Target identification method based on frequency modulated continuous wave laser radar and laser radar system

Abstract

The invention provides a frequency modulated continuous wave laser radar-based target identification method and a laser radar system, which can be applied to the technical field of laser radars. The method comprises the following steps: acquiring an interference beat frequency signal after polarization orthogonal demodulation; performing signal demodulation on the electric signal of the interference beat frequency signal to obtain a target detection feature; performing feature matching on the target detection features and a typical target polarization feature library to obtain a feature matching result; and under the condition that the feature matching result represents that the target feature similarity meets a first threshold condition, performing identification processing on the target detection features based on an air intrusion target identification model to obtain a target type identification result.

Description

Technical Field

[0001] The present disclosure relates to the field of laser radar technology, mainly to the field of laser radar detection technology, and more specifically to a target recognition method and a laser radar system based on frequency modulated continuous wave laser radar. Background Art

[0002] Frequency Modulation Continuous Wave (FMCW) lidar, a high-precision active 3D perception technology, has been widely used in recent years for target recognition, imaging, positioning, and tracking. While traditional FMCW lidar can measure distance and velocity, it lacks the ability to discern precise target characteristics and cannot distinguish between drones and low-threat targets like birds and plastic bags. This results in a high false alarm rate, limiting its effectiveness in complex environments and making it difficult to meet the practical needs of aviation safety and border protection. Summary of the Invention

[0003] In view of the above problems, the present disclosure provides a target recognition method and a lidar system based on a frequency modulated continuous wave lidar.

[0004] According to the first aspect of the present disclosure, a target recognition method based on a frequency modulated continuous wave laser radar is provided, wherein the method includes: obtaining an interference beat signal after polarization orthogonal demodulation, wherein the interference beat signal is coherently generated by the orthogonal polarization components of the intrinsic light and the echo light; performing signal demodulation on the electrical signal of the interference beat signal to obtain target detection features, wherein the target detection features include polarization characteristics, morphological characteristics and dynamic characteristics of the airborne intrusion target; performing feature matching on the target detection features with a typical target polarization feature library to obtain feature matching results; when the feature matching results indicate that the target feature similarity meets a first threshold condition, performing recognition processing on the target detection features based on an airborne intrusion target recognition model to obtain a target type recognition result.

[0005] According to an embodiment of the present disclosure, the generation process of the above-mentioned interference beat signal includes: splitting the linearly polarized light output by the laser light source into the above-mentioned intrinsic light and the detection light; performing polarization orthogonal demodulation on the above-mentioned intrinsic light and the above-mentioned echo light respectively to obtain the orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light, wherein the above-mentioned echo light is obtained by reflecting the above-mentioned detection light from the above-mentioned aerial intrusion target; and coherently processing the orthogonal polarization components of the above-mentioned intrinsic light and the orthogonal polarization components of the above-mentioned echo light to obtain the above-mentioned interference beat signal.

[0006] According to an embodiment of the present disclosure, the method further includes: performing balanced reception processing on the interference beat signal to obtain an electrical signal of the interference beat signal.

[0007] According to an embodiment of the present disclosure, the above-mentioned polarization characteristics include echo polarization state parameters, the above-mentioned morphological characteristics of the above-mentioned airborne intrusion target include target distance, and the above-mentioned dynamic characteristics include target radial velocity, wherein the above-mentioned signal demodulation of the electrical signal of the above-mentioned interference beat signal to obtain target detection characteristics includes: determining the above-mentioned target distance based on the electrical signal frequency difference of the above-mentioned interference beat signal and the sweep slope of the frequency modulated continuous wave; performing Doppler frequency shift analysis on the electrical signal of the above-mentioned interference beat signal to obtain the above-mentioned target radial velocity; and calculating the above-mentioned echo polarization state parameters based on the amplitude ratio of the orthogonal polarization components of the above-mentioned echo light.

[0008] According to an embodiment of the present disclosure, when the above-mentioned feature matching result indicates that the target feature similarity does not meet the first threshold condition, the above-mentioned method further includes: when the above-mentioned feature matching result indicates that the target feature similarity meets the second threshold condition, marking the above-mentioned air intrusion target as an unknown target; when the above-mentioned feature matching result indicates that the target feature similarity does not meet the second threshold condition, marking the above-mentioned air intrusion target as a noise interference target.

[0009] According to an embodiment of the present disclosure, the method further includes: updating the typical target polarization feature library based on the target detection feature of the unknown target to obtain an updated typical target polarization feature library.

[0010] According to an embodiment of the present disclosure, the above-mentioned typical target polarization feature library is obtained by the following operations: obtaining multiple typical target detection features; performing feature extraction on the above-mentioned multiple typical target detection features respectively to obtain a target feature vector containing polarization characteristics, morphological characteristics and dynamic characteristics; associating the above-mentioned target feature vector with a category label to obtain the above-mentioned typical target polarization feature library.

[0011] According to an embodiment of the present disclosure, the above-mentioned airborne intrusion target recognition model is obtained by the following operations: using the above-mentioned typical target polarization feature library to optimize and train a multi-classification model based on an ensemble learning algorithm to obtain the above-mentioned airborne intrusion target recognition model.

[0012] The second aspect of the present disclosure provides a laser radar system based on polarization orthogonal demodulation, comprising: a laser light source, an optical beam splitter, a first polarization beam splitter, a second polarization beam splitter, an optical amplifier, an optical circulator, a quarter wave plate, an optical transceiver system, a first coupler, a second coupler, a balanced detector, a field programmable gate array, and a host computer, wherein the input end of the optical beam splitter is connected to the laser light source, and the output end of the optical beam splitter is respectively connected to the first polarization beam splitter and the optical amplifier; the output end of the first polarization beam splitter is respectively connected to the first coupler and the second coupler; the optical The output end of the amplifier is connected to the above-mentioned optical circulator; the output end of the above-mentioned optical circulator is respectively connected to the second polarization beam splitter and the 1 / 4 wave plate, and the output end of the above-mentioned 1 / 4 wave plate is connected to the optical transceiver system; the output end of the above-mentioned second polarization beam splitter is respectively connected to the above-mentioned first coupler and the above-mentioned second coupler; the output end of the above-mentioned first coupler is connected to the first balanced detector; the output end of the above-mentioned second coupler is connected to the second balanced detector; the output ends of the above-mentioned first balanced detector and the above-mentioned second balanced detector are both connected to the above-mentioned field programmable gate array; and the output end of the above-mentioned field programmable gate array is connected to the above-mentioned host computer.

[0013] According to an embodiment of the present disclosure, the above-mentioned optical beam splitter is used to split the linearly polarized light emitted by the above-mentioned laser light source into detection light and intrinsic light; the above-mentioned first polarization beam splitter is used to perform polarization orthogonal demodulation on the above-mentioned intrinsic light to obtain the orthogonal polarization components of the intrinsic light; the above-mentioned optical amplifier is used to power amplify the above-mentioned detection light to obtain the amplified detection light; the above-mentioned 1 / 4 wave plate is used to adjust the polarization state of the above-mentioned amplified detection light to obtain the scanning detection light; the above-mentioned optical transceiver system is used to transmit the above-mentioned scanning detection light to the target area and receive the echo light reflected by the target area; the above-mentioned second polarization beam splitter is used to perform polarization orthogonal demodulation on the above-mentioned echo light to obtain the orthogonal polarization components of the echo light; the above-mentioned first coupler is used to coherently interfere the first orthogonal polarization component of the intrinsic light with the first orthogonal polarization component of the echo light to generate a first interference beat signal signal; the second coupler is used to perform coherent interference on the second orthogonal polarization component of the intrinsic light and the second orthogonal polarization component of the echo light to generate a second interference beat signal; the first balanced detector and the second balanced detector are respectively used to perform balanced reception and conversion processing on the first interference beat signal and the second interference beat signal to obtain the electrical signal of the interference beat signal; the field programmable gate array is used to perform signal demodulation on the electrical signal of the interference beat signal to obtain the target detection feature; the host computer is used to perform feature matching on the target detection feature with the typical target polarization feature library to obtain a feature matching result. When the feature matching result indicates that the target feature similarity meets the first threshold condition, the target detection feature is identified and processed based on the air intrusion target recognition model to obtain a target type recognition result.

