Day and night detection method and system based on laser transparent window

Abstract

The invention discloses a day and night detection method and system based on a laser transparent window, and relates to the technical field of optical detection, and the method comprises the following steps: 1, transmitting a laser beam through a laser radar, transmitting the laser beam through the transparent window, and building an exponential decay model of the laser beam to quantitatively describe the influence of the transparent window; according to the method, the exponential decay model of the laser beam is established, the influence of the transparent window on the laser beam is quantitatively described, the actual wavefront phase is monitored in real time through the radar wavefront sensor and compared with the ideal wavefront, wavefront distortion can be rapidly recognized, beam distortion caused by wavefront distortion is reduced, and the accuracy of wavefront distortion detection is improved. A self-adaptive optical correction model is introduced, so that the system can dynamically adjust a deformable mirror according to wavefront distortion monitored in real time, the wavefront distortion of a laser beam is corrected, and efficient target detection and classification can be realized in a complex environment by effectively mapping an intensity label of a reflected beam and a category label of target characteristics.

Description

Technical Field

[0001] The present invention relates to the field of optical detection technology, and in particular to a day and night detection method and system based on a laser window. Background Art

[0002] Day and night detection technology refers to the technology that can continuously and effectively detect, monitor and identify targets under different day and night conditions. With the acceleration of globalization and the continuous advancement of science and technology, society's demand for all-weather detection technology has become more urgent. The acceleration of urbanization and the advancement of intelligent society require that monitoring and detection need to have higher detection efficiency and intelligence level. With global warming and the frequent occurrence of extreme weather events, relying on a variety of detection methods to conduct all-weather monitoring of weather changes and natural disasters has become the norm.

[0003] The existing technology has the following deficiencies: the existing day and night detection technology relies on different types of sensors for data capture, such as infrared sensors and visible light cameras. The difference in lighting conditions during the day and at night has a significant impact on the image quality and sensor performance, resulting in a significant decrease in detection capabilities. In complex environments, the accuracy and effectiveness of the sensor may be disturbed. When there are obstructions, complex terrain or dynamic changes in the environment, it is difficult to ensure sufficient accuracy and robustness. LiDAR is easily affected by the wavefront distortion of the medium in the target environment, which in turn affects the pointing accuracy of the laser beam and reduces the detection effect. The performance under day and night lighting conditions is poor, affecting the reflection signal of the laser beam, thereby affecting the detection accuracy. In a dynamic environment, it is often unable to quickly adapt to changing lighting, targets or environments.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a day and night detection method and system based on a laser window to solve the problems in the above-mentioned background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a day and night detection method based on a laser window, comprising the following steps:

[0007] Step 1: Use the laser radar to emit a laser beam and propagate it through the transparent window, and establish an exponential attenuation model of the laser beam to quantitatively describe the influence of the transparent window;

[0008] Step 2: Monitor the actual wavefront phase in real time through the radar wavefront sensor, define the ideal wavefront and calculate the wavefront distortion of the laser beam, and quantify the degree of distortion according to the root mean square error;

[0009] Step 3: Using an adaptive optical correction model to dynamically adjust the deformable mirror to correct the wavefront distortion of the laser beam according to the real-time wavefront distortion;

[0010] Step 4: Record the round-trip propagation time of the laser beam under day and night conditions to calculate the detection distance, and use the support vector machine classification algorithm to construct the optimal hyperplane in the high-dimensional space to classify the target characteristics and construct a three-dimensional model of the target;

[0011] Preferably, a laser beam is generated by selecting a pulse mode through a laser radar and emitted to the direction of the target area under day and night conditions. The laser beam interacts with molecules or particles through a transparent window, resulting in attenuation of the laser beam. An exponential attenuation model is established based on the attenuation coefficient to quantitatively describe the influence of the transparent window. The specific formula is:

[0012] I(d)=I0e -αd

[0013] Wherein, I(d) represents the intensity of the laser beam at a distance d after propagating through the window, I0 represents the initial intensity of the laser beam emitted in the direction of the target area after propagating through the window, and α represents the attenuation coefficient of propagation through the window, wherein the attenuation coefficient reflects the attenuation degree of the intensity of the laser beam as the propagation distance increases after propagating through the window, and the attenuation factors of the laser beam in the process of propagation through the window are divided into absorption influence and scattering influence. After the laser beam propagates through the window and reaches the target area under day and night conditions, it is reflected back to the receiving end of the lidar, and the reflected beam is detected by the photoelectric detector at the receiving end.

