Image acquisition and processing method

Abstract

The present invention provides a low-power image sensor component, a detector and an image acquisition and processing method, which relate to intelligent transportation and the Internet of Things, especially the field of fully automatic smart parking. In view of the problems that existing products produce visual blind spots in irregular parking situations, out-of-focus images or incomplete license plates, frequently start videos to dynamically track and identify vehicles entering and exiting during operation, consume a lot of power and must rely on city power supply or manual battery replacement for battery life, and the installation of products involves many stakeholders, resulting in great difficulty in coordination, long construction period, and high investment cost, a new detector mechanism with a low-power single-axis dual-chip directional geomagnetic and radar ranging module as the core and a low-power multi-channel video image switching and acquisition component are proposed. The combined application product does not need to be installed under the vehicle chassis and can achieve high-precision vehicle detection and high-quality image acquisition with extremely low power consumption near the berth edge, providing more choices and a more solid technical foundation for the development of the core business of fully automatic smart parking.

Description

Technical Field

[0001] The present invention relates to a fully automatic smart parking information device and related methods in the field of intelligent transportation and the Internet of Things, and in particular to a low-power image sensor component and detector for static traffic management, and an image acquisition and processing method for combined application thereof in different berth scenarios. Background Art

[0002] Fully automatic unmanned smart parking is an important development direction of static traffic management in the future, and accurate and reliable parking image acquisition, analysis and processing that are suitable for complex urban environments are important supports and prerequisites. At present, the following outstanding problems need to be overcome and improved: 1. Most existing parking license plate recognition devices use a single lens with a megapixel sensor image acquisition solution to capture clear and recognizable license plate images during the entry and exit of parking vehicles, but this solution is difficult to fully take into account various irregular parking situations, especially illegal parking across the line. There are many detection and recognition errors caused by visual blind spots, out-of-focus blur or missed shooting opportunities, which are often prominent; 2. Since parking spaces are usually located in busy urban areas with large traffic and are generally small and compact, the existing solutions basically set the image acquisition and processing equipment 6-50 meters outside the parking space. For example, the high-position and mid-position video acquisition solutions that are currently widely used, the installation of related products involves many merchants and a series of units and departments along the road, and there are many problems to varying degrees, such as difficulty in coordination and approval, long construction period, and high investment cost. The call for "solving parking problems within the parking space" is getting louder and louder; 3. Low-position video image acquisition and processing equipment generally consumes a lot of power. Existing products mainly rely on two methods: mains power supply or manual battery replacement for battery life. The new method of solar power supply has just begun to become popular; after analysis, the main consumption of electricity is the frequent activation of the daily video module for target detection and tracking, and / or the video module linkage caused by the false triggering of the built-in detection mechanism. For example, the existing new patent "A license plate cloud recognition camera, image sensor component and cleaning and power supply unit" (authorization announcement number: CN 214410276 U), although the radar ranging module is more power-saving than the single video, if used in combination with the solar power supply unit, the solution of relying on short-cycle radar detection in exchange for a better shooting time is relatively power-consuming, and in actual use, a single detection method is also more likely to cause false triggering and start the video. On the other hand, although the two-to-two orthogonal three-axis geomagnetic power consumption is extremely low and is widely used to detect vehicles on the ground of the parking space under the chassis, it is difficult to collect the license plate image at this location. At present, when the power supply is tight and energy conservation and emission reduction are advocated, there is an urgent need for a new low-power software and hardware solution to optimize and reduce the number of unnecessary video startups and improve the current high energy consumption of the equipment; 4. Existing new patents "A parking space type roadside device" (authorization announcement number: CN 206574141 U) and "A simplified parking space type roadside device" (authorization announcement number: CN 212809295 U) The disclosed technical solution does not disclose in detail the optimal structure, parameters and image acquisition and processing methods of image sensor components and detectors in different subdivided scenarios and their combined application equipment. It is necessary to further propose targeted solutions based on the current architecture to meet the needs of new business development and promote the upgrading of product equipment. Summary of the invention

[0003] The technical problem to be solved by the present invention is to avoid the shortcomings of the above-mentioned prior art and propose a low-power image sensor component, a detector and an image acquisition and processing method. The specific scheme is:

[0004] Design and use a low-power image sensor component suitable for license plate cloud recognition cameras or parking space-type roadside equipment, including: n groups of image sensor components and multiple-choice switching units connected to the signal source end by wire, where 1≤n≤8; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through the image sensor and its supporting circuit as a signal required for the processor to collect the image; when each image acquisition and processing device made of the image sensor component completes the relevant settings including the lens at a suitable place around the center of the boundary line at the end of each non-character parking space, or at each one-character parking space curb platform or curb wall, then the n groups of the image sensor components Used to switch and collect image data of high / low license plates of parking spaces related to the device, or license plates after standard / irregular parking, or license plates in front / rear of parking spaces, or during the process of vehicle parking / leaving; when the device completes relevant settings including lenses on the ceiling in the center of n non-character parking spaces in the room, the nth group of image sensor components is selected to collect the front license plate image data of the nth parking space; the lens settings include: selecting a fixed aperture fixed focal length lens that matches the corresponding parking space within the aperture F0.95-F2.8 and focal length 1.9mm-16mm, and setting the average center of the parking space-related license plates and pointing the central axis of the relevant lens to the spherical space within a radius of 0.6 meters from the center.

[0005] As an improvement, the above-mentioned image sensor component specifically includes: a polarizing plate that also serves as a lens viewfinder and can suppress the reflection of the license plate, a fixed aperture fixed focal length lens with an aperture of F0.95-F2.8 and a focal length of 1.9mm-16mm that matches the use environment, a lens holder including a neutral gray density dimming film and a high light transmittance white glass sheet and can be controlled to switch between the two through a glass sheet switching end, a 400nm-650 or 700nm visible light passing infrared light cutoff filter attached to the lens holder, a GalaxyCore GC0403 1 / 3 inch 768 x 576 resolution CMOS image sensor, automatic exposure and automatic white balance camera controller chip connected to the image sensor through an internal terminal; the external terminal of the camera controller chip is connected to the access terminal of the multiple-choice switching unit as the information source terminal of the image sensor component, and the control terminal of the multiple-choice switching unit is connected to the processor to provide the processor with image data required for multi-angle license plate recognition, evidence filing and / or vehicle occupancy image detection; when n=1, the multiple-choice switching unit may be unnecessary and the external terminal of the camera controller chip may be directly connected to the processor; the camera controller chip may be preset as required to crop the JPG image with a resolution of 640×480 in the center, left or right on the basis of the original frame to further save data traffic and bandwidth; the camera controller chip is directly connected to the processor through its external terminal using a USB bus or connected through the multiple-choice switching unit, and provides the processor with the required image information in the UVC standard protocol; the camera controller chip may be preset as required to crop the JPG image with a resolution of 640×480 in the center, left or right on the basis of the original frame 480 JPG pictures to further save data traffic and bandwidth; the camera controller chip is directly connected to the processor through its external terminal using a USB bus or through the multiple-choice switching unit, and provides the required image information to the processor using the UVC standard protocol.

