the laser output module in the laser detection system are high-power pulsed laser systems that can output multiple spectral bands. The echo receiving system and the echo fine spectrum screening module are laser detection receiving systems, which are responsible for collecting the target reflected laser information and simultaneously receiving the spectral change information after the interaction between the laser and the target;

[0022] In the above S4 step, the target simulation module can be selected but is not limited to various standard black bodies, targets, etc.;

[0023] In the above S5 step, the image acquisition subsystem can be selected but is not limited to acquisition devices such as infrared cameras, spectral imagers, and spectrometers;

[0024] In the above S6 to S7 steps, it is a debugging and information acquisition process of laser detection and induction. The laser wavelength in this process can be obtained through but not limited to a monochromator, and the laser power can be obtained through but not limited to adjusting parameters such as the laser output power and frequency;

[0025] In the above S9 step, the image acquisition subsystem can be selected but is not limited to acquisition devices such as infrared cameras, spectral imagers, and spectrometers;

[0026] In the above S10 step, the image acquisition subsystem can be selected but is not limited to acquisition devices such as infrared cameras, spectral imagers, and spectrometers.

[0027] Advantages of this method: Through the careful construction and coordinated operation of multiple modules, this system demonstrates excellent performance in aspects such as spectral acquisition, laser detection, image acquisition, target simulation, and data processing and analysis. The functions of each module cooperate with each other, the work process is rigorous and orderly, the data acquisition is comprehensive and accurate, the data fusion and analysis methods are scientific and reasonable, providing an efficient and reliable solution for the field of target classification and recognition, solving the target recognition problems brought about by the phenomena of homogenous isospectrum and heterogeneous isospectrum in the target characteristic spectrum, and realizing the dynamic recognition of targets by using the spatio-temporal characteristic information of images, with broad application prospects and important scientific research value. Brief Description of the Drawings

[0028] Figure 1 It is a schematic diagram of a target recognition method based on the mechanism of nonlinear induced characteristic spectral drift.

[0029] Figure 2 It is a schematic diagram of spectral drift in a typical scenario.

[0030] Figure 3 It is a schematic diagram of the dynamic change process of characteristic images in a typical scenario. Detailed Implementation Manner

[0031] The following further elaborates on the present invention in conjunction with the Figures 1-3 accompanying drawings and implementation manners.

[0032] Example 1

[0033] As Figure 1 , the design steps of the method for generating and verifying the dynamic information of laser-induced detection spectral drift are as follows:

[0034] S1. Analyze the target constituent materials, collect / gather / arrange the characteristic spectral information of relevant constituent materials, and at the same time collect the target to be recognized for subsequent spectral matching;

[0035] S2. Build a laser detection and induction composite test system, debug the multi-spectrum laser system, laser emission module, echo receiving system, and echo fine spectrum screening module, and test the output power, wavelength adjustment, frequency adjustment and other functions of the laser module to effectively form rich spectral drift excitation conditions;

[0036] S3. Build an image acquisition subsystem, test the imaging functions, image quality, image resolution, etc. of the infrared camera and signal acquisition module in the subsystem, and collect the whole-process image information of the spectral change of the laser-induced target;

[0037] S4. Build and debug the target simulation module, with the focus on debugging the simulation parameter output control module and functions of the target radiation characteristics to ensure stable output of simulation characteristics;

[0038] S5. Turn on the image acquisition subsystem to acquire the dynamic image of the target radiation;

[0039] S6. Turn on the laser detection system. First, conduct a low-power laser output test. By adjusting the laser output wavelength, preliminarily obtain the variation relationship between the target characteristic radiation spectrum and the laser detection spectral band;

[0040] S7. Adjust the spectral distribution of the target radiation characteristics to further obtain the variation relationship between the target characteristic radiation spectrum and the laser detection spectral band;

[0041] S8. Continuously increase the laser output power to further obtain the target laser detection spectra under different induced wavelengths and different target spectra. By comparing and analyzing with the low-power laser detection spectra, obtain the characteristic frequency spectrum drift characteristic distribution of the target;

[0042] S9. Continuously repeat the above S8 step;

[0043] S10. Based on the image information collected by the acquisition system set in step S5, obtain the target image change characteristics under the fine spectral changes induced;

[0044] S11. By comparing with the characteristic information in step S1, obtain the fine characteristics of the characteristic frequency spectrum difference;

[0045] S12. Refer to the fine characteristic information of the frequency spectrum difference in step S11, and fuse the dynamic characteristic information obtained in step S10 to form multi-source data of dynamic characteristics, spectral characteristics, and frequency spectrum perturbation;

[0046] S13. Analyze the multi-source data in step S12. While gradually improving the data characteristics, complete the target classification and recognition through characteristic comparison;

[0047] During the process of spectral perturbation acquisition and analysis research, continuously repeat steps S6 to S13.