[0014] Preferably, the actual wavefront phase caused by the medium environment of the target area under day and night conditions for the propagation of the laser beam is monitored in real time by a radar wavefront sensor, and compared with the ideal wavefront phase, the ideal wavefront is defined to represent the wavefront when the laser beam is emitted from the laser radar, and the wavefront distortion of the laser beam is calculated, and the specific formula is:

[0015] Δφ(r)=φ m (r)-φ0(r)

[0016] Where Δφ(r) represents the wavefront distortion, φ m(r) represents the actual wavefront phase, φ0(r) represents the ideal wavefront phase, r represents the spatial position of the transmitted wavefront distortion, and the distortion mode is checked. The phase diagram and amplitude diagram of the wavefront are drawn according to the wavefront distortion. The wavefront distortion is defined as a phase jump according to the instantaneous change of the phase value of the laser beam at a certain point in the wavefront. By performing differential analysis on the phase diagram of the wavefront, it is observed whether there is a discontinuous phase value. The wavefront distortion is defined as a focal shift according to the change in the focal position when the laser beam is emitted to the target area. It is identified by the change in the focal position of the coherent interference pattern. The wavefront data is analyzed by Fourier transform to check whether there is a periodic pattern of a specific frequency. The root mean square error of the wavefront distortion is calculated to quantify the degree of distortion. The specific formula is:

[0017]

[0018] Among them, RMS represents the root mean square error of wavefront distortion, represents the measure of the wavefront distortion quantification result, N represents the total number of points where the radar wavefront sensor monitors the actual wavefront phase in real time, Δφ(r i ) represents the number of points corresponding to the i-th actual wavefront phase monitored by the radar wavefront sensor in real time r i The wavefront distortion at the position, where the smaller the RMS, the smaller the wavefront distortion, the higher the wavefront quality, and the enhanced propagation and focusing capabilities of the laser beam. The larger the RMS, the greater the wavefront distortion, resulting in a decrease in the quality of the laser beam during propagation.

[0019] Preferably, the core component of the adaptive optics correction model is defined as a deformable mirror, and the deformable mirror is used to control the displacement of each point on the surface to achieve correction of wavefront distortion. In the adaptive optics correction model, there is a physical relationship that the shape of the deformable mirror produces reverse wavefront distortion to deduce the adjustment amount of the deformable mirror, and an objective function is defined to make the wavefront after adaptive optics correction consistent with the ideal wavefront by adjusting the shape of the deformable mirror. The specific formula is:

[0020]

[0021] Where Cost represents the error after wavefront adaptive optics correction, D(r) represents the control signal of each point on the deformable mirror surface, It represents the mapping process from the deformable mirror shape to the wavefront. The gradient of the calculated objective function represents the sensitivity to the deformable mirror shape, which is used to adjust each point of the deformable mirror to reduce and minimize the error after wavefront adaptive optics correction. The specific formula is:

[0022]

[0023] in, Represents the gradient of the objective function, and updates the shape of the deformable mirror by gradient descent. The specific formula is:

[0024]

[0025] Among them, D new (r) represents the updated shape of the deformable mirror, D old (r) represents the initial shape of the deformable mirror, η represents the learning rate that controls the iterative step size of the gradient descent, and the objective function converges to the minimum value through iterative optimization, outputs the laser beam corrected by adaptive optics, and ideally restores the laser beam corrected by adaptive optics. The specific formula is:

[0026] I cd =I(d)*T day *|ψ corrected (r)| 2

[0027] I cn =I(d)*T night *|ψ corrected (r)| 2

[0028] Among them, I cd represents the wavefront intensity of the ideally restored laser beam under daytime conditions, I cn represents the wavefront intensity of the ideally restored laser beam under nighttime conditions, T day Indicates the medium transmittance under daytime conditions, T night represents the medium transmittance under night conditions, ψ corrected (r) represents the wavefront intensity of the laser beam after adaptive optics correction.