[0014] According to an embodiment of the present disclosure, a multi-dimensional target detection feature including polarization characteristics, morphological characteristics, and dynamic characteristics of an airborne intrusion target is obtained by demodulating the interference beat signal. Information fusion is performed on the multi-dimensional target detection feature to achieve complementarity between different features, thereby improving the accuracy of subsequent target recognition. In addition, the target detection feature is feature matched with a typical target polarization feature library. Only when the matching result exceeds a first threshold, the target recognition process of the subsequent airborne intrusion target recognition model is triggered to quickly exclude targets that are obviously mismatched, thereby avoiding low-confidence data interfering with classification decisions, reducing the model calculation load, and solving the problem that traditional FMCW laser radars are difficult to perform fine target recognition due to the lack of information reflecting characteristics such as target surface shape, material, and texture. The detection range and detection accuracy of the radar are improved, as well as the ability to identify targets in complex regional spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0016] Figure 1 The flowchart of the target recognition method based on the frequency modulated continuous wave laser radar according to the embodiment of the present disclosure is schematically shown;

[0017] Figure 2 A schematic diagram schematically illustrates typical target polarization characteristics according to an embodiment of the present disclosure;

[0018] Figure 3 The following schematically shows a flow chart of a target recognition method according to an embodiment of the present disclosure;

[0019] Figure 4 The following schematically shows a training flow chart of an airborne intrusion target model according to an embodiment of the present disclosure;

[0020] Figure 5 The figure schematically shows a schematic diagram of a laser radar system based on polarization orthogonal demodulation according to an embodiment of the present disclosure;

[0021] Figure 6 The following schematically shows a working flow diagram of the laser radar system according to an embodiment of the present disclosure;

[0022] Figure 7 A block diagram schematically shows a frequency modulated continuous wave laser radar target recognition device according to an embodiment of the present disclosure; and

[0023] Figure 8 A block diagram of an electronic device suitable for implementing a target recognition method based on a frequency modulated continuous wave laser radar according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0024] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0025] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0026] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0027] When expressions such as "at least one of A, B, and C, etc." are used, they should generally be interpreted in accordance with the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0028] In the technical solutions disclosed herein, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0029] Frequency Modulation Continuous Wave (FMCW) lidar, a high-precision active 3D sensing technology, has recently been widely used in target recognition, imaging, positioning, and tracking. Its core advantage lies in its ability to achieve high-precision measurement of target range, position, and velocity, while also offering strong anti-interference capabilities and high sensitivity. It is considered a promising lidar technology. FMCW lidar's detection principle is based on the interferometry of a linear frequency-modulated continuous wave (FMCW) with the target's reflected echo. Specifically, the system transmits a FMCW wave toward the target and mixes the echo signal with an intrinsic signal to generate a beat frequency signal. The target's range and velocity are then calculated by analyzing the frequency offset of this signal. More specifically, as the optical signal propagates from emission to reflection by the target, the echo signal's frequency shifts relative to the intrinsic signal due to the target's distance and motion, resulting in a beat frequency signal. By accurately analyzing the frequency offset of this beat frequency signal and combining it with the modulation bandwidth and propagation time, the target's range can be accurately calculated. In addition, the relative motion of the target will cause Doppler frequency shift, and FMCW lidar can directly measure the target's velocity by analyzing this frequency shift.

[0030] FMCW lidar not only provides three-dimensional spatial information about a target (such as range, direction, and altitude), but also acquires motion information (such as velocity) through Doppler frequency shift, thereby acquiring four-dimensional (4D) information. This fusion of multidimensional information provides crucial data support for target recognition, trajectory prediction, and target tracking in complex environments. Compared to traditional TOF lidar, FMCW lidar boasts higher sensitivity and anti-interference capabilities, enabling stable detection in complex environments. Furthermore, its continuous wave modulation method enables the system to achieve both high precision and low power consumption, making it suitable for a variety of fields, including autonomous driving, drone navigation, robotic perception, military reconnaissance, and industrial inspection.

[0031] Currently, the identification of aerial foreign objects (such as drones, kites, birds, hot air balloons, vegetation, and plastic bags) plays a crucial role in airport security, low-altitude economic development, and perimeter security. Its importance lies not only in addressing the potential threat posed by drones in airport cleared zones, border patrols, and sensitive areas, but also in effectively mitigating the serious impact of illegal drone flights and bird strikes on aviation safety. With the rapid adoption of drone technology and the booming low-altitude economy, the types and number of aerial foreign objects are increasing, and their complexity and diversity are placing higher demands on existing detection technologies. While traditional FMCW lidar can measure the distance and velocity of targets, it lacks the ability to precisely identify target characteristics. It cannot distinguish drones from low-threat objects such as birds and plastic bags, resulting in a high false alarm rate. This limits its effectiveness in complex environments and makes it difficult to meet the practical needs of aviation safety and perimeter protection.

[0032] In view of this, the embodiments of the present disclosure provide a target recognition method based on frequency modulated continuous wave laser radar, which obtains multi-dimensional target detection features including polarization characteristics, morphological characteristics and dynamic characteristics of airborne intrusion targets by demodulating the interference beat frequency signal, and achieves complementarity between different features by fusing information of the multi-dimensional target detection features, thereby improving the accuracy of subsequent target recognition. In addition, the target detection features are matched with the typical target polarization feature library, and only when the matching result exceeds the first threshold, the target recognition process of the subsequent airborne intrusion target recognition model is triggered to quickly exclude obviously mismatched targets, avoid low-confidence data interfering with classification decisions, reduce the model calculation load, and solve the problem that traditional FMCW laser radars are difficult to perform fine target recognition due to the lack of information reflecting the characteristics of the target surface shape, material and texture, thereby improving the radar's detection range and detection accuracy and the ability to identify targets in complex regional spaces.

[0033] Specifically, an embodiment of the present disclosure provides a target recognition method based on a frequency modulated continuous wave lidar, including: obtaining an interference beat signal after polarization orthogonal demodulation, wherein the interference beat signal is coherently generated by the orthogonal polarization components of the intrinsic light and the echo light; performing signal demodulation on the electrical signal of the interference beat signal to obtain target detection features, wherein the target detection features include polarization characteristics, morphological characteristics and dynamic characteristics of the airborne intrusion target; performing feature matching on the target detection features with a typical target polarization feature library to obtain feature matching results; when the feature matching result indicates that the target feature similarity meets a first threshold condition, performing recognition processing on the target detection features based on an airborne intrusion target recognition model to obtain a target type recognition result.

[0034] Figure 1The flowchart of the target recognition method based on frequency modulated continuous wave laser radar according to an embodiment of the present disclosure is schematically shown.

[0035] like Figure 1 As shown, the method includes operations S11 to S14.

[0036] In operation S11 , an interference beat signal after polarization orthogonal demodulation is acquired.

[0037] According to an embodiment of the present disclosure, polarization orthogonal demodulation can be used to decompose the echo light into two orthogonal polarization components by utilizing the difference in polarization characteristics of the target reflected light, and generate an interference beat signal coherently with the corresponding components of the intrinsic light to enhance the material recognition capability.

[0038] According to embodiments of the present disclosure, an interference beat signal is a signal generated by the interaction of two optical signals of different frequencies. For example, when two original signals (e.g., sinusoidal signals with frequencies f1 and f2, where f1 ≠ f2) are added together, a beat phenomenon occurs, and a low-frequency signal with a frequency of |f1-f2| appears in the composite signal. This low-frequency signal is the interference beat signal.

[0039] In a specific embodiment of the present disclosure, an interference beat signal can be obtained by cohering the orthogonal polarization components of the intrinsic light and the echo light, wherein the interference beat signal carries a target detection feature of an aerial intrusion target.

[0040] Among them, the intrinsic light can represent the reference light generated inside the lidar and not propagated externally. Therefore, the intrinsic light can be used as a reference signal to interfere with the echo light to generate an interference beat signal carrying target information.

[0041] Echo light refers to the laser light emitted by a lidar radar that reflects off a target (such as a drone or bird) and returns to the system. This echo light carries information about the target's distance, speed, material, or polarization characteristics. Its frequency, phase, and polarization state can change depending on the target's motion, material, and surface characteristics.

[0042] In operation S12, the electrical signal of the interference beat signal is demodulated to obtain a target detection feature.

[0043] According to embodiments of the present disclosure, target detection features may include polarization characteristics, morphological characteristics, and dynamic characteristics of an aerial intrusion target. Polarization characteristics may include degree of polarization (DOP) and polarization angle (θ). Morphological characteristics may include size, volume, surface roughness, symmetry, material, and texture. Dynamic characteristics may include velocity, acceleration, and trajectory curvature.

[0044] According to an embodiment of the present disclosure, by demodulating the electrical signal of the interference beat signal, multi-dimensional target detection features including polarization characteristics, morphological characteristics, and dynamic characteristics of the intrusion target carried in the interference beat signal are obtained.

[0045] According to an embodiment of the present disclosure, before signal demodulation, the interference beat frequency signal may be converted into an electrical signal through photoelectric conversion to facilitate subsequent signal amplification, filtering, analog-to-digital conversion, and other processing.

[0046] In operation S13 , feature matching is performed between the target detection feature and a typical target polarization feature library to obtain a feature matching result.

[0047] According to embodiments of the present disclosure, a typical target polarization feature library can be used to pre-store feature information of known typical targets. Typical targets can include drones, birds, aircraft, vegetation, etc. The feature information of typical targets can include their polarization characteristics, morphological characteristics, and dynamic characteristics.

[0048] According to an embodiment of the present disclosure, the feature matching result can, for example, represent the similarity calculation result between the aerial intrusion target and multiple typical targets, wherein the similarity calculation result can be obtained by performing similarity calculation on the target detection feature of the aerial intrusion target and the feature information of multiple typical targets in the typical target polarization feature library.

