Strong background light license plate recognition system and method

Abstract

The invention discloses a strong background light license plate recognition system and method, relates to the technical field of optics, and aims to solve the problem that a driving vehicle license plate is difficult to recognize under the conditions of strong light interference and low light in traffic monitoring. The system is composed of a laser illuminator, a polarizing film, a visible light color camera, a visible light color polarization camera, an attenuation sheet, a data acquisition system, a power supply module, a controller, an image processor, a laser beam expanding lens group, a visible light imaging lens group and a visible light polarization imaging lens group. Compared with a traditional traffic monitoring probe, the system can effectively restrain strong light interference generated by an automobile headlamp at night, and the imaging contrast ratio and recognition precision of a license plate in the strong light background are improved. The visible light color camera and the visible light color polarization camera are combined for use, the vehicle contour and the surrounding background image obtained by the visible light color camera are reserved, detail information and background information lost by the polarization image can be effectively supplemented, a basis is provided for judgment of management personnel, and advantage complementation is achieved.

Description

Technical Field

[0001] The invention relates to the field of optical technology, and in particular to a system and method for recognizing a license plate under strong background light, which is used for fast capture and accurate recognition of a license plate under strong background light. Background Art

[0002] The application of license plate detection and recognition technology has been very extensive, especially in complex environments. Due to the influence of many uncertain factors in some complex scenes (such as strong light background, high speed, weak light, etc.), the existing technologies currently available are either less targeted or have poor recognition effects.

[0003] The Chinese patent publication number is "CN211878342U", and the patent name is "A license plate recognition polarization device". It fixes the polarized lens directly in front of the camera lens, and uses the filtering effect of the polarized lens on the reflected light to reduce the intensity of sunlight reflected light, improve the efficiency of license plate recognition, and finally realize the recognition of the license plate. This method depends largely on the illumination of the license plate surface by natural light during the day, especially when the recognition environment changes (such as weak light, high headlight brightness), it is necessary to make corresponding adjustments according to different environmental conditions. At the same time, if the polarized lens is directly installed in front of the camera, the clarity will be weakened and the contrast will be reduced. It does not have the characteristics of dynamic real-time imaging of the target and actively enhancing the polarization characteristics of the target.

[0004] A single polarizer can only filter one polarization direction. For dynamic targets, the relative position and posture change in real time. Imaging with only one polarization direction can only reduce the glare for the relative position and posture of a specific target. If multi-angle polarization filtering is required for dynamic targets, vehicles, etc., the polarizer direction needs to be rotated and matched one-to-one with the detector frame rate, which will cause problems such as shooting delay and reduced structural stability, and cannot meet the imaging requirements of dynamic high-speed targets. Summary of the invention

[0005] In order to solve the problem that the license plate of a moving vehicle is difficult to identify under strong light interference and low light conditions in traffic monitoring, the present invention provides a strong background light license plate recognition system and a method for quickly capturing and accurately identifying specific information of the license plate in the strong background light direction.

[0006] A strong background light license plate recognition system, the system is composed of a laser illuminator, a polarizer, a visible light color camera, a visible light color polarization camera, an attenuation plate, a data acquisition system, a power module, a controller and an image processor, a laser beam expander lens group, a visible light imaging lens group and a visible light polarization imaging lens group;

[0007] The laser illuminator is connected to a laser beam expander lens group, and a narrow-band polarizer is installed in front of a light outlet of the laser beam expander lens group;

[0008] The visible light color camera is connected to the visible light imaging lens group;

[0009] The visible light color polarization camera is connected to the visible light polarization imaging lens group, and the attenuation plate is installed in front of the visible light polarization imaging lens group;

[0010] The controller is connected to the laser illuminator, the visible light color camera, and the visible light color polarization camera respectively;

[0011] The visible light color camera and the visible light color polarization camera are respectively connected to the data acquisition system;

[0012] The data acquisition system is connected to the image processor;

[0013] The controller is connected to the laser illuminator, the visible light color camera, and the visible light color polarization camera;

[0014] The image processor is connected to the controller;

[0015] The power supply module is connected to the controller (8) and the image processor respectively.