[0029] Preferably, the time it takes for the laser beam to propagate through the window and reach the target area under day and night conditions and then be reflected to the receiving end of the laser radar is recorded, and the detection distance from the laser radar to the target area is calculated, and the specific formula is:

[0030]

[0031] Where R represents the detection distance from the laser radar to the target area, C represents the speed of light, and t round-tripIt represents the time when the laser beam is reflected to the receiving end of the laser radar after propagating through the window to the target area under day and night conditions. It associates the reflected beams of multiple laser beams at different angles and positions after propagating through the window to the target area under day and night conditions with the corresponding target characteristics, annotates the intensity label of the reflected beam and the category label of the target characteristic, extracts the time delay feature of the reflected beam and the reflectivity feature of the target characteristic, separates the intensity label of the reflected beam through the SVM classification algorithm, and uses the kernel function to map the category label of the target characteristic to the high-dimensional space to construct the optimal hyperplane. The projection point of the intensity label of the reflected beam on the optimal hyperplane is maximized through the high-dimensional space. The specific formula is:

[0032]

[0033] Where f(x) represents the classification result of the target feature, y i represents the intensity label of the i-th reflected beam, α i represents the Lagrange multiplier, K(x i ,x) represents the kernel function that maps the time delay characteristics of the i-th reflected light beam to the reflectivity characteristics of the target characteristics of different category labels, b represents the bias term that adjusts the distance between the hyperplane and the origin, and the three-dimensional model of the target is constructed according to the classification results of the target characteristics.

[0034] A day and night detection system based on a laser window, comprising a laser propagation attenuation module, a wavefront distortion quantification module, a wavefront distortion correction module and a target classification detection module;

[0035] Laser propagation attenuation module: uses the laser radar to emit a laser beam and propagate it through a transparent window, and establishes an exponential attenuation model of the laser beam to quantitatively describe the influence of the transparent window;

[0036] Wavefront distortion quantification module: monitors the actual wavefront phase in real time through the radar wavefront sensor, defines the ideal wavefront and calculates the wavefront distortion of the laser beam, and quantifies the degree of distortion based on the root mean square error;

[0037] Wavefront distortion correction module: uses the adaptive optical correction model to dynamically adjust the deformable mirror according to the real-time wavefront distortion to correct the wavefront distortion of the laser beam;

[0038] Target classification detection module: records the round-trip propagation time of the laser beam under day and night conditions to calculate the detection distance, and uses the support vector machine classification algorithm to construct the optimal hyperplane in the high-dimensional space to realize the classification of target characteristics and construct a three-dimensional model of the target.

[0039] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0040] 1. In the laser radar system, the emission and propagation of the laser beam are the core links. The laser radar is used to emit the laser beam and propagate it through the transparent window. An exponential attenuation model of the laser beam is established to quantitatively describe the influence of the transparent window on the laser beam. During the propagation process through the window, the attenuation of the laser beam is mainly affected by two factors: absorption and scattering. The absorption effect refers to the part of the light energy absorbed by the medium when the laser beam passes through the transparent window; and the scattering effect refers to the part of the light energy scattered to other directions due to the presence of molecules or particles in the medium during the propagation of the laser beam. The two effects work together to weaken the intensity of the laser beam, thereby affecting the detection capability. By establishing an exponential attenuation model to quantify the influence of the transparent window on the propagation of the laser beam, the design and application of the laser radar are optimized to ensure the best detection effect under different day and night conditions.