[0049] For example, during a target recognition process, the typical targets in the typical target polarization feature library include birds, drones, and balloons. By calculating the similarity of target detection features between aerial intrusion target A and these known typical targets, the feature matching results are: 83% for birds, 75% for drones, and 56% for balloons.

[0050] In operation S14, when the feature matching result indicates that the target feature similarity satisfies a first threshold condition, the target detection feature is identified based on the air intrusion target identification model to obtain a target type identification result.

[0051] According to an embodiment of the present disclosure, the first threshold condition is used to screen out potential target categories to determine whether to call the air intrusion target recognition model to identify the air intrusion target.

[0052] According to an embodiment of the present disclosure, the airborne intrusion target recognition model is used to more precisely identify the target type of the airborne intrusion target based on the target detection characteristics.

[0053] Among them, the air intrusion target recognition model can be a multi-classification model based on ensemble learning (such as random forest, XGBoost), which can establish a mapping relationship from feature vectors to target categories by learning target characteristics in a typical target feature library.

[0054] According to an embodiment of the present disclosure, when the feature matching result represents that the target feature similarity satisfies the first threshold condition, it indicates that there are types and / or features similar to the aerial intrusion target in the typical feature library. Therefore, the aerial intrusion target recognition model can be called to further identify the target detection features of the aerial intrusion target to obtain a target type recognition result.

[0055] According to embodiments of the present disclosure, the model input for an aerial intrusion target recognition model may be a target feature vector of an aerial intrusion target. The target feature vector of an aerial intrusion target may be a feature vector composed of target detection features of aerial intrusion targets that have a high similarity to typical targets, selected through feature matching with typical targets. The model output may be a target type recognition result.

[0056] In a specific embodiment, the target type recognition result may include information such as the target category, confidence score, and additional attribute tags.

[0057] Among them, the target categories obtained by identifying birds can be, for example, subcategories such as birds of prey and migratory birds, and the target categories obtained by identifying aircraft can be, for example, subcategories such as cargo aircraft, civil aircraft, and helicopters.

[0058] The confidence score can be used to quantify the reliability of the recognition result. For example, if the target category is migratory birds, the output confidence score can be 85%.

[0059] Additional attribute tags can include the target's physical familiarity or behavioral attributes. For example, they can include the target's material (such as metal, plastic, etc.), motion state (such as hovering, constant speed flight, etc.), and threat level, such as high threat (such as a fast-approaching drone), medium threat (such as a bird), or low threat (such as a kite).

[0060] According to the embodiments of the present disclosure, by fusing multi-dimensional information of polarization characteristics, morphological characteristics and dynamic characteristics, and combining feature matching with deep learning models, accurate identification of airborne intrusion targets in complex scenarios is achieved. By demodulating the interference beat signal, multi-dimensional target detection features including polarization characteristics, morphological characteristics and dynamic characteristics of the airborne intrusion target are obtained. By fusing information of the multi-dimensional target detection features, complementarity is achieved between different features, thereby improving the accuracy of subsequent target identification. In addition, the target detection features are feature matched with the typical target polarization feature library. Only when the matching result exceeds the first threshold, the target identification process of the subsequent airborne intrusion target identification model is triggered to quickly exclude obviously mismatched targets, avoid low-confidence data interfering with classification decisions, reduce the model calculation load, and solve the problem that traditional FMCW laser radars are difficult to perform fine target identification due to the lack of information reflecting the characteristics of the target surface shape, material and texture. The detection range and detection accuracy of the radar and the ability to identify targets in complex regional spaces are improved.

[0061] According to an embodiment of the present disclosure, the process of generating an interference beat signal includes: splitting the linearly polarized light output by a laser light source into intrinsic light and detection light; performing polarization orthogonal demodulation on the intrinsic light and the echo light respectively to obtain the orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light, wherein the echo light is obtained by reflecting the detection light from an invading target in the air; and coherently processing the orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light to obtain an interference beat signal.

[0062] According to the embodiments of the present disclosure, linearly polarized light can be used to represent light that vibrates only in a single plane. The linearly polarized light output by the laser light source is split into two beams, one of which is intrinsic light, which serves as a reference signal for coherent detection, and the other is detection light, which is emitted into free space to illuminate the airborne intrusion target. The detection light is irradiated by the airborne intrusion target and reflected by the airborne intrusion target to produce echo light. Different characteristics of the airborne intrusion targets can have different effects on the polarization characteristics of the detection light, causing the reflected echo light to have different changes relative to the polarization characteristics of the detection light. The polarization characteristics may include: degree of polarization, polarization angle, etc.; the characteristics of the airborne intrusion target may include: target surface roughness, material, texture, shape, etc.

[0063] According to the embodiments of the present disclosure, since the reflection of the probe light by the target changes its polarization state (e.g., polarization rotation, depolarization, etc.), the difference in the orthogonal polarization components directly reflects the polarization scattering characteristics of the target (e.g., target surface roughness, material anisotropy, etc.). Therefore, the return light can be subjected to polarization orthogonal decomposition to obtain a set of orthogonal polarization components of the return light, and the intrinsic light can be subjected to polarization orthogonal decomposition to obtain a set of orthogonal polarization components of the intrinsic light. The orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the return light are coherently superimposed, and an interference beat signal can be obtained by superimposing the two beams of light with the same frequency and polarization state. The interference beat signal contains target detection characteristics of an aerial intrusion target, such as polarization characteristics, morphological characteristics, dynamic characteristics, and other information.

[0064] In one specific embodiment, both the intrinsic light and the echo light can be decomposed into horizontally polarized TE light and vertically polarized TM light. The TE light of the intrinsic light and the TE light of the echo light are coherently interfered to obtain an interference beat signal in the TE direction. The TM light of the intrinsic light and the TM light of the echo light are coherently interfered to obtain an interference beat signal in the TM direction. Feature extraction of the interference beat signals in the TE and TM directions can be performed to obtain target detection signatures of airborne intruders carried by the echo light.

[0065] Specifically, coherent interference is performed on the TE light of the intrinsic light and the TE light of the echo light, and on the TM light of the intrinsic light and the TM light of the echo light, respectively, so as to obtain differential components of four-path interference beat signals.

[0066] Among them, the TE light of the intrinsic light and the TE light of the echo light are interfered to output two differential components of the interference beat frequency signal with a phase difference of 180° and , the output differential components are shown in the following formulas (1) and (2):

[0067] (1);

[0068] (2);

[0069] Where, TE light represents the intrinsic light, TE light represents the echo light. represents the differential positive component of the interference beat signal in the TE polarization direction, It represents the negative differential component of the interference beat signal in the TE polarization direction.

[0070] Among them, the TM light of the intrinsic light and the TM light of the echo light can also be interfered to output two differential components of the interference beat signal with a phase difference of 180°. and , the output differential components are shown in the following formulas (3) and (4):

[0071] (3);

[0072] (4);

[0073] Where, TM light represents intrinsic light, TM light represents the echo light, represents the differential positive component of the interference beat signal in the TM polarization direction, It represents the negative differential component of the interference beat signal in the TM polarization direction.

[0074] According to the embodiments of the present disclosure, by performing polarization-orthogonal demodulation on the intrinsic light and the echo light, and coherently processing the orthogonal components of the eigenlight and the echo light, the measurement capabilities of traditional FMCW radar are retained while effectively obtaining an interferometric beat frequency signal that carries the polarization characteristics of airborne intrusion targets, providing a key basis for subsequently distinguishing targets of different materials or shapes. Furthermore, by using the eigenlight and probe light split from the same light source, the eigenlight and probe light have good coherence, avoiding instabilities caused by differences in light source characteristics, such as power fluctuations and frequency drift.

[0075] According to an embodiment of the present disclosure, balanced reception processing is performed on the interference beat signal to obtain an electrical signal of the interference beat signal.

[0076] According to an embodiment of the present disclosure, after the interference beat signal is generated, the interference beat signal can be balanced received and processed. For example, a pair of interference beat signals in a differential signal state can be differentially amplified to eliminate common mode noise, thereby outputting an electrical signal with a high signal-to-noise ratio.

[0077] For example, in the process of balanced reception of the interference beat signal in the differential state in the input TE direction, the differential components of the interference beat signal are firstly and Converted into electrical signals and Then the electrical signal and Perform differential operation to obtain the electrical signal of the interference beat signal , as shown in the following formula (5):

[0078] (5);

[0079] According to the embodiments of the present disclosure, balanced reception of the interference beat signal is performed to achieve conversion of the interference beat signal from an optical signal to an electrical signal and eliminate common-mode noise, thereby ensuring that the electrical signal carrying information such as polarization, distance, and speed enters the subsequent demodulation process with a high signal-to-noise ratio, thereby avoiding feature extraction errors caused by noise.

[0080] According to an embodiment of the present disclosure, the polarization characteristics include echo polarization state parameters, the morphological characteristics of the aerial intrusion target include the target distance, and the dynamic characteristics include the target radial velocity. wherein, signal demodulation is performed on the electrical signal of the interference beat signal to obtain the target detection characteristics including: determining the target distance based on the frequency difference of the electrical signal of the interference beat signal and the sweep slope of the frequency modulated continuous wave; performing Doppler frequency shift analysis on the electrical signal of the interference beat signal to obtain the target radial velocity; and calculating the echo polarization state parameters based on the amplitude ratio of the orthogonal polarization components of the echo light.