[0016] A method for a strong background light license plate recognition system, the method comprising the following steps:

[0017] The first step is target capture and illumination;

[0018] When a vehicle enters the visible light imaging lens group and focuses on the field of view of the visible light color camera, the image processor determines whether the vehicle enters the field of view based on the abnormal light intensity area in the image. When the target vehicle enters the field of view, a trigger signal is quickly sent to the controller. After receiving the trigger signal from the visible light color camera, the controller quickly triggers the laser illuminator to continuously emit laser light, which is collimated and expanded by the laser beam expansion lens group to illuminate the car for rapid exposure until the target moves out of the field of view.

[0019] The required laser illuminator power and the total power reflected back by the target can be calculated based on the license plate laser echo power model:

[0020] The laser illuminator has an output wavelength of 850nm and emits p t The output power of the receiving module is A. r , then the laser emission light is reflected by the license plate surface, and the final total receiving power of the receiving module is:

[0021]

[0022] Where σ is the backscattering cross section; AREA≈π(Rθ B / 2) 2 is the beam coverage area at distance R; θ Bis the beam divergence angle, where: the deviation angle between the reflected light edge ray and the main ray is set as θ c , the expression is θ c =2arctan(r2-r1 / R) and at the edge of the light θ c =θ B , r1 and r2 are the light emitting radius and the spot radius respectively;

[0023] According to the above formula, combined with the actual imaging illumination distance and target size, the laser illuminator can be selected in a targeted manner to determine the laser illuminator power and beam divergence angle;

[0024] The second step is camera exposure;

[0025] The controller receives a trigger signal from the visible light color camera, and synchronously triggers the visible light color camera and the visible light color polarization camera to work, and respectively focuses and images through the visible light imaging lens group and the visible light polarization imaging lens group, so as to continuously take pictures of the target at the same frame rate, and synchronously obtain the visible light color image and the visible light polarization image of the target background;

[0026] The third step is data collection;

[0027] The visible light color camera and the visible light color polarization camera collect and store dual-channel images through the data acquisition system. The visible light color camera obtains color images in the 400nm-900nm band; the visible light color polarization camera obtains images of different polarization directions of 0°, 45°, 90°, and 135° in the 400nm-900nm band, and transmits the two-channel images to the image processor through a wired or wireless network transmission interface;

[0028] Step 4: Polarization image processing;

[0029] The image processor solves the polarization image, and the relationship between the intensity information of different polarization directions and the degree of polarization is:

[0030]

[0031]

[0032] According to the laser illumination, the bidirectional reflection distribution function that characterizes the incident and reflected light distribution of the license plate is obtained as follows:

[0033]

[0034] Using the relationship between the degree of polarization and the intensity information of different polarization directions, the Mueller matrix that characterizes the surface characteristics of the license plate can be obtained:

[0035]

[0036] The bidirectional reflection distribution function model of the laser reflected light after passing through the license plate surface is:

[0037]

[0038]

[0039] At this time, the Stokes vector of the laser reflected light after the license plate can be expressed as:

[0040]

[0041] According to the Stokes vector and polarization calculation formula of the reflected light from the license plate, the polarization expression of the reflected light based on the laser incident target surface can be obtained as follows:

[0042]

[0043] Ignore the energy loss during the propagation process, that is, when P0=P t When , the polarization degree expression is:

[0044]

[0045] According to the above formula, the polarization degree of the license plate surface is affected by multiple parameters such as the incident angle, detection angle, beam divergence angle and medium refractive index. Using the above formula, the laser illuminator parameters and the color polarization camera observation angle can be set in a targeted manner to achieve higher polarization image acquisition and improve the polarization image effect.