[0041] 2. By establishing an exponential attenuation model of the laser beam and quantitatively describing the effect of the window on the laser beam, the target can be effectively identified and the detected laser beam can be optimized, so that the laser radar can dynamically adjust its parameters under different day and night conditions, thereby minimizing the attenuation of the laser beam, improving the propagation efficiency of the laser beam, and enhancing the detection capability of the system. It also enhances the application adaptability in complex environments. The radar wavefront sensor monitors the actual wavefront phase in real time and compares it with the ideal wavefront, which can quickly identify the wavefront distortion, so that the laser beam can maintain high-quality wavefront characteristics during propagation. The quantification of the root mean square error provides an objective wavefront quality evaluation standard, thereby ensuring that the propagation and focusing capabilities of the laser beam are enhanced, reducing the beam distortion caused by wavefront distortion, and introducing adaptive The optical correction model enables the system to dynamically adjust the deformable mirror according to the wavefront distortion monitored in real time, correct the wavefront distortion of the laser beam, improve the response speed and adaptability of the lidar, and ensure that the quality of the laser beam is always maintained in the best state under various environmental conditions. Through iterative optimization, the shape of the deformable mirror can be accurately adjusted, further improving the focusing ability and detection accuracy of the laser beam. By recording the round-trip propagation time of the laser beam to calculate the detection distance, combined with the support vector machine classification algorithm, the classification of target characteristics can be realized in high-dimensional space, which not only improves the accuracy of target recognition, but also can construct a three-dimensional model of the target. By effectively mapping the intensity label of the reflected light beam with the category label of the target characteristic, efficient target detection and classification can be achieved in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0043] Figure 1 The present invention is a method flow chart of a day and night detection method based on a laser window.

[0044] Figure 2 The module schematic diagram of a day and night detection system based on a laser window according to the present invention. DETAILED DESCRIPTION

[0045] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0046] Example 1

[0047] The present invention provides Figure 1 A day and night detection method based on a laser window is shown, comprising the following steps:

[0048] Step 1: Use the laser radar to emit a laser beam and propagate it through the transparent window, and establish an exponential attenuation model of the laser beam to quantitatively describe the influence of the transparent window;

[0049] The laser radar selects the pulse mode to generate a laser beam and emits it to the target direction. The laser beam interacts with molecules or particles through the transparent window, resulting in the attenuation of the laser beam. An exponential attenuation model is established based on the attenuation coefficient to quantitatively describe the influence of the transparent window. The specific formula is:

[0050] I(d)=I0e -αd α

[0051] Wherein, I(d) represents the intensity of the laser beam at a distance d after propagating through the window, I0 represents the initial intensity of the laser beam emitted to the target direction after propagating through the window, and α represents the attenuation coefficient of propagation through the window, wherein the attenuation coefficient reflects the attenuation degree of the intensity of the laser beam as the propagation distance increases after propagating through the window. The attenuation factors of the laser beam in the process of propagation through the window are divided into absorption effect and scattering effect. Among them, the absorption effect is that the medium through which the laser beam passes in the process of propagation through the window absorbs part of the light energy, and the scattering effect is that the molecules or particles in the atmosphere scatter during the propagation of the laser beam through the window, causing part of the light energy to be transferred to other directions. After the laser beam propagates through the window and reaches the target area, it is reflected back to the receiving end of the lidar. The reflected beam is detected by the photoelectric detector at the receiving end, and its intensity is related to the distance and characteristics of the target reflection.

[0052] Step 2: Monitor the actual wavefront phase in real time through the radar wavefront sensor, define the ideal wavefront and calculate the wavefront distortion of the laser beam, and quantify the degree of distortion according to the root mean square error;

[0053] In the medium environment of the target area under day and night conditions, the laser beam propagates through the window, and the changes in temperature, pressure and density at the day and night levels cause random fluctuations in the propagation path of the laser beam, which manifests as wavefront distortion of the laser beam. For a plane wave, the ideal wavefront is a plane, and for a spherical wave, the ideal wavefront is a sphere. When propagating in the free space of the target area, the wavefront is optically ideal. However, the wavefront actually monitored by the wavefront sensor will change due to the medium environment of the target.

[0054] The actual wavefront phase caused by the medium environment of the target area under day and night conditions on the propagation of the laser beam is monitored in real time by the radar wavefront sensor, and compared with the ideal wavefront phase. The ideal wavefront is defined as the wavefront when the laser beam is emitted from the lidar, and the wavefront distortion of the laser beam is calculated. The specific formula is:

[0055] Δφ(r)=φ m (r)-φ0(r)