[0081] According to an embodiment of the present disclosure, the morphological characteristics of an airborne intrusion target may include a target distance. The target distance may be used to represent the distance between the target and an FMCW (Frequency Modulated Continuous Wave) laser radar. The target distance may be obtained based on the electrical signal frequency difference of the interferometric beat signal and the sweep slope of the FMCW. For example, in an FMCW (Frequency Modulated Continuous Wave) laser radar, the target distance may be calculated using the following formula (6):

[0082] (6);

[0083] in, is the speed of light; The frequency difference is determined by mixing the echo light with the intrinsic light to obtain the frequency of the difference frequency signal; The frequency sweep slope defines the rate at which the frequency changes over time.

[0084] According to an embodiment of the present disclosure, the morphological characteristics of an aerial intrusion target also include target topographical features, which may include dimensional information such as the target's length, width, and surface area, as well as shape complexity information such as the target's surface curvature and edge sharpness. A method for determining target topographical features may include controlling the scanning angle of a laser beam to obtain distance information from the target in different directions, combining this distance information with the corresponding scanning angle to determine the position of each point on the target in three-dimensional space, reconstructing the target's three-dimensional outline based on the position of each point on the target in three-dimensional space, and extracting its topographical features.

[0085] According to an embodiment of the present disclosure, the dynamic characteristics of the airborne intrusion target include the target radial velocity, wherein the target radial velocity can be used to represent the velocity component of the airborne intrusion target in the direction of the line connecting the frequency modulated continuous wave laser radar and the airborne intrusion target.

[0086] According to an embodiment of the present disclosure, the radial velocity of a target can be obtained by performing Doppler shift analysis on the electrical signal of the interference beat signal. Specifically, the interference beat signal in the time domain can be converted to the frequency domain to obtain the signal's spectrum distribution. The interference beat signal is then filtered to remove noise and other interfering signals. In the filtered signal spectrum, the peak frequency (Doppler shift) corresponding to the Doppler shift is found. , the radial velocity of the target can be calculated by the following formula (7);

[0087] (7);

[0088] Where, is the Doppler shift, λ is the wavelength of the light source, is the radial velocity of the target.

[0089] According to embodiments of the present disclosure, the dynamic characteristics of an aerial intrusion target may also include its trajectory. This trajectory can be derived by fitting radial velocity data with time series data. The target's motion state can be determined from the trajectory. Different aerial intrusion targets may have different motion states. For example, a drone's motion state may be hovering (low velocity variance) or turning at right angles (with sudden acceleration changes). A bird's motion state may be periodic wing flapping (with a velocity fluctuation frequency of 0.5 to 5 Hz).

[0090] According to an embodiment of the present disclosure, the polarization characteristics of the airborne intrusion target include echo polarization state parameters, and the echo polarization state parameters may include polarization extinction ratio, polarization degree, and the like.

[0091] According to the embodiments of the present disclosure, the polarization state of a light wave is determined by the trajectory form and orientation of the electric field vector in a plane perpendicular to the propagation direction. The electric field vibration direction of unpolarized light is random, while the electric field vector of polarized light has a specific pattern. From the perspective of mathematical formulas, both linearly polarized light and circularly polarized light can be regarded as special cases of elliptically polarized light. In elliptically polarized light, the resultant electric field vector The direction and amplitude of will change. The endpoint trajectory will form an ellipse, and the components in the x and y directions can be written as the following formulas (8) and (9):

[0092] (8);

[0093] (9);

[0094] Where, Indicates the amplitude of the x-direction component, Indicates the amplitude of the y-direction component, represents the resultant electric field vector The angular frequency, represents the wave number, and They represent the initial phase in the x direction and the initial phase in the y direction respectively.

[0095] Eliminate the formula The influence of , we can get the polarization ellipse equation formula (10):

[0096] (10);

[0097] in, is the phase difference between the x-direction and the y-direction. From formula (10), we can see that the phase difference Amplitude ratio It can be used to represent the polarization state of elliptically polarized light.

[0098] According to an embodiment of the present disclosure, the angle between the major axis of the ellipse and the x-axis is Satisfies the following formula (11);

[0099] (11);

[0100] According to an embodiment of the present disclosure, the angle between the major axis of the ellipse and the x-axis is Determines the orientation of the ellipse, that is, the degree of inclination of the main direction of light vector vibration relative to the x-axis. Corresponding to different elliptical polarization states.

[0101] According to the embodiments of the present disclosure, the degree of polarization (P) can be used to describe the degree of polarization of light. Based on the amplitude ratio of the TE / TM polarization components of the echo light, the degree of polarization is calculated as shown in the following formula (12):

[0102] (12);

[0103] in, and are the amplitudes of the two orthogonal polarization components, Greater than .

[0104] According to embodiments of the present disclosure, the polarization extinction ratio (PER) is the ratio of the optical power of light in two mutually perpendicular polarization directions. The polarization extinction ratio (PER) reflects the anisotropic reflectivity of the target surface to polarized light. Based on the ratio of the optical power of the return light in the two mutually perpendicular polarization directions (TE / TM), the polarization extinction ratio is calculated as shown in the following formula (13):

[0105] (13);

[0106] in, Indicates the optical power of TE light, Indicates the optical power of TM light.

[0107] For example, for drones and birds, the metal surface of drones maintains good polarization reflection, resulting in a higher PER value, such as PER>20dB. Birds, on the other hand, are non-metallic targets, and their feathers have a high degree of random polarization scattering, resulting in a lower PER value, such as PER<15dB. Therefore, drones and birds can be distinguished based on the PER value.

[0108] According to embodiments of the present disclosure, polarization characteristics may also include a side mode suppression ratio (SMSR). This can be obtained by analyzing the power ratio between the main mode and side mode of the echo light. The SMSR can be used to distinguish the material of an aerial intrusion target. For example, the SMSR of a metallic drone surface is relatively high, such as SMSR > 40dB; while the SMSR of non-metallic materials, such as bird feathers and plastic interference objects, is relatively low, such as SMSR < 30dB.

[0109] Figure 2 A schematic diagram schematically illustrates typical target polarization characteristics according to an embodiment of the present disclosure.

[0110] like Figure 2 As shown in the figure, the polarization characteristics obtained for detecting four typical targets, including drones, balloons, birds, and vegetation, include the degree of polarization (PER) and side mode suppression ratio (SMSR). The polarization characteristics of the drone are: PER = 0.8, SMSR = -12.281 dB; the polarization characteristics of the balloon are: PER = 0.6, SMSR = -9.746 dB; the polarization characteristics of the bird are: PER = 0.5, SMSR = -8.580 dB; and the polarization characteristics of the vegetation are: PER = 0.2, SMSR = -5.374 dB. Based on the PER and SMSR of each target, the four typical targets can be distinguished.

[0111] According to the embodiments of the present disclosure, by demodulating the interference beat signal as an electrical signal, multi-dimensional target detection characteristics of the aerial intrusion target, including echo polarization state parameters, side mode suppression ratio, morphological characteristics, dynamic characteristics, etc., are obtained. Through comprehensive analysis of the multi-dimensional target detection characteristics, different aerial intrusion targets can be distinguished more precisely.

[0112] According to an embodiment of the present disclosure, when the feature matching result characterizing the target feature similarity does not meet the first threshold condition, the method also includes: when the feature matching result characterizing the target feature similarity meets the second threshold condition, marking the air intrusion target as an unknown target; when the feature matching result characterizing the target feature similarity does not meet the second threshold condition, marking the air intrusion target as a noise interference target.

[0113] According to an embodiment of the present disclosure, if the feature matching result characterizing that the target feature similarity does not meet the first threshold condition, it means that the airborne intrusion target is preliminarily judged to be a non-known typical target, but non-known typical targets are not all noise interference, and non-known typical targets may still be targets that need attention. For example, in an airport application scenario, both quadcopter drones and fixed-wing drones are airborne intrusion targets that may affect the flight safety of passenger aircraft. Among them, if the typical target polarization feature library only includes the target detection features of quadcopter drones, but does not include the target detection features of fixed-wing drones, when a fixed-wing drone is detected, although the specific target type of the fixed-wing drone cannot be identified, the fixed-wing drone will still pose a threat to airport safety and is an unknown target that needs attention.

[0114] Therefore, the second threshold condition is used to distinguish non-known typical targets from unknown targets or noise interference targets. If the feature matching results indicate a target feature similarity greater than or equal to the second threshold, the aerial intrusion target is marked as an unknown target. Unknown targets can be used to indicate targets that partially match known categories, such as targets with similar polarization to drones but inconsistent morphology, or targets with unusual motion patterns (e.g., trajectories that do not conform to known models). For unknown targets, their feature data can be recorded for subsequent analysis or to trigger manual verification or multi-sensor verification.

[0115] If the feature matching results indicate that the similarity between the aerial intrusion target and multiple typical targets is less than a second threshold, the aerial intrusion target is marked as a noise interference target. A noise interference target is a target that does not need attention or a false target caused by an interference signal. It can be directly discarded or recorded for rapid identification the next time the same interference target is received.