[0046] The fifth step is image fusion processing;

[0047] The license plate image dataset is processed, fused and reconstructed using a visible light color camera and a visible light color polarization camera. First, the R, G and B three-channel color image data collected by the visible light color camera and the R, G and B three-channel DOLP image data collected by the visible light color polarization image are extracted. Secondly, the two images are divided into three channels of R, G and B for data feature point registration to achieve target alignment of images on different channels. Then, the aligned R, G and B three-channel data are subjected to weighted average image fusion processing to form polarization fused images of R, G and B channels respectively. Finally, the polarization fused image data of R, G and B channels are used to reconstruct the color image, complete the color polarization fusion image processing, highlight the vehicle profile, personnel face and license plate details, and achieve feature enhancement in the image detail area.

[0048] The beneficial effects of the present invention are:

[0049] 1) Compared with the traditional single visible light color traffic monitoring probe, this system can effectively suppress the strong light interference generated by car headlights at night and improve the imaging contrast and recognition accuracy of license plates in strong light backgrounds.

[0050] 2) A laser illuminator is added to illuminate the vehicle license plate. The system combines a laser illuminator, a laser beam expander and a polarizer to purify the polarization of the illumination light source, which can increase the polarization degree of the emitted illumination light and effectively reduce the polarization attenuation of the laser illumination light source during atmospheric transmission. When the purified polarized light is irradiated onto the license plate, the license plate can produce specific polarized light, thereby enhancing the polarization characteristics of the license plate. This method can effectively improve the imaging illumination in a dim environment, while improving the polarization imaging effect, and ultimately improve the signal-to-noise ratio of the polarization image.

[0051] 3) The system adopts the split-focus plane polarization imaging method. Compared with the traditional rotating polarizer imaging method, this method can simultaneously obtain polarization images in four polarization directions with a single exposure, ensuring the position consistency of the target in images of different polarization directions, and effectively solving the field of view error caused by the rotating polarizer when imaging dynamic targets. It effectively improves the polarization solution accuracy of high-speed dynamic targets, has the characteristics of stronger real-time polarization imaging, higher resolution, no moving parts in the system, higher reliability, and more compact structure, and is more suitable for polarization imaging of moving vehicles.

[0052] 4) This method combines the attenuation plate and the visible light color polarization camera, and combines it with active polarization laser illumination. While effectively attenuating strong light such as car lights entering the color polarization camera, it avoids overexposure of the color polarization camera, and actively and controllably ensures that the color polarization camera can obtain effective polarization information. Since the polarization image is independent of the absolute intensity of the target and background, it is only related to the intensity difference between the target and the background. Since the license plate material is different from the surface material of the car and the roughness is different, its polarization characteristics are different. Therefore, the polarization intensity of the license plate target and the car body background is different. When the attenuation plate is used to filter the light intensity of the target and the background incident on the camera at the same time, the traditional visible light color camera will reduce the responsiveness of the target and the background at the same time, resulting in a deterioration of the target signal-to-noise ratio; but for the color polarization camera, the target and the background are imaged at the same time under the premise of ensuring that the color polarization camera is not overexposed, and the target imaging contrast can be improved by only using the polarization difference between the target and the background. Therefore, this method uses the "weak light enhancement" feature of polarization to effectively suppress the strong light background, rather than the "strong light weakening" effect of polarization used in the usual method.

[0053] 5) This method combines a visible light color camera and a visible light color polarization camera to synchronously acquire dual-channel images and fuse and enhance them. While highlighting the target contour information, it retains the vehicle contour and surrounding background image acquired by the visible light color camera. It can effectively supplement the lost detail information and background information of the polarization image, provide a basis for managers to make judgments, and achieve complementary advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The present invention is a structural schematic diagram of a strong background light license plate recognition system.

[0055] Figure 2 is a schematic diagram of a polarization focal plane array.

[0056] Figure 3 This is the effect of taking a backlit license plate photo with a visible light color camera.

[0057] Figure 4 It is the flow chart of color polarization fusion image fusion processing.

[0058] Figure 5 This is the effect picture of backlit license plate recognition using color polarization fusion image without laser fill light.