[0006] Preferably, the camera controller chip is SN9C2788 produced by SONIX Technology; the multiple-choice switching unit is a low-resistance analog switch chip CH440 / 442 / 443 / 444 / 445 / 448 produced by Nanjing Qinheng Microelectronics.

[0007] In addition, a low-power detector is designed and used, which can be built-in and / or externally placed in a license plate cloud recognition camera or a parking space-type roadside device, and includes at least: a processor and a detector sensor connected thereto by wire; the detector sensor consists of a single-axis directional geomagnetic and a radar ranging module; the single-axis directional geomagnetic adopts a single-axis double-chip structure, that is, the sensitivity direction of the first single-axis directional geomagnetic points to a spherical space within a radius of 0.6 meters from the average center of the front or rear license plate of the vehicle in the parking space, and the sensitivity direction of the second single-axis directional geomagnetic is parallel or perpendicular to the sensitivity direction of the first single-axis directional geomagnetic, and a directional geomagnetic field electrical signal is transmitted to the processor through a matching circuit; the radar ranging module is a 24GHz or 77GHz working frequency. The invention relates to a special component comprising an antenna, a microwave signal processing part and an intermediate frequency signal processing part, which can measure the distance of a close-range target object in a certain frequency band; the antenna and the microwave signal processing part are 24GHz or 77GHz radar ranging sensors; the intermediate frequency signal processing part of the radar ranging module can complete the target distance calculation by the processor through an interface, or complete the target distance calculation through a built-in MCU and then output the result to the processor; the processor regularly uses the uniaxial directional geomagnetism to scan and detect magnetic field changes in a short period, and uses the radar ranging module to scan and detect changes in the distance of the target in a long period, and immediately complements and confirms the changes that meet the preset conditions, and outputs new berth status change trigger linkage information in a wired or wireless manner.

[0008] Preferably, the above-mentioned uniaxial directional geomagnetism is the TMR26xx series or TMR2102 tunnel magnetoresistance magnetic linear sensor of Jiangsu Multidimensional Technology, wherein the TMR26xx series linear sensor has a built-in signal conditioning circuit that can directly transmit the geomagnetic field electrical signal, while the TMR2102 must be equipped with an external signal conditioning circuit or operational amplifier circuit to transmit the geomagnetic field electrical signal.

[0009] At the same time, the present invention proposes a non-character parking space pole-type dual-camera internal inspection image acquisition and processing method, which is applicable to a license plate cloud recognition camera or a parking space type roadside device, comprising the following steps:

[0010] A-1: A parking limiter is set on the ground at a distance of 1.1-1.6 meters from the parking end boundary line of the berth to standardize the parking posture of the vehicle, and a pole-type dual-camera internal inspection image acquisition and processing device is set at a suitable place around the center of the end boundary line, which at least includes: a processor and an image sensor assembly including the first and second image sensor components and a two-choice switching unit connected to it by wire, and the first and second fill lights and their measurement and control units, a detector sensing unit and an online data communication interface; the processor is composed of a first processor connected to the detector sensing unit for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor assembly The first and second processors are interconnected and matched; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensing part is composed of the first and second single-axis directional geomagnetic and radar ranging modules; the image sensor component is a general term for a series of components including the lens that output the optical image captured by the lens through the image sensor and its supporting circuits as the signal required for the second processor to collect the image; the online data communication interface is 433, 470-510MHz, 2 .4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna, one or more combinations thereof; the first image sensor component on the pole for photographing low-position license plates has a lens height of 0.48±0.18 meters from the berth ground, the second image sensor component for photographing high-position license plates has a lens height of 0.78±0.18 meters from the berth ground, the first fill light with a depression angle of 20°±7° used in conjunction with the first image sensor component has a height of 0.78±0.3 meters from the berth ground, and the second fill light used in conjunction with the second image sensor component has a height of 0.68±0.2 horizontal forward fill light from the berth ground; the built-in detector sensor part for detecting changes in the state of vehicles entering and exiting the berth has a height of 0.68±0 .2 meters; the processor sends the relevant information of the berth to the back-end system and the cloud platform through the online data communication interface; the lens center axis of the first and second image sensor components, the sensitivity direction of the first uniaxial directional geomagnetic, and the beam center axis of the radar ranging module are basically parallel to the center axis of the left and right boundaries of the berth (deviation ≤ 10°), and the sensitivity direction of the second uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic;

[0011] B-1: Under normal circumstances, the second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the radar ranging module to scan and detect changes in the distance of targets in the berth, and after being awakened by its short-cycle timer, uses the first and second single-axis directional geomagnetic scans to detect changes in the magnetic field, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing is performed, and if there is a new berth status change that triggers linkage information, the second processor is awakened in a wired manner and relevant information is output to it, and after the output is completed, it enters a low-power sleep state;

[0012] C-1: If the second processor receives a new state change trigger linkage information of the berth after being awakened, the first and second image sensor components, the two-choice switching unit, and the first and second fill lights and their measurement and control units that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor are powered on in sequence, firstly the first image sensor component is selected to collect the low-position image data of the berth, and then the second image sensor component is selected to collect the high-position image data of the berth; after the image acquisition is completed, the ETC or RFID functional components that need to perform wireless communication business processing can be further executed, and then the second processor powers off the relevant components, packages and submits the image data to the background to implement license plate cloud recognition, or immediately implements license plate recognition on the spot (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best after the data is reported), controls the online data communication interface to complete the reporting of all business data including the image data, and then enters a low-power sleep state.

[0013] The present invention proposes a method for collecting and processing images of a vertical pole type dual-camera mixed inspection of a straight parking space, which is applicable to a license plate cloud recognition camera or a parking space type roadside device, and comprises the following steps:

[0014] A-2: Set up the average center of the rear license plate of the parking space for standard parking and irregular parking: Based on a certain number of representative parking examples on site, measure and record the three-dimensional coordinates of the center of the rear license plate when the parked vehicle does not exceed the rear line, and measure and record the three-dimensional coordinates of the center of the rear license plate when the parked vehicle exceeds the rear line; after the sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is calculated according to the rear license plate of standard parking and the rear license plate of irregular parking, and this is used as the average center of the rear license plate of standard parking and irregular parking;