[0116] For example, when applied to airport airspace target identification, the first threshold is set at 85% and the second threshold is set at 70%. The typical target feature library consists of target detection features of quadcopters, balloons, and passenger aircraft. If the highest similarity among the feature matching results obtained by matching the airborne intrusion target with the quadcopter, balloon, or passenger aircraft is 86%, the feature matching result meets the first threshold condition, and the airborne intrusion target identification model is directly called for model identification. If the highest similarity is 80%, the feature matching result does not meet the first threshold condition, but meets the second threshold condition, and the airborne intrusion target is marked as an unknown target. If the highest similarity is 68%, the feature matching result does not meet both the first and second threshold conditions, and the airborne intrusion target is marked as a noise interference target.

[0117] Figure 3The flowchart of the target recognition method according to the embodiment of the present disclosure is schematically shown.

[0118] like Figure 3 As shown, the method includes operations S31 to S38.

[0119] In operation S31 , an interference beat signal after polarization orthogonal demodulation is acquired.

[0120] In operation S32 , the electrical signal of the interference beat signal is demodulated to obtain a target detection feature.

[0121] In operation S33 , feature matching is performed between the target detection feature and a typical target polarization feature library to obtain a feature matching result.

[0122] In operation S34, it is determined whether the feature matching result meets the first threshold condition. If yes,

[0123] Then proceed to operation S36; if not, proceed to operation S35.

[0124] In operation S36 , the target detection features are identified based on the air intrusion target identification model to obtain a target type identification result.

[0125] In operation S35, it is determined whether the feature matching result meets the second threshold condition. If not, operation S37 is performed; if yes, operation S38 is performed.

[0126] In operation S37 , the airborne intrusion target is marked as a noise interference target.

[0127] In operation S38, the airborne intrusion target is marked as an unknown target.

[0128] Based on this, the embodiment of the present disclosure introduces a first threshold and a second threshold to make two judgments on the feature matching results of the aerial intrusion target. When the feature matching result does not meet the first threshold condition, the second threshold condition is used to achieve refined classification and response to the target. The second threshold is used to filter out noise interference (such as leaves and raindrop reflections) to avoid misjudging environmental clutter as real targets, and also reduce the probability of missed detection of targets that require attention.

[0129] According to the embodiments of the present disclosure, by setting a second threshold condition, aerial intrusion targets that are determined to be non-known typical targets are secondary screened, thereby reducing the probability of missed detection of non-known typical targets that require attention, and realizing the feature collection of noise interference targets, which is conducive to quickly eliminating noise interference targets in subsequent detections.

[0130] According to an embodiment of the present disclosure, based on the target detection characteristics of the unknown target, the typical target polarization feature library is updated to obtain an updated typical target polarization feature library.

[0131] According to an embodiment of the present disclosure, after detecting that an airborne intrusion target is an unknown target and not a noise interference target, the target detection features of the detected unknown target are added to the typical target polarization feature library, thereby obtaining an updated typical target polarization feature library. For example, the target detection features obtained by target detection of unknown target A may be: polarization degree: 0.8, side mode suppression ratio: 0.7, surface roughness: 0.56, .... The target detection features of the above-mentioned unknown target A and the corresponding target type name can be added to the typical target polarization feature library to achieve an update of the typical target polarization feature library. Unknown target A can participate in the feature matching of airborne intrusion targets as a typical target.

[0132] According to embodiments of the present disclosure, the type name of an unknown target can be a simple label, such as: unknown target 1, unknown target 2, etc. The type name of an unknown target can also be an actual type name obtained by confirming the type through various methods, such as visual identification using a telescope, manual approach to the target, etc.

[0133] According to an embodiment of the present disclosure, by adding the target detection features of an unknown target to a typical target polarization feature library, the recognition speed when the same unknown target is detected again can be improved, saving computing resources.

[0134] According to an embodiment of the present disclosure, a typical target polarization feature library is obtained by the following operations: obtaining multiple typical target detection features; performing feature extraction on the multiple typical target detection features respectively to obtain a target feature vector containing polarization characteristics, morphological characteristics and dynamic characteristics; associating the target feature vector with a category label to obtain the typical target polarization feature library.

[0135] According to an embodiment of the present disclosure, when constructing a library of typical target polarization signatures, multiple typical targets must first be identified. These typical targets are aerial intrusion targets that may appear in the monitoring area, such as different types of drones, birds (e.g., pigeons, eagles), and so on. The detection signatures of multiple typical targets can be acquired by performing multi-scene, multi-angle scanning of known typical targets using a frequency-modulated continuous-wave lidar for each typical target. For example, data collection can be performed on a variety of typical aerial intrusion targets (e.g., different types of drones, birds, balloons, etc.) under different environmental conditions (e.g., sunny, cloudy, rainy, etc.), different time periods (daytime, nighttime), and different geographical regions (cities, rural areas, mountainous areas, etc.).

[0136] According to the embodiments of the present disclosure, since the acquired raw detection features may contain a large amount of redundant information, and the dimensions and numerical ranges of different features may vary greatly, feature extraction can be performed on multiple typical target detection features to extract the most representative and discriminative features from the raw detection features, and then combine them into a target feature vector for subsequent efficient feature matching and target recognition.

[0137] Methods for extracting features from multiple typical target detection features may include feature dimensionality reduction. Feature dimensionality reduction can select the most representative and discriminative features from the original target detection features, removing redundant and irrelevant features to reduce the dimensionality of the data. Feature dimensionality reduction methods include principal component analysis (PCA) and linear discriminant analysis (LDA).

[0138] According to embodiments of the present disclosure, the method for extracting features from multiple typical target detection features may further include feature combination. Feature combination may be combining different types of features to form a new feature. For example, polarization characteristics and dynamic characteristics may be combined to form a new target feature to better describe the characteristics of the target.

[0139] According to an embodiment of the present disclosure, a target feature vector may be a vector composed of multiple target detection features. Multiple target detection features may include polarization characteristics, morphological characteristics, dynamic characteristics, etc. Among them, the polarization characteristics may be polarization extinction ratio, polarization angle, side mode suppression ratio, etc. Morphological characteristics may be target morphology and size reconstructed in combination with scanning data. Dynamic features may be motion trajectories. Arranging these extracted features in a certain order constitutes a target feature vector containing polarization characteristics, morphological characteristics, and dynamic characteristics. Each target feature vector corresponds to a specific air invasion target sample, and these feature vectors will serve as input for subsequent model training. For example, a target feature vector can be expressed as F=[PER, SMSR, size, velocity variance, morphology complexity, …].

[0140] According to an embodiment of the present disclosure, the target detection features in the target feature vector can be target detection features directly obtained based on the echo light interference beat frequency electrical signal, such as polarization degree, extinction ratio, speed, etc.; or they can be target detection features indirectly derived based on single or multiple directly obtained target detection features, such as surface roughness, material, etc.

[0141] According to an embodiment of the present disclosure, a unique category label is assigned to each typical target to identify the type of target. For example, a multi-rotor drone is marked as "category 1", a fixed-wing drone is marked as "category 2", a pigeon is marked as "category 3", and so on. The target feature vector of each typical target is associated with the corresponding category label to form a labeled sample library, that is, a typical target polarization feature library. In the subsequent target recognition process, after the detection features of the airborne intrusion target are obtained and the target feature vector is extracted, the vector can be matched with the feature vector in the typical target polarization feature library, and the most similar typical target and its corresponding category label are found according to the matching results, thereby realizing the type recognition of the airborne intrusion target.

[0142] Based on this, the embodiments of the present disclosure utilize multi-dimensional feature fusion to form a library of typical target polarization signatures, avoiding the ambiguity of single features. In complex environments (such as strong light, haze, and multi-target scenes), the multi-dimensional feature combination within the library provides stable target representation and enhances system robustness. Furthermore, this library, constructed by collecting a large amount of detection data from typical targets offline, provides a template library for online recognition. This allows the system to achieve rapid classification by matching pre-stored typical features without relying on real-time learning.

[0143] According to an embodiment of the present disclosure, an airborne intrusion target recognition model is obtained by the following operations: a typical target polarization feature library is used to optimize and train a multi-classification model based on an ensemble learning algorithm to obtain an airborne intrusion target recognition model.

[0144] Figure 4 The following schematically shows a training flow chart of an airborne intrusion target model according to an embodiment of the present disclosure.

[0145] like Figure 4 As shown, the training process includes operations S401 to S411.

[0146] In operation S401 , a plurality of typical target detection features are acquired.

[0147] In operation S402 , feature extraction is performed on a plurality of typical target detection features to obtain target feature vectors including polarization characteristics, morphological characteristics, and dynamic characteristics.

[0148] In operation S403 , the target feature vector is associated with the category label to obtain a typical target polarization feature library.

[0149] In operation S404 , the typical target polarization feature library is divided into a training set, a validation set, and a test set.