[0059] Figure 6 This is the effect picture of laser fill light color polarization fusion image of backlit license plate recognition. DETAILED DESCRIPTION

[0060] like Figure 1 As shown, a strong background light license plate recognition system is composed of a laser illuminator 1, a polarizer 2, a visible light color camera 3, a visible light color polarization camera 4, an attenuation plate 5, a data acquisition system 6, a power module 7, a controller 8, an image processor 9, a laser beam expander lens group 10, a visible light imaging lens group 11 and a visible light polarization imaging lens group 12.

[0061] The laser illuminator 1 is connected to the laser beam expander lens group 10, and the narrow-band polarizer 2 is installed in front of the light outlet of the laser beam expander lens group 10;

[0062] The visible light color camera 3 is connected to the visible light imaging lens group 11;

[0063] The visible light color polarization camera 4 is connected to the visible light polarization imaging lens group 12, and the attenuation plate 5 is installed in front of the visible light polarization imaging lens group 12;

[0064] The controller 8 is connected to the laser illuminator 1, the visible light color camera 3, and the visible light color polarization camera 4 respectively;

[0065] The visible light color camera 3 and the visible light color polarization camera 4 are respectively connected to the data acquisition system 6;

[0066] The data acquisition system 6 is connected to the image processor 9;

[0067] The controller 8 is connected to the laser illuminator 1, the visible light color camera 3, and the visible light color polarization camera 4;

[0068] The image processor 9 is connected to the controller 8;

[0069] The power module 7 is connected to the controller 8 and the image processor 9 respectively.

[0070] The laser illuminator 1 actively illuminates the moving vehicle, and the polarizer is used to further purify the polarization degree of the laser illumination light source, improve the polarization degree of the emitted illumination light, and enhance the polarization characteristics of the target when it is irradiated on the target;

[0071] The visible light color camera 3 uses a CMOS 5MP visible light non-biased camera. The module can realize high-resolution (2464*2056) recognition function and provide a speed of up to 97 frames per second through the camera link output. The camera has excellent sensitivity and dynamic range, and has functions such as automatic exposure. It can quickly capture high-speed moving vehicles, and quickly capture and monitor and record the vehicle shape and surrounding environment background. At the same time, it gives a trigger signal to the laser illuminator to make the laser illuminator power on and emit light;

[0072] The visible light color polarization camera 4 has high pixel (3.45μm*3.45μm), high resolution (2464*2056), and high frame rate (163fps). The camera adopts a split-focal plane polarization imaging method. Each detector pixel corresponds to a polarization direction, which are 0°, 45°, 90°, and 135° linear polarization directions. 2×2 pixels form a superpixel. The polarization degree image is solved by extracting the grayscale of pixels with the same polarization direction. The pixel polarization direction is arranged as follows: Figure 2 As shown. The visible light color polarization camera can obtain the polarization echo signal of the laser illuminator 1 reflected by the vehicle, complete the photo of the vehicle license plate, and simultaneously obtain the polarization images of the three channels of the license plate R, G and B in four polarization directions of 0°, 45°, 90°, and 135° in a single exposure, which is used for subsequent polarization image solution. The camera can provide images of different polarization directions (0°, 45°, 90°, 135°) of the three channels of R, G and B and different formats (AoP, DoLP, Intensity) of the three channels of R, G and B to achieve color polarization imaging, which can significantly improve the contrast between the detection target and the background, reduce the influence of reflected light, and has fog-penetrating characteristics. It can overcome harsh environments such as strong light background, dark background, foggy and rainy weather.

[0073] The attenuation sheet 5 is used to attenuate the light intensity of the vehicle light incident on the visible light color polarization camera 4 , so as to prevent the visible light color polarization camera 4 from being overexposed and causing the polarization degree image calculation to fail.

[0074] The controller 8 is used to control the visible light color camera 3 and the visible light color polarization camera 4 to synchronously capture images, and has the function of adjusting parameters such as camera exposure time and data format.

[0075] The data acquisition system 6 includes a dual-channel camerlink FULL acquisition card, a data line, a storage module, a data output interface and other parts, and completes the synchronous acquisition, storage and transmission of the dual-channel images of the visible light color camera 3 and the visible light color polarization camera 4.