[0015] B-2: A pole-type dual-camera mixed inspection image acquisition and processing device is set at a suitable place on the curb platform at a distance of 1±0.25 meters from the front line of the berth and 0.35±0.20 meters from the curb edge, which at least includes: a processor and an image sensor assembly including the first and second image sensor components and a two-choice switching unit connected thereto by wire, a fill light and its measurement and control unit, a detector sensing part and an online data communication interface; the processor is composed of a first processor connected to the detector sensing part for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor assembly for intercommunication; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensing part is composed of the first and second single-axis directional geomagnetic and the first and second radar ranging modules; the image sensor component is to output the optical image captured by the lens as the second image sensor through the image sensor and its supporting circuit. The processor is a general term for a series of components including a lens for collecting signals required for images; the online data communication interface is a combination of one or more of a 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna; the first image sensor component used to photograph the rear license plate of the front parking space in accordance with the regulations has a lens height of 0.77±0.2 meters from the parking space ground, the second image sensor component used to photograph the rear license plate of the front parking space in accordance with the regulations has a lens height of 0.87±0.2 meters from the parking space ground, and the fill light used in conjunction with the first and second image sensor components has a height of 0.87±0.2 meters from the parking space ground. .2 meters, the first and second uniaxial directional geomagnetic and first and second radar ranging modules of the built-in detector sensing part for detecting the change of the entry and exit status of the present berth and the front berth are 0.85±0.2 meters above the berth ground; the processor receives the trigger linkage wireless signal sent to it by the preset associated berth image acquisition and processing device through the online data communication interface, and sends the relevant information of the present berth to the preset associated berth image acquisition and processing device and / or the back-end system and cloud platform; the central axis of the lens of the first image sensor component and the central axis of the beam of the matching fill light, and the central axis of the beam of the first radar ranging module must all point to the spherical space within 0.6 meters of the average center radius of the license plate after the standard parking of the front berth; the central axis of the lens of the second image sensor component must point to the spherical space within 0.6 meters of the average center radius of the license plate after the irregular parking of the front berth.6 meters of spherical space; the sensitivity direction of the first uniaxial directional geomagnetic and the beam center axis of the second radar ranging module are basically horizontal and perpendicular to the left and right sides of the parking vehicle body (allowable deviation ≤ 15°), and the sensitivity direction of the second uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic;.

[0016] C-2: Under normal circumstances, the second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the first and second radar ranging modules to scan and detect the distance changes of the targets in the front berth and the current berth respectively; after being awakened by its short-cycle timer, the first processor uses the first and second single-axis directional geomagnetic scanning to detect magnetic field changes and uses the online data communication interface to receive the trigger linkage information sent to it by the preset front and rear berth image acquisition and processing equipment in a wireless manner, and uses multiple internal and external detectors to summarize various conditions of the associated berths, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing, if there is a new state change in the current berth triggers linkage information, it can be further executed and completed with ETC or RFID functional components and wireless communication business processing is required, and the relevant business information is wirelessly reported to the image acquisition and processing equipment of the rear berth through the online data communication interface; if there is a new state change in the front berth triggers linkage information, the second processor is awakened by wire and relevant information is output to it; after all output matters are completed, it enters a low-power sleep state;

[0017] D-2: If the second processor receives a new state change trigger linkage information of the front berth after being awakened, the first and second image sensor components, the two-choice switching unit, and the fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor are powered on, and the first image sensor component is selected to collect the front berth image data 1; then the second image sensor component is selected to collect the front berth image data 2; after the image acquisition is completed, the second processor powers off the relevant components, packages and submits the image data 1 and 2 to the background to implement license plate cloud recognition, or immediately implements license plate recognition on the spot (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, the background will use another set of candidate algorithms to further implement license plate cloud recognition to select the best after the data is reported), controls the online data communication interface to complete the reporting of all business data including the image data 1 and / or image data 2, and then enters a low-power sleep state.

[0018] The present invention proposes a method for collecting and processing non-character parking space ground-mounted single-camera mixed inspection images, which is applicable to a license plate cloud recognition camera, and comprises the following steps:

[0019] A-3: A ground-mounted single-camera mixed-inspection image acquisition and processing device is set at a suitable place around the center of the boundary line at the end of the parking space, which at least includes: a processor and an image sensor component and a fill light connected to it by wire and its measurement and control unit, the first and second detector sensing parts and an online data communication interface; the processor is composed of a first processor connected to the first detector sensing part for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the second detector sensing part and the image sensor component; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the first detector sensing part is composed of the first and second single-axis directional geomagnetic or integrated AMR three-axis geomagnetic, and the first radar ranging module; the The sensing part of the second detector is composed of the third and fourth uniaxial directional geomagnetic and the second radar ranging module; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor to collect the image; the online data communication interface is a combination of one or more of a 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna; the image sensor component for photographing the license plate has a lens located at the center of the terminal boundary line, and the left or right side is used for matching with an elevation angle of 20°±7° The first external detector sensor and the first processor for detecting the change of the parking space vehicle entry and exit status can add digital codes and / or QR codes on the surface of the product so that it can be installed at the center of the parking space parking front end boundary line 0.4±0 outside the parking space with the appearance of a parking space sign. .4 meters, and the built-in second detector sensor and the second processor are located in the image acquisition and processing equipment at the center of the terminal boundary line; the second processor sends the relevant information of the berth to the back-end system and the cloud platform through the online data communication interface; the lens center axis of the image sensor component, the sensitivity direction of the third uniaxial directional geomagnetic and the beam center axis of the second radar ranging module must all point to the direction of the vehicle entering and exiting and form an elevation angle of 20°±7° with the ground; the sensitivity direction of the first uniaxial directional geomagnetic and the beam center axis of the first radar ranging module are both facing the sky and basically perpendicular to the ground (deviation ≤10°); the sensitivity directions of the second and fourth uniaxial directional geomagnetic are respectively perpendicular or parallel to the sensitivity directions of the first and third uniaxial directional geomagnetic.

[0020] B-3: Under normal circumstances, the second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the first radar ranging module to scan and detect changes in the distance of targets outside the berth, and after being awakened by its short-cycle timer, uses the first and second single-axis directional geomagnetic scans to detect changes in the magnetic field, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing is performed, and if there is a new change in the state outside the berth that triggers linkage information, the second processor is awakened in a wired manner and relevant information is output to it, and after the output is completed, it enters a low-power sleep state;

[0021] C-3: If the second processor receives a new state change trigger linkage information outside the berth after being awakened, it will power on and pre-start the image sensor component and the fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor. At the same time, the second detector sensor part is used in a short period to detect the change of the distance and magnetic field of the target in the berth. Combined with the real-time change information provided by the first processor, the two internal and external detectors are mixed to summarize various situations, and the changes that meet the preset conditions are immediately complemented and verified. Comprehensive processing is given. If it is determined that a car is entering or leaving, the image sensor component is immediately used to detect the change of the distance and magnetic field of the target in the berth. Parking space image data is collected continuously at intervals until the vehicle completely stops or leaves; after image collection is completed, the second processor powers off relevant components, screens and packages the required selected image data according to the distance of the target and the law of magnetic field changes and submits them to the background to implement license plate cloud recognition, or immediately implements license plate recognition on the spot (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, after the data is reported, the background will use another set of candidate algorithms to further implement license plate cloud recognition to select the best one), controls the online data communication interface to complete the reporting of all business data including the selected image data, and then enters a low-power sleep state.