[0056] Where Δφ(r) represents the wavefront distortion, φ m (r) represents the actual wavefront phase, φ0(r) represents the ideal wavefront phase, r represents the spatial position of the transmitted wavefront distortion, and the distortion mode is checked. The phase diagram and amplitude diagram of the wavefront are drawn according to the wavefront distortion. The wavefront distortion is defined as a phase jump according to the instantaneous change of the phase value of the laser beam at a certain point in the wavefront. By performing differential analysis on the phase diagram of the wavefront, it is observed whether there is a discontinuous phase value. The wavefront distortion is defined as a focal shift according to the change in the focal position when the laser beam is emitted to the target area. It is identified by the change in the focal position of the coherent interference pattern. The wavefront data is analyzed by Fourier transform to check whether there is a periodic pattern of a specific frequency. The root mean square error of the wavefront distortion is calculated to quantify the degree of distortion. The specific formula is:

[0057]

[0058] Among them, RMS represents the root mean square error of wavefront distortion, represents the measure of the wavefront distortion quantification result, N represents the total number of points where the radar wavefront sensor monitors the actual wavefront phase in real time, Δφ(r i ) represents the number of points corresponding to the i-th actual wavefront phase monitored by the radar wavefront sensor in real time r i The wavefront distortion at the position, where the smaller the RMS, the smaller the wavefront distortion, the higher the wavefront quality, and the enhanced propagation and focusing capabilities of the laser beam. The larger the RMS, the greater the wavefront distortion, resulting in a decrease in the quality of the laser beam during propagation.

[0059] Step 3: Using an adaptive optical correction model to dynamically adjust the deformable mirror to correct the wavefront distortion of the laser beam according to the real-time wavefront distortion;

[0060] The core component of the adaptive optics correction model is defined as a deformable mirror. The deformable mirror is used to control the displacement of each point on the surface to achieve the correction of wavefront distortion. In the adaptive optics correction model, there is a physical relationship that the shape of the deformable mirror produces reverse wavefront distortion to deduce the adjustment amount of the deformable mirror. The objective function is defined to make the wavefront after adaptive optics correction consistent with the ideal wavefront by adjusting the shape of the deformable mirror. The specific formula is:

[0061]

[0062] Where Cost represents the error after wavefront adaptive optics correction, D(r) represents the control signal of each point on the deformable mirror surface, It represents the mapping process from the deformable mirror shape to the wavefront. The gradient of the calculated objective function represents the sensitivity to the deformable mirror shape, which is used to adjust each point of the deformable mirror to reduce and minimize the error after wavefront adaptive optics correction. The specific formula is:

[0063]

[0064] in, Represents the gradient of the objective function, and updates the shape of the deformable mirror by gradient descent. The specific formula is:

[0065]

[0066] Among them, D new (r) represents the updated shape of the deformable mirror, D old (r) represents the initial shape of the deformable mirror, η represents the learning rate that controls the iterative step size of the gradient descent, and the objective function converges to the minimum value through iterative optimization, outputs the laser beam corrected by adaptive optics, and ideally restores the laser beam corrected by adaptive optics. The specific formula is:

[0067] I cd =I(d)*T day *ψ corrected (r) 2

[0068] I cn =I(d)*T night *|ψ corrected (r)| 2

[0069] Among them, I cd represents the wavefront intensity of the ideally restored laser beam under daytime conditions, I cn represents the wavefront intensity of the ideally restored laser beam under nighttime conditions, T day Indicates the medium transmittance under daytime conditions, Tnight represents the medium transmittance under night conditions, ψ corrected (r) represents the wavefront intensity of the laser beam after adaptive optics correction.

[0070] Step 4: Record the round-trip propagation time of the laser beam under day and night conditions to calculate the detection distance, and use the support vector machine classification algorithm to construct the optimal hyperplane in the high-dimensional space to classify the target characteristics and construct a three-dimensional model of the target;

[0071] Record the time it takes for the laser beam to propagate through the window and reach the target area under day and night conditions and then reflect to the receiving end of the laser radar, and calculate the detection distance from the laser radar to the target area. The specific formula is:

[0072]

[0073] Where R represents the detection distance from the laser radar to the target area, C represents the speed of light, and t round-trip It represents the time when the laser beam is reflected to the receiving end of the laser radar after propagating through the window to the target area under day and night conditions. It associates the reflected beams of multiple laser beams at different angles and positions after propagating through the window to the target area under day and night conditions with the corresponding target characteristics, annotates the intensity label of the reflected beam and the category label of the target characteristic, extracts the time delay feature of the reflected beam and the reflectivity feature of the target characteristic, separates the intensity label of the reflected beam through the SVM classification algorithm, and uses the kernel function to map the category label of the target characteristic to the high-dimensional space to construct the optimal hyperplane. The projection point of the intensity label of the reflected beam on the optimal hyperplane is maximized through the high-dimensional space. The specific formula is:

[0074]

[0075] Where f(x) represents the classification result of the target feature, y i represents the intensity label of the i-th reflected beam, α i represents the Lagrange multiplier, K(x i ,x) represents the kernel function that maps the time delay characteristics of the i-th reflected light beam to the reflectivity characteristics of the target characteristics of different category labels, b represents the bias term that adjusts the distance between the hyperplane and the origin, and the three-dimensional model of the target is constructed according to the classification results of the target characteristics.

[0076] Example 2

[0077] The present invention provides Figure 2 A day and night detection system based on a laser window is shown, comprising a laser propagation attenuation module, a wavefront distortion quantification module, a wavefront distortion correction module and a target classification detection module;

[0078] Laser propagation attenuation module: uses the laser radar to emit a laser beam and propagate it through a transparent window, and establishes an exponential attenuation model of the laser beam to quantitatively describe the influence of the transparent window;

[0079] In this embodiment, the laser propagation attenuation module is specifically required to be explained. Assuming that the laser radar selects the pulse mode to generate the laser beam, the initial intensity of which is 100w after being transmitted through the window to the target direction, and the distance from the laser beam being emitted to the target direction to the photoelectric detector at the receiving end detecting the reflected beam is 500m. Assuming that under certain meteorological conditions, the attenuation coefficient of the laser beam transmitted through the window is 0.01m-1, the intensity of the laser beam at a distance of 500m after being transmitted through the window is I(d)=100e -0.01*500 ≈0.674w. It can be seen that although the initial intensity of the laser is 100w, due to the long propagation distance and the absorption and scattering effects of the window medium, the photodetector at the receiving end finally detects that the intensity of the reflected light beam is only 0.674w.

[0080] Wavefront distortion quantification module: monitors the actual wavefront phase in real time through the radar wavefront sensor, defines the ideal wavefront and calculates the wavefront distortion of the laser beam, and quantifies the degree of distortion based on the root mean square error;

[0081] In this embodiment, what needs to be explained specifically is the wavefront distortion quantification module. Wavefront distortion refers to the change in the shape of the wavefront caused by the medium environment during the propagation of the laser beam. Ideally, when the laser is emitted from the radar, the wavefront of the beam is an ideal plane wave or spherical wave, but in actual propagation it will be disturbed by environmental factors, resulting in wavefront distortion. If the laser beam is emitted from a plane source, the ideal wavefront phase is a flat phase distribution. The actual wavefront phase refers to the wavefront phase formed after the laser beam passes through the medium change during actual propagation. Different types of wavefront distortion modes are identified by real-time monitoring of the wavefront phase change, where a phase jump refers to a sudden change in the phase value of certain areas in the wavefront, and a focus shift refers to the fact that when the wavefront is distorted, the focal position of the laser beam may shift, affecting the focusing ability.

[0082] Wavefront distortion correction module: uses the adaptive optical correction model to dynamically adjust the deformable mirror according to the real-time wavefront distortion to correct the wavefront distortion of the laser beam;

[0083] In this embodiment, what needs to be explained specifically is the wavefront distortion correction module, which ensures high-quality three-dimensional construction and transmission during the propagation of the light beam by correcting the distortion of the light wavefront, and corrects the wavefront distortion caused by the medium environment by dynamically adjusting the deformable mirror. The deformable mirror is composed of multiple independently controllable tiny elements. According to the feedback error information, the surface of the deformable mirror will undergo a tiny displacement, thereby correcting the light wavefront, which can effectively eliminate the wavefront distortion of the light beam.

[0084] Target classification detection module: records the round-trip propagation time of the laser beam under day and night conditions to calculate the detection distance, and uses the support vector machine classification algorithm to construct the optimal hyperplane in the high-dimensional space to realize the classification of target characteristics and construct a three-dimensional model of the target.