[0076] The image processor 9 is used to determine whether the target enters the visible light color camera 3. At the same time, it can also solve the polarization image obtained by the visible light color polarization camera, and fuse and enhance the visible light color camera 3 and the solved visible light polarization image to complete the target details and background color information, so as to facilitate the staff to judge the vehicle model, violation location, etc. according to the background information and vehicle outline information.

[0077] The laser beam expander lens group 10 is used to collimate and expand the light beam emitted by the laser illuminator 1 to achieve uniform irradiation of a distant target;

[0078] The visible light imaging lens group 11 and the visible light polarization imaging lens group 12 are both used to collect target reflected echoes, focus them on the visible light color camera 3 and the visible light color polarization camera 4 respectively, and obtain the target image;

[0079] The power module 7 is connected to the visible light color camera 3 , the visible light color polarization camera 4 , the laser illuminator 1 , the controller 8 and the image processor 9 for supplying power.

[0080] When the vehicle enters the field of view of the visible light color camera 3 , the image processor 9 determines that the vehicle has entered the field of view and sends a trigger signal to the controller 8 , and the controller 8 triggers the laser illuminator 1 to continuously and rapidly expose.

[0081] At the same time, the controller 8 synchronously triggers the visible light color camera 3 and the visible light color polarization camera 4 to take pictures.

[0082] The images acquired by the visible light color camera 3 and the visible light color polarization camera 4 are transmitted to the image processor 9 via a wired or wireless network.

[0083] The image processor 9 calculates the polarization image and fuses and enhances the two camera images.

[0084] A method for recognizing a license plate under strong background light, the method comprising the following steps:

[0085] The first step is target capture and illumination;

[0086] When the vehicle enters the visible light imaging lens group 11 and focuses on the field of view of the visible light color camera 3, as shown in FIG. Figure 3As shown, the image processor 9 determines whether a vehicle enters the field of view based on the abnormal light intensity area in the image. When the target vehicle enters the field of view, the controller 8 quickly sends a trigger signal. After receiving the trigger signal from the visible light color camera 3, the controller 8 quickly triggers the laser illuminator 1 to continuously emit laser light, which is collimated and expanded by the laser beam expander group 10 to irradiate the car for rapid exposure until the target moves out of the field of view.

[0087] The required laser illuminator power and the total power reflected back by the target can be calculated based on the license plate laser echo power model:

[0088] The laser illuminator 1 has an output wavelength of 850nm and emits p t The output power of the receiving module is A. r , then the laser emission light is reflected by the license plate surface, and the final total receiving power of the receiving module is:

[0089]

[0090] Where σ is the backscattering cross section; AREA≈π(Rθ B / 2) 2 is the beam coverage area at distance R; θ B is the beam divergence angle (where the deviation angle between the reflected light edge ray and the main ray is set as θ c , the expression is θ c =2arctan(r2-r1 / R) and at the edge of the light θ c =θ B , r1 and r2 are the light exit radius and spot radius respectively).

[0091] According to the above formula, combined with the actual imaging illumination distance and target size, the laser illuminator can be selected in a targeted manner to determine the laser illuminator power and beam divergence angle.

[0092] The second step is camera exposure;

[0093] The controller 8 receives the trigger signal of the visible light color camera 3, and synchronously triggers the visible light color camera 3 and the visible light color polarization camera 4 to work, and respectively focuses and images through the visible light imaging lens group 11 and the visible light polarization imaging lens group 12, so as to complete continuous photography of the target at the same frame rate, and synchronously obtain the visible light color image and the visible light polarization image of the target background.

[0094] The third step is data collection;

[0095] The visible light color camera 3 and the visible light color polarization camera 4 collect and store dual-channel images through the data acquisition system 6. The visible light color camera 3 obtains color images in the 400nm~900nm band; the visible light color polarization camera 4 obtains images of different polarization directions (0°, 45°, 90°, 135°) in the 400nm~900nm band, and transmits the two-channel images to the image processor 9 through a wired or wireless network transmission interface.