[0150] According to embodiments of the present disclosure, the typical target polarization signature library can be divided into a specific proportion, for example, 70% training set, 15% validation set, and 15% test set. The training set is used to train the model, the validation set is used to adjust the model's hyperparameters, and the test set is used to evaluate the model's final performance.

[0151] In operation S405 , an initial model is selected and the model is initialized.

[0152] According to embodiments of the present disclosure, the initial model may select a multi-classification model of an ensemble learning algorithm. For example, a random forest, gradient boosted tree, or stacking model. Model initialization may include initializing the multi-classification model based on the selected ensemble learning algorithm and setting some initial hyperparameters. For example, for a random forest, hyperparameters such as the number of decision trees, maximum depth, and minimum number of sample splits need to be set.

[0153] In operation S406 , the selected model is trained using the training set data.

[0154] According to an embodiment of the present disclosure, training the selected model may include inputting the target feature vectors and corresponding class labels of the training set into an initialized model for training. The model learns the mapping relationship between features and classes based on the training data and continuously adjusts the model parameters to minimize prediction error.

[0155] In operation S407 , the trained model is evaluated on the validation set to observe the performance indicators of the model.

[0156] According to embodiments of the present disclosure, evaluating a trained model on a validation set and observing the model's performance metrics involves inputting feature vectors from the validation set into the trained model and comparing the class labels corresponding to the feature vectors with the predictions output by the model to obtain the model's performance metrics. Model performance metrics may include accuracy, precision, recall, and F1 value.

[0157] According to the embodiments of the present disclosure, the F1 value combines precision and recall and is the harmonic mean of the two. The higher the F1 value, the better the balance between precision and recall of the model. The F1 value can be calculated using the following formula (14):

[0158] (14);

[0159] Where, represents the accuracy, Represents the recall rate.

[0160] In operation S408 , it is determined whether the performance index of the model meets the performance requirements.

[0161] According to the embodiments of the present disclosure, the preset model performance indicators can be compared with the actually generated model performance indicators to determine whether the performance indicators of the model meet the requirements.

[0162] If the answer is no, then operation S409 is executed to optimize the hyperparameters of the model. If the answer is yes, then operation S410 is executed.

[0163] According to an embodiment of the present disclosure, the validation set can be used to optimize the hyperparameters of the model. For example, a grid search, random search, or Bayesian optimization method can be used to find the optimal hyperparameter combination within the range of hyperparameter values, reset the hyperparameters of the model, and return to operation S406 to retrain the model.

[0164] In operation S410 , the optimized model is evaluated using a test set and the model is output.

[0165] According to the embodiments of the present disclosure, the optimized model is evaluated using a test set, and common evaluation metrics such as accuracy, precision, recall, F1 value, etc. are calculated. Based on the evaluation results, the actual performance of the model on unknown data is analyzed.

[0166] In operation S411 , the model is deployed.

[0167] According to an embodiment of the present disclosure, a trained model that meets performance requirements is saved and deployed in an actual recognition system.

[0168] Based on this, the initial model is trained using a typical target polarization feature library that includes typical target polarization characteristics, morphological characteristics, and dynamic characteristics. The obtained aerial intrusion target recognition model has the ability to perform fine target type recognition based on the multi-dimensional target detection characteristics of aerial intrusion targets.

[0169] An embodiment of the present disclosure also provides a laser radar system based on polarization orthogonal demodulation, comprising: a laser light source, an optical beam splitter, a first polarization beam splitter, a second polarization beam splitter, an optical amplifier, an optical circulator, a quarter-wave plate, an optical transceiver system, a first coupler, a second coupler, a balanced detector, a field programmable gate array, and a host computer, wherein the input end of the optical beam splitter is connected to the laser light source, and the output end of the optical beam splitter is respectively connected to the first polarization beam splitter and the optical amplifier; the output end of the first polarization beam splitter is respectively connected to the first coupler and the second coupler; the output end of the optical amplifier is connected to the optical circulator; the output end of the optical circulator is respectively connected to the second polarization beam splitter and the quarter-wave plate, and the output end of the quarter-wave plate is connected to the optical transceiver system; the output end of the second polarization beam splitter is respectively connected to the first coupler and the second coupler; the output end of the first coupler is connected to the first balanced detector; the output end of the second coupler is connected to the second balanced detector; the output ends of the first balanced detector and the second balanced detector are both connected to the field programmable gate array; and the output end of the field programmable gate array is connected to the host computer.

[0170] Figure 5 A schematic diagram of a laser radar system based on polarization orthogonal demodulation according to an embodiment of the present disclosure is schematically shown.

[0171] like Figure 5As shown, linearly polarized light emitted by a laser light source 501 is transmitted to an optical beam splitter 502. Optical beam splitter 502 is used to split the linearly polarized light into intrinsic light and probe light. The intrinsic light is then transmitted to a first polarization beam splitter 507, which splits the intrinsic light into two intrinsic light polarization components with orthogonal polarization directions. These components can be represented as TE light and TM light. Optical beam splitter 502 transmits the probe light to an optical amplifier 503. The probe light passes through an optical circulator 504, a quarter-wave plate 505, and an optical transceiver system 506 before irradiating a target. The target reflects the probe light, generating echo light. This echo light is received by the optical transceiver system 506 and passes through the quarter-wave plate 505 to the optical circulator 504. The optical circulator 504 transmits the echo light to a second polarization beam splitter 508. The second polarization beam splitter 508 splits the echo light into two echo light polarization components with orthogonal polarization directions. For example, it can be represented as TE light and TM light. The first coupler 509 receives a set of intrinsic light polarization components and echo light polarization components with the same polarization direction sent by the first polarization beam splitter 507 and the second polarization beam splitter 508. For example, it receives TE light of the intrinsic light and TE light of the echo light. The second coupler 510 receives another set of intrinsic light polarization components and echo light polarization components with the same polarization direction sent by the first polarization beam splitter 507 and the second polarization beam splitter 508. For example, it receives TM light of the intrinsic light and TM light of the echo light. The first coupler 509 outputs an interference beat signal (in the form of a differential signal) with a polarization direction, such as the TE direction, which is received by the first balanced detector 511; the first balanced detector 511 transmits the interference beat signal, which has undergone balanced reception and photoelectric conversion, to the field programmable gate array 513. The second coupler 510 outputs an interference beat signal (in the form of a differential signal) in another polarization direction, such as the TM direction, which is received by the second balanced detector 512. The first balanced detector 512 transmits the balanced received and photoelectrically converted interference beat signal to the field programmable gate array 513. The field programmable gate array 513 receives the interference beat signals in two orthogonal polarization directions, performs calculations on them, and transmits the calculation results to the host computer 514 for further processing.

[0172] According to an embodiment of the present disclosure, an optical beam splitter is used to split the linearly polarized light emitted by a laser light source into a detection light and an intrinsic light; a first polarization beam splitter is used to perform polarization orthogonal demodulation on the intrinsic light to obtain an orthogonal polarization component of the intrinsic light; an optical amplifier is used to power amplify the detection light to obtain amplified detection light; a 1 / 4 wave plate is used to adjust the polarization state of the amplified detection light to obtain a scanning detection light; an optical transceiver system is used to transmit the scanning detection light to a target area and receive an echo light reflected by the target area; a second polarization beam splitter is used to perform polarization orthogonal demodulation on the echo light to obtain an orthogonal polarization component of the echo light; a first coupler is used to perform coherent interference on the first orthogonal polarization component of the intrinsic light and the first orthogonal polarization component of the echo light to generate a first interference beat signal ; The second coupler is used to perform coherent interference on the second orthogonal polarization component of the intrinsic light and the second orthogonal polarization component of the echo light to generate a second interference beat signal; the first balanced detector and the second balanced detector are respectively used to perform balanced reception and conversion processing on the first interference beat signal and the second interference beat signal to obtain an electrical signal of the interference beat signal; the field programmable gate array is used to perform signal demodulation on the electrical signal of the interference beat signal to obtain a target detection feature; the host computer is used to perform feature matching on the target detection feature with a typical target polarization feature library to obtain a feature matching result. When the feature matching result indicates that the target feature similarity meets the first threshold condition, the target detection feature is identified and processed based on the air intrusion target recognition model to obtain a target type recognition result.

[0173] Figure 6 The working flow chart of the laser radar system of an embodiment of the present disclosure is schematically shown.

[0174] like Figure 6 As shown, the workflow of the laser radar system of the embodiment of the present disclosure includes steps S601 to S611.

[0175] In step S601, the linearly polarized light emitted by the laser light source is split into probe light and intrinsic light.

[0176] According to the embodiments of the present disclosure, the linearly polarized light emitted by the laser light source can be split into high-power detection light and low-power intrinsic light by a beam splitter, thereby using most of the energy of the linearly polarized light emitted by the laser light source for target detection and improving the detection distance.

[0177] In step S602 , polarization orthogonal demodulation is performed on the intrinsic light to obtain orthogonal polarization components of the intrinsic light.