[0085] In this embodiment, what needs to be specifically explained is the target classification detection module. The time required for the laser beam to be emitted and received reflects the distance information of the target. The intensity of the reflected light beam is related to the material, shape and target characteristics of the target. Through multiple laser beam reflections, the spatial distribution information of the target area can be obtained. The reflected light beam data obtained from the lidar is analyzed to extract effective features, and classified through machine learning algorithms, so as to identify and distinguish different types of targets.

[0086] By establishing an exponential attenuation model of the laser beam and quantitatively describing the effect of the window on the laser beam, the target can be effectively identified and the detected laser beam can be optimized, so that the laser radar can dynamically adjust its parameters under different day and night conditions, thereby minimizing signal attenuation, improving the propagation efficiency of the laser beam, and enhancing the detection capability of the system. It also enhances the application adaptability in complex environments. The radar wavefront sensor monitors the actual wavefront phase in real time and compares it with the ideal wavefront, which can quickly identify wavefront distortion, so that the laser beam can maintain high-quality wavefront characteristics during propagation. The quantification of the root mean square error provides an objective wavefront quality evaluation standard, thereby ensuring that the propagation and focusing capabilities of the laser beam are enhanced, reducing signal distortion caused by wavefront distortion, and introducing adaptive light The mechanical correction model enables the system to dynamically adjust the deformable mirror according to the wavefront distortion monitored in real time, correct the wavefront distortion of the laser beam, improve the response speed and adaptability of the lidar, and ensure that the quality of the laser beam is always maintained in the best state under various environmental conditions. Through iterative optimization, the shape of the deformable mirror can be accurately adjusted, further improving the focusing ability and detection accuracy of the laser beam. By recording the round-trip propagation time of the laser beam to calculate the detection distance, combined with the support vector machine classification algorithm, the classification of target characteristics can be realized in high-dimensional space, which not only improves the accuracy of target recognition, but also can construct a three-dimensional model of the target. By effectively mapping the intensity label of the reflected light beam with the category label of the target characteristic, efficient target detection and classification can be achieved in complex environments.

[0087] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A day and night detection method based on laser window, characterized in that: The following steps are involved: Step 1: Use the laser radar to emit a laser beam and propagate it through the transparent window, and establish an exponential attenuation model of the laser beam to quantitatively describe the influence of the transparent window; Step 2: Monitor the actual wavefront phase in real time through the radar wavefront sensor, define the ideal wavefront and calculate the wavefront distortion of the laser beam, and quantify the degree of distortion according to the root mean square error; Step 3: Using an adaptive optical correction model to dynamically adjust the deformable mirror to correct the wavefront distortion of the laser beam according to the real-time wavefront distortion; Step 4: Record the round-trip propagation time of the laser beam under day and night conditions to calculate the detection distance, and use the support vector machine classification algorithm to construct the optimal hyperplane in the high-dimensional space to classify the target characteristics and construct a three-dimensional model of the target.

2. The method for day and night detection based on a laser window according to claim 1, characterized in that: In the step 1, an exponential decay model is established according to the attenuation coefficient to quantitatively describe the influence of the transparent window, and the specific formula is: I(d)=I0e -αd Wherein, I(d) represents the intensity of the laser beam at a distance d after propagating through the window, I0 represents the initial intensity of the laser beam emitted to the target area after propagating through the window, and α represents the attenuation coefficient of propagation through the window.

3. The method for day and night detection based on a laser window according to claim 1, characterized in that: In the step 2, the ideal wavefront is defined to represent the wavefront when the laser beam is emitted from the laser radar, and the wavefront distortion of the laser beam is calculated. The specific formula is: Δφ(r)=φ m (r)-φ0(r) Where Δφ(r) represents the wavefront distortion, φ m (r) represents the actual wavefront phase, φ0(r) represents the ideal wavefront phase, and r represents the spatial position of the transmitted wavefront distortion.

4. The method for day and night detection based on a laser window according to claim 1, characterized in that: In the step 2, the root mean square error of the wavefront distortion is calculated to quantify the degree of distortion, and the specific formula is: Where RMS represents the root mean square error of wavefront distortion, represents the measure of the wavefront distortion quantification result, N represents the total number of points where the radar wavefront sensor monitors the actual wavefront phase in real time, Δφ(r i ) represents the number of points corresponding to the actual wavefront phase of the radar wavefront sensor monitoring in real time r i Wavefront distortion in position.