[0096] Step 4: Polarization image processing;

[0097] The image processor 9 calculates the polarization image, and the relationship between the intensity information of different polarization directions and the degree of polarization (Dop) is:

[0098]

[0099]

[0100] According to the laser illumination, the bidirectional reflection distribution function that characterizes the incident and reflected light distribution of the license plate is obtained as follows:

[0101]

[0102] Using the relationship between the degree of polarization and the intensity information of different polarization directions, the Mueller matrix that characterizes the surface characteristics of the license plate can be obtained:

[0103]

[0104] The bidirectional reflection distribution function model of the laser reflected light after passing through the license plate surface is:

[0105]

[0106] At this time, the Stokes vector of the laser reflected light after the license plate can be expressed as:

[0107]

[0108] According to the Stokes vector and polarization calculation formula of the reflected light from the license plate, the polarization expression of the reflected light based on the laser incident target surface can be obtained as follows:

[0109]

[0110] Ignore the energy loss during the propagation process, that is, when P0=P t When , the polarization degree expression is:

[0111]

[0112] According to the above formula, the polarization degree of the license plate surface is affected by multiple parameters such as the incident angle, detection angle, beam divergence angle and medium refractive index. Therefore, using the above formula, the laser illuminator parameters and the observation angle of the color polarization camera can be set in a targeted manner to achieve higher polarization image acquisition and improve the polarization image effect.

[0113] The fifth step is image fusion processing;

[0114] The license plate image dataset is processed, fused and reconstructed using a visible light color camera and a visible light color polarization camera. First, the R, G and B three-channel color image data collected by the visible light color camera and the R, G and B three-channel DOLP image data collected by the visible light color polarization image are extracted. Secondly, the two images are divided into three channels of R, G and B for data feature point registration to achieve target alignment of images on different channels, and then the aligned R, G and B three-channel data are subjected to weighted average image fusion processing to form R, G and B channel polarization fusion images respectively. Finally, the R, G and B channel polarization fusion image data is used to reconstruct the color image, complete the color polarization fusion image processing, highlight the vehicle profile, personnel face, and license plate details, and achieve feature enhancement in the image detail area. The fusion process is as follows: Figure 4 shown.

[0115] Figure 5 and Figure 6 The comparison of the backlit license plate recognition effect of color polarization fusion image before and after laser fill light is given. Figure 3 Compared with images without polarization imaging, color polarization fusion images can effectively suppress background strong light interference, but the license plate numbers are difficult to identify, such as Figure 5 After the fill light is applied, the imaging effect of the license plate area can be significantly improved, and the license plate numbers are clearly visible. Figure 6 As shown, the above verifies the feasibility of the method of the present invention.

Claims

1. A strong background light license plate recognition system, characterized in that: The system consists of a laser illuminator (1), a polarizing plate (2), a visible light color camera (3), a visible light color polarization camera (4), an attenuation plate (5), a data acquisition system (6), a power module (7), a controller (8), an image processor (9), a laser beam expander lens group (10), a visible light imaging lens group (11), and a visible light polarization imaging lens group (12); The laser illuminator (1) is connected to a laser beam expander lens group (10), and a narrow-band polarizer (2) is installed in front of a light outlet of the laser beam expander lens group (10); The visible light color camera (3) is connected to the visible light imaging lens group (11); The visible light color polarization camera (4) is connected to the visible light polarization imaging lens group (12), and the attenuation plate (5) is installed in front of the visible light polarization imaging lens group (12); The controller (8) is connected to the laser illuminator (1), the visible light color camera (3), and the visible light color polarization camera (4) respectively; The visible light color camera (3) and the visible light color polarization camera (4) are respectively connected to a data acquisition system (6); The data acquisition system (6) is connected to the image processor (9); The controller (8) is connected to the laser illuminator (1), the visible light color camera (3), and the visible light color polarization camera (4); The image processor (9) is connected to the controller (8); The power module (7) is connected to the controller (8) and the image processor (9) respectively.