[0178] According to the embodiments of the present disclosure, to demodulate the polarization characteristics of an airborne intrusion target carried in the return light, it is necessary to generate a beat frequency signal by coherently interfering with the orthogonal polarization components of the return light. Polarization orthogonal demodulation is performed on the intrinsic light to obtain the orthogonal polarization components of the intrinsic light, which are then coherently interfering with the orthogonal polarization components of the return light. A first polarization beam splitter can be used to split the intrinsic light into a TE polarization component and a TM polarization component, both of which have orthogonal polarizations.

[0179] In step S603, the detection light is power amplified to obtain amplified detection light.

[0180] According to an embodiment of the present disclosure, an optical amplifier may be used to amplify the detection light to increase the detection light energy, so as to improve the detection distance and detection accuracy.

[0181] In step S604, the polarization state of the amplified detection light is adjusted to obtain scanning detection light.

[0182] According to embodiments of the present disclosure, a quarter-wave plate can be used to adjust the polarization state of the amplified probe light, converting linearly polarized probe light into circular polarization. Linearly polarized light is unstable and easily changes when encountering a target. Adjusting it to circularly polarized light through a quarter-wave plate reduces the likelihood of polarization changes.

[0183] In step S605 , scanning probe light is emitted to the target area, and echo light reflected by the target area is received.

[0184] According to an embodiment of the present disclosure, the scanning detection light can be transmitted to the target area through an optical transceiver system, and the echo light reflected by the target area is received, wherein the echo light carries the target detection characteristics of the airborne intrusion target.

[0185] In step S606, polarization orthogonal demodulation is performed on the echo light to obtain orthogonal polarization components of the echo light.

[0186] According to an embodiment of the present disclosure, the echo light reaches the optical circulator via an optical path opposite to the scanning detection light, and is transmitted to the second polarization beam splitter through the circulator to split the intrinsic light into a TE polarization component and a TM polarization component with orthogonal polarizations.

[0187] In step S607 , coherent interference is performed on the orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light to generate first and second interference beat signals.

[0188] According to an embodiment of the present disclosure, by interfering the polarization components of the same polarization direction of the eigenlight and the echo light, first and second interference beat signals carrying target detection characteristics including polarization characteristics of the aerial intrusion target are generated.

[0189] In step S608, balanced reception and conversion processing are performed on the first interference beat signal and the second interference beat signal to obtain electrical signals of the interference beat signals.

[0190] According to the embodiment of the present disclosure, by performing photoelectric conversion and differential amplification on the first and second interference beat frequency signals, electrical signals of the first and second interference beat frequency signals with common mode interference eliminated are obtained, so as to facilitate the next step of target detection feature extraction.

[0191] In step S609, the electrical signal of the interference beat signal is demodulated to obtain the target detection feature.

[0192] According to an embodiment of the present disclosure, target detection characteristics can be obtained by processing the electrical signals of the first and second interference beat signals through a field programmable gate array.

[0193] In step S610, feature matching is performed between the target detection feature and the typical target polarization feature library to obtain a feature matching result.

[0194] According to the embodiments of the present disclosure, the target detection features obtained by the field programmable gate array can be matched with the typical target feature information in the typical target polarization feature library to preliminarily and quickly screen the similarity of the air intrusion targets.

[0195] In step S611, when the feature matching result indicates that the target feature similarity satisfies a first threshold condition, the target detection feature is identified based on the air intrusion target identification model to obtain a target type identification result.

[0196] According to an embodiment of the present disclosure, when the similarity between an aerial intrusion target and a typical target exceeds a first threshold, it is considered that the aerial intrusion target may be one of the typical targets. In this case, the aerial intrusion target recognition model is called to perform fine recognition of the aerial intrusion target and identify the target category.

[0197] Based on this, the embodiments of the present disclosure use a beam splitter to split the linearly polarized light emitted by the laser light source into detection light and intrinsic light, thereby ensuring a high correlation between the intrinsic light, the detection light and the echo light, and improving the detection sensitivity of aerial intrusion targets; through the mutual cooperation of the polarization beam splitter, the coupler and the balanced receiver, the interference beat signal of the echo is obtained and the common mode interference is weakened; the extraction of target detection features and the preliminary screening of targets are realized through the field programmable gate array; and the model recognition and recognition result output of aerial intrusion targets are realized through the host computer.

[0198] Figure 7 A block diagram of a frequency modulated continuous wave laser radar target recognition device according to an embodiment of the present disclosure is schematically shown.

[0199] like Figure 7 As shown, the FMCW laser radar target recognition device 700 includes an interference beat signal acquisition module 710 , a target detection feature acquisition module 720 , a feature matching module 730 , and a target type recognition module 740 .

[0200] The interference beat signal acquisition module 710 is used to acquire the interference beat signal after polarization orthogonal demodulation.

[0201] The target detection feature acquisition module 720 is used to demodulate the electrical signal of the interference beat signal to obtain the target detection feature.

[0202] The feature matching module 730 is used to perform feature matching on the target detection features and the typical target polarization feature library to obtain a feature matching result.

[0203] The target type identification module 740 is used to identify the target detection features based on the air intrusion target identification model to obtain a target type identification result when the feature matching result indicates that the target feature similarity meets the first threshold condition.

[0204] According to an embodiment of the present disclosure, the interference beat signal acquisition module 710 may include a beam splitting unit, a polarization decomposition unit, and a beat signal acquisition unit.

[0205] The beam splitting unit is used to split the linearly polarized light output by the laser light source into intrinsic light and detection light.

[0206] The polarization decomposition unit is used to perform polarization orthogonal demodulation on the intrinsic light and the echo light respectively to obtain the orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light.

[0207] The beat signal acquisition unit is used to perform coherent processing on the orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light to obtain an interference beat signal.

[0208] According to an embodiment of the present disclosure, the interference beat signal acquisition module 710 may further include a balanced receiving unit.

[0209] The balanced receiving unit is used to perform balanced receiving processing on the interference beat frequency signal to obtain an electrical signal of the interference beat frequency signal.

[0210] According to an embodiment of the present disclosure, the target detection feature acquisition module 720 may include a target distance determination unit, a target radial velocity determination unit, and an echo polarization state parameter determination unit.

[0211] The target distance determination unit is used to determine the target distance based on the electrical signal frequency difference of the interference beat signal and the sweep slope of the frequency modulated continuous wave.

[0212] The target radial velocity determination unit is used to perform Doppler frequency shift analysis on the electrical signal of the interference beat frequency signal to obtain the target radial velocity.

[0213] The echo polarization state parameter determination unit is used to calculate the echo polarization state parameter based on the amplitude ratio of the orthogonal polarization components of the echo light.

[0214] According to an embodiment of the present disclosure, the FMCW laser radar target recognition device 700 may further include an unknown target screening module.

[0215] The unknown target screening module is used to mark the air intrusion target as an unknown target when the feature matching result indicates that the target feature similarity meets the second threshold condition.

[0216] The noise interference identification module is used to mark the air intrusion target as a noise interference target when the feature matching result indicates that the target feature similarity does not meet the second threshold condition.

[0217] According to an embodiment of the present disclosure, the FMCW laser radar target recognition device 700 may further include a typical target polarization feature library update module.

[0218] The typical target polarization feature library updating module is used to update the typical target polarization feature library based on the target detection characteristics of the unknown target to obtain an updated typical target polarization feature library.

[0219] According to an embodiment of the present disclosure, the FMCW laser radar target recognition device 700 may further include a typical target polarization feature library generation module, wherein the typical target polarization feature library generation module may include a feature vector generation unit and a feature library generation unit.

[0220] The feature vector generating unit is used to obtain multiple typical target detection features; and extract features from the multiple typical target detection features to obtain a target feature vector including polarization characteristics, morphological characteristics and dynamic characteristics.

[0221] The feature library generation unit is used to associate the target feature vector with the category label to obtain a typical target polarization feature library.

[0222] According to an embodiment of the present disclosure, the FMCW laser radar target recognition device 700 may further include an air intrusion target recognition model generation module.

[0223] The airborne intrusion target recognition model generation module is used to optimize and train the multi-classification model based on the ensemble learning algorithm using the typical target polarization feature library to obtain the airborne intrusion target recognition model.

[0224] According to the embodiments of the present invention, any number of modules, sub-modules, units, and sub-units, or at least part of the functions of any number of them, can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, sub-modules, units, and sub-units can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, sub-modules, units, and sub-units can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, sub-modules, units, and sub-units can be at least partially implemented as a computer program module, which can perform the corresponding functions when the computer program module is executed.

[0225] For example, any number of the interference beat signal acquisition module 710, the target detection feature acquisition module 720, the feature matching module 730, and the target type identification module 740 can be combined into a single module / unit / sub-unit, or any one of these modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functionality of one or more of these modules / units / sub-units can be combined with at least part of the functionality of other modules / units / sub-units and implemented in a single module / unit / sub-unit. According to an embodiment of the present disclosure, at least one of the interference beat signal acquisition module 710, the target detection feature acquisition module 720, the feature matching module 730, and the target type identification module 740 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented in hardware or firmware by any other reasonable means of circuit integration or packaging, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or any appropriate combination of any of these. Alternatively, at least one of the interference beat signal acquisition module 710, the target detection feature acquisition module 720, the feature matching module 730, and the target type identification module 740 can be at least partially implemented as a computer program module, which can perform corresponding functions when executed.