[0022] The present invention proposes a method for collecting and processing images of a ground-mounted dual-camera mixed inspection of a straight parking space, which is applicable to a license plate cloud recognition camera or a parking space type roadside device, and comprises the following steps:

[0023] A-4: Set the average center of the front and rear license plates in a straight parking space: Based on a certain number of representative parking examples on site, measure and record the 3D coordinates of the center of the front license plate when the front of the parking vehicle does not exceed the front line, and measure and record the 3D coordinates of the center of the rear license plate when the front of the parking vehicle exceeds the front line; after sample collection is completed, the average value of the relevant samples in the 3D coordinate system of the parking space is calculated according to the front and rear license plates, and this is used as the average center of the front and rear license plates;

[0024] B-4: A ground-mounted dual-camera mixed inspection image acquisition and processing device is arranged on the front parking space curb wall outside the front line where the front line of the parking space and the curb line intersect vertically, and the device at least comprises: a processor and an image sensor assembly including the first and second image sensor components and a two-choice switching unit connected thereto by wire, the first and second fill lights and their measurement and control units, a detector sensing part and an online data communication interface; the processor is composed of a first processor connected to the detector sensing part for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor assembly for intercommunication; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensing part is composed of the first and second single-axis directional geomagnetic and the first and second radar ranging modules; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor to collect the image; the online data communication interface is 433, A combination of one or more of a 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, a NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna; a first image sensor component for photographing the front license plate of the parking space and a second fill light paired with the second image sensor component are located at the far end of the image acquisition and processing device, a second image sensor component for photographing the rear license plate of the parking space in front and a first fill light paired with the first image sensor component are located at the far end of the image acquisition and processing device The proximal end of the built-in detector sensor for detecting the change of the entry and exit status of the vehicle between the present berth and the front berth is located at the proximal end of the image acquisition and processing device; the processor receives the trigger linkage wireless signal sent to it by the preset associated berth image acquisition and processing device through the online data communication interface, and sends the relevant information of the present berth to the preset associated berth image acquisition and processing device and / or the back-end system and the cloud platform; the central axis of the lens of the first image sensor component and the central axis of the light beam of the paired first fill light, the sensitivity direction of the first uniaxial directional geomagnetic, and the central axis of the beam of the first radar ranging module must all point to the spherical space within 0.6 meters of the average center radius of the front license plate of the present berth; the central axis of the lens of the second image sensor component and the central axis of the light beam of the paired second fill light, and the central axis of the beam of the second radar ranging module must all point to the spherical space within 0.6 meters of the average center radius of the rear license plate of the front berth, and at the same time, the sensitivity direction of the second uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic;

[0025] C-4: Under normal circumstances, the second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the first and second radar ranging modules to scan and detect the distance changes of the targets in the current berth and the front berth respectively; after being awakened by its short-cycle timer, the first processor uses the first and second single-axis directional geomagnetic scanning to detect magnetic field changes and uses the online data communication interface to receive the trigger linkage information sent to it by the preset front and rear berth image acquisition and processing equipment in a wireless manner, and uses multiple internal and external detectors to summarize various situations of the associated berths, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing, if there is a new state change in the current berth to trigger linkage information, or if a "request for assistance in capturing the rear license plate" trigger linkage information is received from the front berth, the second processor is awakened in a wired manner and relevant information is output to it; if there is a state change in the front berth to trigger linkage information, it is reported to the image acquisition and processing equipment of the front berth in a wireless manner through the online data communication interface; after all output matters are completed, it enters a low-power sleep state;

[0026] D-4: If the second processor receives a new state change trigger linkage information of the current berth after being awakened, the first image sensor component, the two-choice switching unit, and the first fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor, and the first image sensor component is selected to collect the image data of the current berth; if the trigger linkage information of "requesting assistance in capturing the rear license plate" is received from the front berth, the second image sensor component, the two-choice switching unit, and the second fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor, and the second image sensor component is selected Collect the front parking space image data; after the image collection is completed, the ETC or RFID functional components and the wireless communication business processing can be further executed, and then the second processor powers off the relevant components, packages and submits the parking space image data to the background to implement license plate cloud recognition, or immediately implements license plate recognition on the spot (cloud recognition or on-site recognition has no results, or the results come out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best), controls the online data communication interface to complete the reporting of all business data including the parking space image data, and then enters a low-power sleep state.

[0027] The present invention proposes a method for collecting and processing images of a ceiling-mounted multi-camera external inspection of an indoor non-character parking space, which is applicable to a license plate cloud recognition camera or a parking space type roadside device, and comprises the following steps:

[0028] A-5: Setting the average center of the front license plate of indoor non-character parking spaces: First, install a parking limiter at a suitable place before the parking space end boundary line of the parking space, so that the distance between the front or rear of the parked vehicle and the front boundary line of the parking space is less than 1 meter. Then, after the installation is completed and put into use, based on a certain number of representative parking examples on site, measure and record the three-dimensional coordinates of the center of the license plate of the parked vehicle near the front boundary line of the parking space; after the sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is statistically obtained, which is used as the average center of the front license plate;

[0029] B-5: A ceiling-mounted multi-camera external inspection image acquisition and processing device (2≤n≤8) with lenses corresponding to parking spaces is set on the ceiling in the center of the n parking space area, which at least includes: a 9th processor and an image sensor assembly including 1st to nth image sensor components and an n-to-one switching unit connected to it by wire, a fill light and its measurement and control unit, and a 9th online data communication interface; in addition, n external detectors need to be further set near the front boundary line of the n parking spaces, which at least include: an nth processor and an nth detector sensor part connected to it by wire, and an nth online data communication interface. The detector can add digital codes and / or QR codes on its product surface so that it can be set at the center of the front boundary line of the parking space with the appearance of a parking space identification plate, 0.4±0 outside the parking space. .4 meters; the nth detector sensing part is composed of the n_1th and n_2th single-axis directional geomagnetic and nth radar ranging modules; the processor is a SOC transceiver integrated single-chip microcomputer, a microprocessor, a RISC-V core microcontroller, an ARM or a DSP processor; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through an image sensor and its supporting circuits as a signal required for the 9th processor to collect the image; the online data communication interface includes a wired and / or wireless data communication interface, the wired data communication interface includes a UART, RS232 / RS485 / RS422, USB or TCP / IP network interface, and the wireless data communication interface is 433, 470-510MHz, 2 .4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna, one or more combinations thereof; the 9th processor sends the relevant information of the parking spaces in this area to the back-end system and the cloud platform through the 9th online data communication interface; the lens center axis of the nth image sensor component, the sensitivity direction of the n_1th uniaxial directional geomagnetic, and the beam center axis of the nth radar ranging module must all point to the spherical space within 0.6 meters of the average center radius of the license plate at the front end of the nth parking space; the sensitivity direction of the n_2th uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the n_1th uniaxial directional geomagnetic;