[0048] This specific embodiment is only an explanation of the present invention, and it is not a limitation of the present invention. Those skilled in the art can make modifications without creative contributions to this embodiment according to needs after reading this specification, but as long as it is within the scope of the claims of the present invention, it is protected by the patent law.

Claims

1. A target recognition method based on the mechanism of nonlinear induced feature spectral drift, characterized in that, The specific implementation method steps are as follows: S1. Collect, gather, and organize the spectral distribution information of the target component materials for subsequent spectral matching; S2. Construct a laser detection system, which consists of a multi-spectral laser system, a laser output module, an echo receiving system, and an echo fine spectral screening module, and has the functions of adjustable power, wavelength, and frequency, forming rich spectral drift-induced excitation conditions; S3. Construct an image acquisition module for obtaining the target feature spectral perturbation process, mainly composed of an infrared camera and a signal acquisition module, responsible for collecting the whole-process dynamic information of the laser-induced target spectral change; S4. Construct a target simulation module, with the focus on the output control of the simulation parameters of the target radiation characteristics; S5. Turn on the image acquisition subsystem to collect the dynamic image of the target radiation; S6. Turn on the laser detection system, use low-power laser, and initially obtain the relationship between the target feature radiation spectrum and the laser detection spectral band by adjusting the laser output wavelength; S7. Adjust the spectral distribution of the radiation characteristics of the simulated target to further obtain the relationship between the target feature radiation spectrum and the laser detection spectral band; S8. Gradually increase the laser output power, further obtain the target laser detection spectra under different induced wavelengths and different target spectra, and obtain the target feature spectral drift feature distribution through comparative analysis with the low-power laser detection spectrum; S9. Repeat the above S8 step, and in the obtained target image, obtain the target image change characteristics under the induced fine spectral change; S10. Compare with the feature information in S1 step to obtain the fine features of the feature spectral difference; S11. Refer to the fine feature information of the spectral difference in S10 step, and fuse the dynamic feature information obtained in S9 step to form multi-source data of dynamic features, spectral features, and spectral perturbations; S12. Analyze the multi-source data in S11 step, improve the data features, and support the target classification and recognition.

2. The target recognition method based on the non-linear induced feature spectral drift mechanism according to claim 1, characterized in that, In the S1 step, collect the existing target feature spectra and screen the spectral data of specific target component materials.

3. The object recognition method based on the non-linear induced feature spectrum drift mechanism according to claim 1, characterized in that, In the S2 step, the multi-spectral laser system and the laser output module in the laser detection system are high-power pulsed laser systems that can output multiple spectral bands, and the echo receiving system and the echo fine spectral screening module are laser detection receiving systems, responsible for collecting the target reflected laser information and at the same time receiving the spectral change information after the interaction between the laser and the target.

4. The object recognition method based on the non-linear induced feature spectral drift mechanism according to claim 1, characterized in that, In the S4 step, the target simulation module selects a standard black body or a target.

5. The object recognition method based on the nonlinear induced feature spectral drift mechanism according to claim 1, wherein In the S5 step, the image acquisition subsystem selects an acquisition device such as an infrared camera, a spectral imager, or a spectrometer.

6. The target recognition method based on the non-linear induced feature spectrum drift mechanism according to claim 1, characterized in that, In the S6 to S8 steps, it is a process of laser detection, induction debugging, and information acquisition. The laser wavelength in this process is obtained through a monochromator, and the laser power is obtained by adjusting parameters such as the laser output power and frequency.

7. The object recognition method based on the non-linear induced feature spectral drift mechanism according to claim 1, characterized in that In the S9 step, the image acquisition subsystem selects an acquisition device such as an infrared camera, a spectral imager, or a spectrometer.

8. The object recognition method based on the non-linear induced feature spectral drift mechanism according to claim 1, characterized in that In the S10 step, the image acquisition subsystem selects an acquisition device such as an infrared camera, a spectral imager, or a spectrometer.

Images