[0226] It should be noted that the frequency modulated continuous wave laser radar target recognition device part in the embodiment of the present disclosure corresponds to the target recognition method part based on the frequency modulated continuous wave laser radar in the embodiment of the present disclosure. The description of the frequency modulated continuous wave laser radar target recognition device part specifically refers to the target recognition method part based on the frequency modulated continuous wave laser radar, which will not be repeated here.

[0227] Figure 8 A block diagram of an electronic device suitable for implementing a target recognition method based on a frequency modulated continuous wave laser radar according to an embodiment of the present disclosure is schematically shown.

[0228] Figure 8 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0229] like Figure 8 As shown, the electronic device 800 according to an embodiment of the present disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage unit 808 into a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0230] Various programs and data required for the operation of the electronic device 800 are stored in the RAM 803. The processor 801, ROM 802, and RAM 803 are connected to each other via a bus 804. The processor 801 executes the various operations of the method flow according to the embodiment of the present disclosure by executing the programs in the ROM 802 and / or RAM 803. It should be noted that the programs may also be stored in one or more memories other than the ROM 802 and RAM 803. The processor 801 may also execute the various operations of the method flow according to the embodiment of the present disclosure by executing the programs stored in the one or more memories.

[0231] According to an embodiment of the present disclosure, electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to bus 804. Electronic device 800 may also include one or more of the following components connected to I / O interface 805: an input section 806 including a keyboard, mouse, etc.; an output section 807 including devices such as a cathode ray tube (CRT), liquid crystal display (LCD), and speakers; a storage section 808 including a hard disk; and a communication section 809 including a network interface card such as a LAN card or modem. Communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to I / O interface 805 as needed. Removable media 811, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed in drive 810 as needed, so that computer programs read from the removable media can be installed into storage section 808 as needed.

[0232] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the processor 801, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, device, apparatus, module, unit, etc. described above can be implemented by a computer program module.

[0233] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, and when executed, implements the method according to the embodiments of the present disclosure.

[0234] According to embodiments of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium. Examples include, but are not limited to, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0235] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 802 and / or the RAM 803 described above and / or one or more memories other than the ROM 802 and the RAM 803 .

[0236] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present disclosure. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the target recognition method based on frequency modulated continuous wave laser radar provided by the embodiment of the present disclosure.

[0237] When the computer program is executed by the processor 801, the above functions defined in the system / device of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0238] In one embodiment, the computer program may be stored on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may be transmitted and distributed in the form of a signal on a network medium, downloaded and installed via the communication portion 809, and / or installed from a removable medium 811. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0239] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include, but are not limited to, languages such as Java, C++, Python, "C" or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect via the Internet).

[0240] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of the systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, which contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the boxes can also occur in an order different from that marked in the accompanying drawings. For example, two boxes shown in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, as well as the combination of boxes in the block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified functions or operations, or can be implemented using a combination of dedicated hardware and computer instructions. It will be understood by those skilled in the art that the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present disclosure. In particular, without departing from the spirit and teachings of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or coupled in various ways, and all of these combinations and / or couplings fall within the scope of the present disclosure.

[0241] The above describes the embodiments of the present disclosure. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be advantageously used in combination. Without departing from the scope of the present disclosure, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A target recognition method based on frequency modulated continuous wave laser radar, wherein: The method comprises: Acquire an interference beat signal after polarization orthogonal demodulation, wherein the interference beat signal is coherently generated by orthogonal polarization components of the intrinsic light and the echo light; Performing signal demodulation on the electrical signal of the interference beat signal to obtain target detection characteristics, wherein the target detection characteristics include polarization characteristics, morphological characteristics, and dynamic characteristics of the aerial intrusion target; Performing feature matching on the target detection feature and a typical target polarization feature library to obtain a feature matching result; and When the feature matching result indicates that the target feature similarity satisfies a first threshold condition, the target detection feature is identified based on an airborne intrusion target identification model to obtain a target type identification result.

2. The method according to claim 1, wherein The generation process of the interference beat signal includes: Splitting the linearly polarized light output by the laser light source into the intrinsic light and the probe light; and performing polarization orthogonal demodulation on the intrinsic light and the echo light respectively to obtain an orthogonal polarization component of the intrinsic light and an orthogonal polarization component of the echo light, wherein the echo light is obtained by reflecting the detection light from the aerial intrusion target; The orthogonal polarization components of the intrinsic light and the orthogonal polarization components of the echo light are coherently processed to obtain the interference beat signal.

3. The method according to claim 2, wherein: The method further comprises: Balanced reception processing is performed on the interference beat signal to obtain an electrical signal of the interference beat signal.

4. The method according to claim 3, wherein: The polarization characteristics include echo polarization state parameters, the morphological characteristics of the aerial intrusion target include target distance, and the dynamic characteristics include target radial velocity. The signal demodulation of the electrical signal of the interference beat frequency signal to obtain the target detection characteristics includes: Determining the target distance based on the electrical signal frequency difference of the interference beat signal and the sweep slope of the frequency modulated continuous wave; Performing Doppler frequency shift analysis on the electrical signal of the interference beat signal to obtain the target radial velocity; and The echo polarization state parameter is calculated based on the amplitude ratio of the orthogonal polarization components of the echo light.

5. The method according to claim 1, wherein When the feature matching result indicates that the target feature similarity does not meet a first threshold condition, the method further includes: If the feature matching result indicates that the target feature similarity satisfies a second threshold condition, marking the airborne intrusion target as an unknown target; When the feature matching result indicates that the target feature similarity does not meet a second threshold condition, the aerial intrusion target is marked as a noise interference target.

6. The method according to claim 5, wherein: The method further comprises: Based on the target detection characteristics of the unknown target, the typical target polarization characteristic library is updated to obtain an updated typical target polarization characteristic library.

7. The method according to claim 1, wherein The typical target polarization feature library is obtained by the following operations: acquiring a plurality of typical target detection features; Extracting features of the plurality of typical target detection features to obtain target feature vectors including polarization characteristics, morphological characteristics, and dynamic characteristics; as well as The target feature vector is associated with the category label to obtain the typical target polarization feature library.

8. The method according to claim 7, wherein: The airborne intrusion target recognition model is obtained by the following operations: The typical target polarization feature library is used to optimize and train a multi-classification model based on an ensemble learning algorithm to obtain the aerial intrusion target recognition model.

9. A laser radar system based on polarization orthogonal demodulation, applicable to the method according to any one of claims 1 to 8, wherein: The laser radar system includes: a laser light source, an optical beam splitter, a first polarization beam splitter, a second polarization beam splitter, an optical amplifier, an optical circulator, a quarter wave plate, an optical transceiver system, a first coupler, a second coupler, a balanced detector, a field programmable gate array, and a host computer, wherein: The input end of the optical beam splitter is connected to the laser light source, and the output end of the optical beam splitter is connected to the first polarization beam splitter and the optical amplifier respectively; The output end of the first polarization beam splitter is connected to the first coupler and the second coupler respectively; The output end of the optical amplifier is connected to the optical circulator; The output end of the optical circulator is connected to the second polarization beam splitter and the quarter wave plate respectively, and the output end of the quarter wave plate is connected to the optical transceiver system; The output end of the second polarization beam splitter is connected to the first coupler and the second coupler respectively; The output end of the first coupler is connected to a first balanced detector; The output end of the second coupler is connected to the second balanced detector; The output terminals of the first balanced detector and the second balanced detector are both connected to the field programmable gate array; and The output end of the field programmable gate array is connected to the host computer.

10. The system according to claim 9, wherein: The optical beam splitter is used to split the linearly polarized light emitted by the laser light source into probe light and intrinsic light; The first polarization beam splitter is used to perform polarization orthogonal demodulation on the intrinsic light to obtain orthogonal polarization components of the intrinsic light; The optical amplifier is used to amplify the power of the detection light to obtain amplified detection light; The quarter wave plate is used to adjust the polarization state of the amplified detection light to obtain scanning detection light; The optical transceiver system is used to transmit the scanning detection light to the target area and receive the echo light reflected by the target area; The second polarization beam splitter is used to perform polarization orthogonal demodulation on the echo light to obtain orthogonal polarization components of the echo light; The first coupler is used to perform coherent interference on the first orthogonal polarization component of the intrinsic light and the first orthogonal polarization component of the echo light to generate a first interference beat signal; The second coupler is used to perform coherent interference on the second orthogonal polarization component of the intrinsic light and the second orthogonal polarization component of the echo light to generate a second interference beat signal; The first balanced detector and the second balanced detector are used to perform balanced reception and conversion processing on the first interference beat signal and the second interference beat signal respectively to obtain electrical signals of the interference beat signals; The field programmable gate array is used to demodulate the electrical signal of the interference beat signal to obtain a target detection feature; and The host computer is used to perform feature matching on the target detection feature with a typical target polarization feature library to obtain a feature matching result. When the feature matching result indicates that the target feature similarity satisfies a first threshold condition, the target detection feature is identified and processed based on an aerial intrusion target recognition model to obtain a target type recognition result.

Images