[0030] C-5: Under normal circumstances, the image sensor component is in a power-off shutdown state; the 9th processor listens to and receives the berth status change trigger linkage information sent by the n external detectors through the 9th online data communication interface in a short period; the nth processor uses the nth radar ranging module to scan and detect the distance changes of the target in the berth after being awakened by its long-period timer, and uses the n_1th and n_2th single-axis directional geomagnetic scans to detect magnetic field changes after being awakened by its short-period timer, and performs instant complementary verification and comprehensive processing on the changes that meet the preset conditions. If there is a new berth status change triggering linkage information, the relevant information is reported to the 9th processor through the nth online data communication interface, and after the reporting is completed, it enters a low-power sleep state; D-5: If the 9th processor receives a new state change trigger linkage information for the nth berth, it powers on the nth image sensor component, the n-to-one switching unit, and the fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor, and selects the nth image sensor component to collect berth image data; after the image acquisition is completed, the components with ETC or RFID functions and need to perform wireless communication business processing can be further executed, and then the 9th processor powers off the relevant components, packages and submits the image data to the background to implement license plate cloud recognition, or immediately implements license plate on-site recognition (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best), and controls the 9th online data communication interface to complete the reporting of all business data including the image data.

[0031] Finally, as an improvement of the present invention, in any of the above-mentioned image acquisition and processing methods, the image sensor component includes: a polarizing plate that also serves as a lens viewfinder and can suppress license plate reflections, a fixed aperture fixed focal length lens with an aperture of F1.2-F2.8 and a focal length of 1.9mm-16mm that matches the use environment, a lens mount that includes a neutral gray density dimming film and a high-transmittance white glass and can be controlled to switch between the two through a glass switching end, a 400nm-650 or 700nm visible light-passing infrared light cutoff filter attached to the lens mount, a GalaxyCore GC0403 1 / 3 inch 768 x576 resolution CMOS image sensor, and an automatic exposure and automatic white balance camera controller chip connected to the image sensor through an internal terminal, and the external terminal of the chip can be used as a signal source terminal of the image sensor component to access the switching unit or the processor.

[0032] Compared with the prior art, the present invention provides a low-power image sensor component, a detector and an image acquisition and processing method, which have the following technical effects: 1. A new multi-lens image sensor component configuration is proposed, which achieves full coverage of the field of view by complementing multiple perspectives, solves the problem of berth visual blind spots to the greatest extent, greatly improves the success rate of image data acquisition, screening and recognition processing in complex scenes, and has strong product robustness and greater flexibility; 2. Through the organic combination of the new image sensor component and the detector, the berth problem can be basically solved autonomously within the berth and within a range of 0.5 meters at the farthest from the berth edge, and the product can be "tailored" to select appropriate equipment parameters and installation methods, which wins more initiative for the project construction party and the operator, shortens the construction period, and saves investment and operation and maintenance costs; 3. In view of the fact that the average current of the video startup is at the level of hundreds of milliamperes and the working time is at least more than ten seconds, while the average current of the radar startup is close to the level of hundreds of milliamperes but the working time is less than 0 .5 seconds, the average startup current of directional geomagnetism is at the level of 10 mA, and the working time is less than 0.05 seconds. Compared with any of the two-to-two orthogonal three-axis geomagnetic solutions of AMR or GMR used in existing parking space detectors, the single-axis directional geomagnetism with the unique advantages of parking orientation directional selection and signal conditioning has stronger targeting, longer detection distance, and better sensing sensitivity. In general, directional geomagnetic scanning uses a short cycle, radar scanning uses a long cycle, and the signal change reaches the threshold and is immediately processed within the device and between devices for complementary verification before it is possible to activate the start video. This series of measures not only ensures the accuracy of vehicle detection, but also greatly reduces the number of video startups; in addition, the commonly used two-to-two orthogonal three-axis geomagnetism is installed under the chassis of the vehicle to detect the presence of the vehicle but cannot collect images. It is improved to use a single-axis directional geomagnetism installed at the side of the parking space that is not covered by the vehicle to detect the vehicle in a directional and low-power manner and further collect images, which solves the long-standing problem of high power consumption of low-position video equipment and the need to move the vehicle to install the detection equipment that has plagued the industry; 4 .For five mainstream business scenarios, based on the original product architecture, a new component structure, combined application parameters, and trigger linkage processes are proposed to supplement and improve the deficiencies of the original solution. While consolidating and improving the performance of existing products, it also provides more choices and a more solid technical foundation for the development of the core business of fully automatic smart parking. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the structural principle of the image sensor assembly and detector of the present invention integrated with the existing product architecture;

[0034] Figure 2 It is a schematic diagram of the structural principle of the image sensor component in the present invention;

[0035] Figure 3It is a plan view of the installation layout of the non-font berth field equipment in the first embodiment of the present invention;

[0036] Figure 4 It is a schematic diagram of the structural principle of the image acquisition and processing device in the first embodiment of the present invention;

[0037] Figure 5 It is a schematic diagram of the spatial positions of various elements in the first embodiment of the present invention;

[0038] Figure 6 Schematic diagram of various parking examples of the straight-line berth in the second embodiment of the present invention;

[0039] Figure 7 It is a schematic diagram of the structural principle of the image acquisition and processing device in the second embodiment of the present invention;

[0040] Figure 8 It is a schematic diagram of the spatial positions of various elements in the second embodiment of the present invention;

[0041] Fig. 9 It is a schematic diagram of the spatial positions of various elements of the non-font berth site in the third embodiment of the present invention;

[0042] Fig.10 It is a schematic diagram of the structural principle of the image acquisition and processing device in the third embodiment of the present invention;

[0043] Fig.11 Schematic diagram of various parking examples of the straight-line berth in the fourth embodiment of the present invention;

[0044] Fig.12 It is a schematic diagram of the structural principle of the image acquisition and processing device in the fourth embodiment of the present invention;

[0045] Fig.13 It is a schematic diagram of the spatial positions of various elements in the fourth embodiment of the present invention;

[0046] Fig.14 This is a schematic diagram of the spatial positions of various elements of the indoor non-font berth in the fifth embodiment of the present invention;

[0047] Fig.15 It is a schematic diagram of the structural principle of the image acquisition and processing device in Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0048] The present invention will firstly introduce in detail the structural composition characteristics of a low-power image sensor component and a detector, and then provide corresponding optimized vehicle image acquisition and processing flows in the form of combined application embodiments for five different business scenarios, and give detailed explanations.