5. The method for day and night detection based on laser window according to claim 1, characterized in that: In the step 3, in the adaptive optics correction model, an objective function is defined to make the wavefront after adaptive optics correction consistent with the ideal wavefront by adjusting the shape of the deformable mirror, and its specific formula is: Where Cost represents the error after wavefront adaptive optics correction, D(r) represents the control signal of each point on the deformable mirror surface, Represents the mapping process from deformable mirror shape to wavefront.

6. The method for day and night detection based on laser window according to claim 5, characterized in that: The gradient of the objective function is calculated to represent the sensitivity to the shape of the deformable mirror. Each point of the deformable mirror is adjusted to minimize the error after wavefront adaptive optics correction. The specific formula is: in, Represents the gradient of the objective function, and updates the shape of the deformable mirror by gradient descent. The specific formula is: Among them, D new (r) represents the updated shape of the deformable mirror, D old (r) represents the initial shape of the deformable mirror, and η represents the learning rate that controls the iterative step size of the gradient descent.

7. The method for day and night detection based on laser window according to claim 6, characterized in that: The specific formula for ideal restoration of the laser beam after adaptive optics correction is: I cd =I(d)*T day *|ψ corrected (r)| 2 I cn =I(d)*T night *|ψ corrected (r)| 2 Among them, I cd represents the wavefront intensity of the ideally restored laser beam under daytime conditions, I cn represents the wavefront intensity of the ideally restored laser beam under nighttime conditions, T day Indicates the medium transmittance under daytime conditions, T night represents the medium transmittance under night conditions, ψ corrected (r) represents the wavefront intensity of the laser beam after adaptive optics correction.

8. The method for day and night detection based on laser window according to claim 1, characterized in that: In step 4, the specific formula for calculating the detection distance is: Where R represents the detection distance from the laser radar to the target area, C represents the speed of light, and t round-trip It indicates the time it takes for the laser beam to propagate through the window, reach the target area under day and night conditions, and then be reflected to the receiving end of the lidar.

9. The method for day and night detection based on laser window according to claim 1, characterized in that: In step 4, the specific steps of using the support vector machine classification algorithm to construct an optimal hyperplane in a high-dimensional space to achieve the classification of target characteristics are: separating the intensity labels of the reflected light beam and mapping the category labels of the target characteristics to the high-dimensional space using a kernel function to construct an optimal hyperplane, and maximizing the separation of the intensity labels of the reflected light beam in the high-dimensional space on the projection point of the optimal hyperplane, and the specific formula is: Where f(x) represents the classification result of the target feature, y i represents the intensity label of the i-th reflected beam, α i represents the Lagrange multiplier, K(x i ,x) represents the kernel function that maps the time delay feature of the i-th reflected beam to the reflectivity feature of the target characteristics of different category labels, and b represents the bias term that adjusts the distance between the hyperplane and the origin.

10. A day and night detection system based on a laser window, used to implement a day and night detection method based on a laser window as described in any one of claims 1 to 9, characterized in that: It includes a laser propagation attenuation module, a wavefront distortion quantification module, a wavefront distortion correction module and a target classification detection module; Laser propagation attenuation module: uses the laser radar to emit a laser beam and propagate it through a transparent window, and establishes an exponential attenuation model of the laser beam to quantitatively describe the influence of the transparent window; Wavefront distortion quantification module: monitors the actual wavefront phase in real time through the radar wavefront sensor, defines the ideal wavefront and calculates the wavefront distortion of the laser beam, and quantifies the degree of distortion based on the root mean square error; Wavefront distortion correction module: uses the adaptive optical correction model to dynamically adjust the deformable mirror according to the real-time wavefront distortion to correct the wavefront distortion of the laser beam; Target classification detection module: records the round-trip propagation time of the laser beam under day and night conditions to calculate the detection distance, and uses the support vector machine classification algorithm to construct the optimal hyperplane in the high-dimensional space to realize the classification of target characteristics and construct a three-dimensional model of the target.

Images