2. A method for a strong background light license plate recognition system according to claim 1, characterized in that: The method comprises the following steps: The first step is target capture and illumination; When a vehicle enters the visible light imaging lens group (11) and is focused on the field of view of the visible light color camera (3), the image processor (9) determines whether the vehicle has entered the field of view based on the abnormal light intensity area in the image. When the target vehicle enters the field of view, a trigger signal is quickly sent to the controller (8). After receiving the trigger signal from the visible light color camera (3), the controller (8) quickly triggers the laser illuminator (1) to continuously emit laser light, which is collimated and expanded by the laser beam expansion lens group (10) to illuminate the vehicle for rapid exposure until the target moves out of the field of view. The required laser illuminator power and the total power reflected back by the target can be calculated based on the license plate laser echo power model: The laser illuminator (1) has an output wavelength of 850 nm and emits p t The output power of the receiving module is A. r , then the laser emission light is reflected by the license plate surface, and the final total receiving power of the receiving module is: Where σ is the backscattering cross section; AREA≈π(Rθ B / 2) 2 is the beam coverage area at distance R; θ B is the beam divergence angle, where: the deviation angle between the reflected light edge ray and the main ray is set as θ c , the expression is θ c =2arctan(r2-r1 / R) and at the edge of the light θ c =θ B , r1 and r2 are the light emitting radius and the spot radius respectively; According to the above formula, combined with the actual imaging illumination distance and target size, the laser illuminator can be selected in a targeted manner to determine the laser illuminator power and beam divergence angle; The second step is camera exposure; The controller (8) receives a trigger signal from the visible light color camera (3), and synchronously triggers the visible light color camera (3) and the visible light color polarization camera (4) to work, and respectively focuses and images through the visible light imaging lens group (11) and the visible light polarization imaging lens group (12), so as to continuously photograph the target at the same frame rate, and synchronously obtain the target background visible light color image and the visible light polarization image; The third step is data collection; The visible light color camera (3) and the visible light color polarization camera (4) collect and store dual-channel images through a data collection system (6), wherein the visible light color camera (3) obtains color images in the 400nm to 900nm band; the visible light color polarization camera (4) obtains images in the 400nm to 900nm band with different polarization directions, i.e., 0°, 45°, 90°, and 135°, and the two-channel images are transmitted to an image processor (9) through a wired or wireless network transmission interface; Step 4: Polarization image processing; The image processor (9) calculates the polarization image, and the relationship between the intensity information of different polarization directions and the polarization degree is: According to the laser illumination, the bidirectional reflection distribution function that characterizes the incident and reflected light distribution of the license plate is obtained as follows: Using the relationship between the degree of polarization and the intensity information of different polarization directions, the Mueller matrix that characterizes the surface characteristics of the license plate can be obtained: The bidirectional reflection distribution function model of the laser reflected light after passing through the license plate surface is: At this time, the Stokes vector of the laser reflected light after the license plate can be expressed as: According to the Stokes vector and polarization calculation formula of the reflected light from the license plate, the polarization expression of the reflected light based on the laser incident target surface can be obtained as follows: Ignore the energy loss during the propagation process, that is, when P0=P t When , the polarization degree expression is: According to the above formula, the polarization degree of the license plate surface is affected by multiple parameters such as the incident angle, detection angle, beam divergence angle and medium refractive index. Using the above formula, the laser illuminator parameters and the color polarization camera observation angle can be set in a targeted manner to achieve higher polarization image acquisition and improve the polarization image effect. The fifth step is image fusion processing; A visible light color camera (3) and a visible light color polarization camera (4) are used to process, fuse and reconstruct the license plate image data set. First, the R, G and B three-channel color image data collected by the visible light color camera and the R, G and B three-channel DOLP image data collected by the visible light color polarization image are extracted. Secondly, the data feature points of the two images are aligned in the three same channels of R, G and B to achieve target alignment of images on different channels. Then, the aligned R, G and B three-channel data are subjected to weighted average image fusion processing to form R, G and B channel polarization fusion images respectively. Finally, the R, G and B channel polarization fusion image data are used to reconstruct the color image, complete the color polarization fusion image processing, highlight the vehicle profile, the face of the person and the license plate details, and achieve feature enhancement of the image detail area.

Images