[0049] like Figure 1The schematic diagram of the structure principle of the low-power image sensor assembly and detector of the present invention integrated with the existing product architecture, in which the image sensor assembly 19 includes: n groups of image sensor components 19n and a multiple-choice switching unit 190 connected to the signal source end by wire, wherein 1≤n≤8; the image sensor components 19n are now combined with Figure 2 As shown, it is a general term for a series of components including a lens that outputs the optical image captured by the lens through an image sensor and its supporting circuits as a signal required for the processor 1020 to collect the image, including: a polarizing plate 19n1 that also serves as a lens viewfinder and can suppress the reflection of the license plate, a fixed aperture fixed focal length lens 19n2 with an aperture of F0.95-F2.8 and a focal length of 1.9mm-16mm that matches the use environment, a lens holder 19n3 that includes a neutral gray density dimming film and a high-transmittance white glass sheet and can be controlled to switch between the two through a glass sheet switching end, a 400nm-650 or 700nm visible light passing infrared light cutoff filter attached to the lens holder, and a GalaxyCore GC0403 1 / 3 inch 768 x 576 resolution CMOS image sensor 19n4, and an automatic exposure and automatic white balance camera controller chip 19n5 connected to the image sensor 19n4 through an internal terminal; the external terminal of the camera controller chip 19n5 is connected to the access terminal of the multiple-choice switching unit 190 as the signal source terminal of the image sensor component 19n, and the control terminal of the multiple-choice switching unit 190 is connected to the processor 1020 to provide the processor 1020 with image data required for multi-angle license plate recognition, evidence collection and filing, and / or vehicle occupancy image detection; in particular, when n=1, the multiple-choice switching unit 190 is not required, and the external terminal of the camera controller chip 19n5 is directly connected to the processor 1020.

[0050] Furthermore, the camera controller chip 19n5 can preset, as required, to crop the JPG image with an output resolution of 640×480 in a centered, left-biased or right-biased manner based on the original frame to further save data traffic and bandwidth; the camera controller chip 19n5 is directly connected to the processor 1020 through its external terminal using a USB bus or is connected through the one-of-a-kind switching unit 190, and provides the required image information to the processor 1020 using the UVC standard protocol.

[0051] Preferably, the camera controller chip 19n5 is SN9C2788 from SONIX Technology Co., Ltd.; the multiple-choice switching unit 190 is a low-resistance analog switch chip CH440 / 442 / 443 / 444 / 445 / 448 from Nanjing Qinheng Microelectronics Co., Ltd.

[0052] In addition, the low-power detector proposed in the present invention can be built-in and / or externally placed in an existing license plate cloud recognition camera or parking space type roadside equipment, which at least includes: a processor 1020 and a detector sensor unit 16 connected thereto by wire; Figure 1 As shown, the detector sensing part 16 is composed of a single-axis directional geomagnetic 161 and a radar ranging module 162; the single-axis directional geomagnetic 161 adopts a single-axis double-chip structure, that is, the sensitivity direction of the first single-axis directional geomagnetic points to the average center radius of the front or rear license plate of the vehicle in the parking space 0 .6 meters in a spherical space, the second uniaxial directional geomagnetic field has a sensitivity direction parallel or perpendicular to the sensitivity direction of the first uniaxial directional geomagnetic field, and transmits a directional geomagnetic field electric signal to the processor 1020 through a supporting circuit; the radar ranging module 162 is a special component that works in the 24GHz or 77GHz frequency band and can measure the distance of close-range target objects, including an antenna and a microwave signal processing part and an intermediate frequency signal processing part; the antenna and microwave signal processing part are 24GHz or 77GHz radar ranging sensors; the intermediate frequency signal processing part of the radar ranging module can complete the target distance calculation through the interface by the processor, or complete the target distance calculation through the built-in MCU and then output the result to the processor 1020; the processor 1020 regularly uses the uniaxial directional geomagnetic field 161 to scan and detect magnetic field changes in a short period, and uses the radar ranging module 162 to scan and detect target distance changes in a long period, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing, and outputs new berth status change trigger linkage information in a wired or wireless manner.

[0053] Preferably, the above-mentioned uniaxial directional geomagnetism is the TMR26xx series or TMR2102 tunnel magnetoresistive magnetic linear sensor of Jiangsu Multidimensional Technology, wherein the TMR26xx series linear sensor has a built-in signal conditioning circuit that can directly transmit the geomagnetic field electrical signal, while the TMR2102 must be equipped with an external signal conditioning circuit or operational amplifier circuit to transmit the geomagnetic field electrical signal.

[0054] Based on the above preferences, the following five combined application embodiments are used for further detailed description.

[0055] Embodiment 1: Non-character parking space pole-type dual-camera internal inspection image acquisition and processing This embodiment aims at the high cost and high coordination difficulty dilemma brought about by the existing products in the industry, which are mostly installed and deployed 5-50 meters outside the parking space to collect images for vehicle detection and identification. A new solution to the image acquisition problem in the parking space is proposed, which is suitable for parking spaces with a depth of not less than 5 meters and can be used for upgrading and transformation of existing license plate cloud recognition cameras or parking space-type roadside equipment. The main working principle is:

[0056] A-1: If Figure 3As shown, a parking limiter is set on the ground 1.1-1.6 meters away from the parking end boundary line to standardize the parking posture of the vehicle, and a pole-type dual-camera internal inspection image acquisition and processing device is set at a suitable place around the center of the end boundary line. Figure 4 As shown, it at least includes: a processor 1020 and an image sensor assembly 19 including a first image sensor component 191, a second image sensor component 192 and a two-choice switching unit 190 connected thereto by wire, a first fill light B1 and its measurement and control unit 171, a second fill light B2 and its measurement and control unit 172, a detector sensor unit 16 and an online data communication interface 121; the processor 1020 is composed of a first processor 10201 connected to the detector sensor unit 16 through a sensor interface for comprehensive analysis of whether there is a car or not, and a second processor 10202 connected to the image sensor assembly 19 and the fill light and its measurement and control unit 17 through a video terminal and a measurement and control terminal; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensor unit 16 is composed of a first uniaxial directional geomagnetic M1 1611, a second uniaxial directional geomagnetic M2 1612 and radar ranging module R 162; the first image sensor component 191 and the second image sensor component 192 are the general term for a series of components including lenses that respectively output the optical images captured by the lens J1 and the lens J2 through their respective image sensors and their supporting circuits as signals required for the second processor 10202 to collect images; the online data communication interface 121 is a 4G wireless communication module 1211 and its transceiver antenna 1221; Figure 5 As shown, the first image sensor component 191 on the pole for photographing the low-position license plate has a lens J1 at a height of 0.48 meters from the berth ground, the second image sensor component 192 for photographing the high-position license plate has a lens J2 at a height of 0.78 meters from the berth ground, and the fill light B1 with a depression angle of 20°±7° used in conjunction with the first image sensor component 191 is at a height of 0. .78 meters, the fill light B2 used in conjunction with the second image sensor component 192 is 0.68 meters above the berth ground and provides horizontal forward fill light; the built-in detector sensor 16 used to detect changes in the entry and exit status of vehicles in the berth is 0.73 meters above the berth ground; the second processor 10202 sends the relevant information of the berth to the back-end system and the cloud platform through the online data communication interface 121; the central axes of the lenses J1 and J2 of the first and second image sensor components, the sensitivity direction of the first uniaxial directional geomagnetic M1, and the central axis of the beam of the radar ranging module R are basically parallel to the central axes of the left and right boundaries of the berth (deviation ≤10°), and the sensitivity direction of the second uniaxial directional geomagnetic M2 is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic M1;

[0057] B-1: Under normal circumstances, the second processor 10202 is in a low-power sleep state, and the image sensor component 19 is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor 10201 uses the radar ranging module R 162 to scan and detect changes in the distance of targets in the berth, and after being awakened by its short-cycle timer, uses the first uniaxial directional geomagnetic M1 and the second uniaxial directional geomagnetic M2 to scan and detect changes in the magnetic field, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing is performed, and if there is a new berth status change that triggers linkage information, the second processor 10202 is awakened in a wired manner and relevant information is output to it. After the output is completed, it enters a low-power sleep state;

[0058] C-1: If the second processor 10202 receives a new state change trigger linkage information after being awakened, it will successively power on the first image sensor component 191, the second image sensor component 192, the two-choice switching unit 190, and the fill light B1, B2 and its measurement and control unit that perform ambient light measurement to switch the slide in advance or light up the fill light when the light is poor. First, the first image sensor component 191 is selected to collect the low-position image data of the berth, and then the second image sensor component 192 is selected to collect the high-position image data of the berth; after the image acquisition is completed, The second processor 10202 powers off relevant components, packages and submits the image data to the background to implement license plate cloud recognition, or implements license plate recognition on-site in real time (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, the background will use another set of candidate algorithms to further implement license plate cloud recognition to select the best one after the data is reported), controls the 4G wireless communication module 1211 and its transceiver antenna 1221 in the online data communication interface 121 to complete the reporting of all business data including the image data, and then enters a low-power sleep state.

[0059] It is worth mentioning that although the above introduction is the working principle of the single-row non-shaped berth equipment, it is also applicable to back-to-back double-row non-shaped berths. It only requires adding a dual-camera combination and a radar ranging module in a symmetrical position on the back of the dual-camera internal inspection image acquisition and processing equipment, and changing the two-to-one switching unit to a four-to-one switching unit, so as to achieve reuse and reduce equipment investment. I will not go into details here.

[0060] Embodiment 2: Image acquisition and processing of double-camera mixed inspection of vertical poles in straight parking spaces. This embodiment aims at the dilemma that most existing products in the industry, especially ETC video poles (a type of parking space type roadside equipment), are heavily dependent on mains power supply or manual regular replacement of lithium batteries for battery life due to high power consumption for image acquisition. By comprehensively applying the low-power image sensor components and detectors proposed in the present invention, the high energy consumption problem caused by frequent startup of the video part of the equipment is greatly reduced under the premise that the parking space vehicle detection performance remains unchanged or even better, and the common visual blind spot problem is solved, which is conducive to accelerating the promotion and deployment of existing license plate cloud recognition cameras or parking space type roadside equipment. The main working principle is:

[0061] A-2: If Figure 6 As shown, the average center of the rear license plate of the parking space with standard parking and irregular parking is set: based on a certain number of representative parking examples on site, the three-dimensional coordinates of the center of the rear license plate in case 1 and case 2 where the parked vehicle does not exceed the rear line of the vehicle are measured and recorded, and the three-dimensional coordinates of the center of the rear license plate in case 3 where the parked vehicle exceeds the rear line of the vehicle are measured and recorded; after the sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is obtained by counting the rear license plate of the standard parking and the rear license plate of the irregular parking, and this is used as the average center of the rear license plate of the standard parking and the irregular parking;

[0062] B-2: A pole-type dual-camera mixed inspection image acquisition and processing device is installed on the curb platform 0.35 meters away from the curb, 1 meter away from the front line of the berth. Figure 7As shown, it at least includes: a processor 1020 and an image sensor assembly 19 including a first image sensor component 191, a second image sensor component 192 and a two-choice switching unit 190 connected thereto by wire, a fill light and its measurement and control unit 17, a detector sensing unit 16 and an online data communication interface 12; the processor 1020 is composed of a first processor connected to the detector sensing unit 16 through a sensor interface for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor assembly 19 and the fill light and its measurement and control unit 17 through a video terminal and a measurement and control terminal; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensing unit 16 is composed of a first uniaxial directional geomagnetic M1 1611, a second uniaxial directional geomagnetic M2 1612, a first radar ranging module R1 1621, and a second radar ranging module R2 1622; the image sensor components 192, 192 are the general term for a series of components including the lenses J1, J2 that output the optical images captured by the lenses J1, J2 through their respective image sensors and their supporting circuits as signals required for the second processor 10202 to collect images; the online data communication interface 12 is a combination of a 2.4GHz wireless transceiver 1212 and its transceiver antenna 1222, and a 4G wireless communication module 1211 and its transceiver antenna 1221; Figure 8As shown, the first image sensor component 191 for photographing the rear license plate of the front parking space in accordance with regulations has its lens J1 at a height of 0.95 meters from the parking space ground, the second image sensor component for photographing the rear license plate of the front parking space in accordance with regulations has its lens at a height of 1.05 meters from the parking space ground, the fill light used in conjunction with the first and second image sensor components is at a height of 1.05 meters from the parking space ground, the built-in detector sensor unit 16 for detecting the change in the entry and exit status of the vehicle in the present parking space and the front parking space has its first uniaxial directional geomagnetic M1 1611, the second uniaxial directional geomagnetic M2 1612 and the first radar ranging module R1 1621, the second radar ranging module R2 1622 at a height of 0.75 meters from the parking space ground; the processor 10201 uses the 2 .4GHz wireless transceiver 1212 and its transceiver antenna 1222 trigger linkage wireless signal transmission and reception with preset associated parking space image acquisition and processing equipment, and the processor 10202 sends the relevant information of the parking space to the back-end system and cloud platform through the 4G wireless communication module 1211 and its transceiver antenna 1221; the central axis of the lens J1 of the first image sensor component 191 and the central axis of the light beam of the matching fill light B, and the central axis of the beam of the first radar ranging module R1 must all point to the spherical space within 0.6 meters of the average center radius of the license plate after the front parking space is parked in a standardized manner; the central axis of the lens J2 of the second image sensor component 192 must point to the spherical space within 0.6 meters of the average center radius of the license plate after the front parking space is parked in an irregular manner .6 meters; the sensitivity direction of the first uniaxial directional geomagnetic M1 and the beam center axis of the second radar ranging module R2 are substantially horizontal and perpendicular to the left and right sides of the parking vehicle body (allowable deviation ≤ 15°), and the sensitivity direction of the second uniaxial directional geomagnetic M2 is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic M1;

[0063] C-2: Under normal circumstances, the second processor 10202 is in a low-power sleep state, and the image sensor component 19 is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor 10201 uses the first radar ranging module R1 and the second radar ranging module R2 to scan and detect the distance changes of the front berth and the current berth targets respectively; after being awakened by its short-cycle timer, it uses the first uniaxial directional geomagnetic M1 and the second uniaxial directional geomagnetic M2 to scan and detect the magnetic field changes and uses the 2.4GHz wireless transceiver 1212 and its transceiver antenna 1222 to receive the trigger linkage information sent to it by the preset front and rear berth image acquisition and processing equipment in a wireless manner, and mixes multiple internal and external detectors to summarize various situations of the associated berths, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing, if there is a new state change in the current berth triggers linkage information, it has ETC or RFID functional components (such as Figure 7 and Figure 8As shown, in this embodiment, the image acquisition and processing device further includes a microwave radio frequency communicator 11, which has a built-in national standard ETC 5.8GHz transceiver and wake-up device 111 and a transceiver antenna 112. At the same time, there is also an ETC vehicle-mounted electronic tag on the vehicle) and wireless communication business processing can be further executed according to the existing business process. The relevant business information is wirelessly reported to the image acquisition and processing device of the rear berth through the 2.4GHz wireless transceiver 1212 and its transceiver antenna 1222; if there is a new state change in the front berth that triggers linkage information, the second processor 10202 is awakened in a wired manner and relevant information is output to it; after all output matters are completed, it enters a low-power sleep state;

[0064] D-2: If the second processor 10202 receives a new state change trigger linkage information of the front berth after being awakened, the first and second image sensor components 191 and 192, the two-choice switching unit, and the fill light and its measurement and control unit 17 that perform ambient light measurement to switch the slide in advance or light up the fill light when the light is not good are powered on in sequence, and the first image sensor component 191 is selected to collect the front berth image data 1; then the second image sensor component 192 is selected to collect the front berth image data 2; after the image acquisition is completed, the second image sensor component 192 is powered on. The processor powers off relevant components, packages and submits the image data 1 and 2 to the background to implement license plate cloud recognition, or immediately implements license plate recognition on-site (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, the background will use another set of candidate algorithms to further implement license plate cloud recognition to select the best one after the data is reported), controls the 4G wireless communication module 1211 and its transceiver antenna 1221 to complete the reporting of all business data including the image data 1 and / or image data 2, and then enters a low-power sleep state.

[0065] Embodiment 3: Non-character parking space ground-mounted single-camera mixed inspection image acquisition and processing This embodiment is aimed at the short and compact parking spaces below 4.7 meters commonly seen in bustling urban areas. Existing products in the industry can only use high-position aerial acquisition of images to solve the various on-site construction and later maintenance problems. Instead, a hybrid solution is proposed that uses an external detector on the parking space edge to trigger first to start the equipment to continuously capture, and a built-in detector to dynamically track changes. This solution is obviously easier to install and deploy quickly. Its working principle is as follows:

[0066] A-3: If Fig. 9 As shown in the figure, a ground-mounted single-camera mixed inspection image acquisition and processing device is set at a suitable place around the center of the parking space end boundary line. Fig.10, which at least includes: a processor 1020 and an image sensor component 191 connected to it by wire, a fill light B and its measurement and control unit 17, a first detector sensor unit 161, a second detector sensor unit 162 and an online data communication interface 121; the processor 1020 is composed of a first processor 10201 connected to the first detector sensor unit 161 for comprehensive analysis and processing of whether there is a car or not, and a second processor 10202 connected to the second detector sensor unit 162 and the image sensor component 191, which are connected and communicated by wire; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the first detector sensor unit 161 is composed of a first uniaxial directional geomagnetic M1 16111, a second uniaxial directional geomagnetic M2 16112, and a first radar ranging module R1 The second detector sensing portion 162 is composed of a third uniaxial directional geomagnetic M3 16211, a fourth uniaxial directional geomagnetic M4 16212 and the second radar ranging module 1622; the image sensor component 191 is a general term for a series of components including the lens J, which outputs the optical image captured by the lens J through the image sensor and its supporting circuit as the signal required for the second processor 10202 to collect the image; the online data communication interface 121 is a 4G wireless communication module 1211 and its transceiver antenna 1221; the image sensor component 191 for photographing the license plate has a lens J located at the center of the end boundary line, and the left or right side is a supporting fill light B with an elevation angle of 20°±7°; the external first detector sensor part 161 and the first processor 10201 for detecting the change of the parking vehicle in and out state can add digital codes and / or two-dimensional codes on the surface of the product so that it can be installed at the center of the parking head end boundary line of the parking space with the appearance of a parking space identification plate, 0.4 meters, and the built-in second detector sensor 162 and the second processor 10202 are located in the image acquisition and processing equipment at the center of the terminal boundary line; the second processor 10202 sends the relevant information of the berth to the back-end system and the cloud platform through the 4G wireless communication module 1211 and its transceiver antenna 1221; the image sensor component 191, its lens J center axis, the third uniaxial directional geomagnetic M3 16211, its sensitivity direction and the second radar ranging module R2 1622, its beam center axis must all point to the direction of the vehicle entering and exiting and form an elevation angle of 20°±7° with the ground; the first uniaxial directional geomagnetic M1 16111, its sensitivity direction, and the first radar ranging module R1 1612, its beam center axis are all facing the sky and are basically perpendicular to the ground (deviation ≤10°); the second uniaxial directional geomagnetic M216112, the fourth uniaxial directional geomagnetic M4 16212, whose sensitivity direction is respectively perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic M1 16111 and the third uniaxial directional geomagnetic M3 16211;.

[0067] B-3: Under normal circumstances, the second processor 10202 is in a low-power sleep state, and the image sensor component 191 is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor 10201 uses the first radar ranging module R1 to scan and detect changes in the distance of targets outside the berth, and after being awakened by its short-cycle timer, uses the first and second uniaxial directional geomagnetic M1 and M2 to scan and detect changes in the magnetic field, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing is performed, and if there is a new change in the state outside the berth that triggers linkage information, the second processor 10202 is awakened in a wired manner and relevant information is output to it. After the output is completed, it enters a low-power sleep state;

[0068] C-3: If the second processor 10202 receives a new state change trigger linkage information outside the berth after being awakened, it will power on and pre-start the image sensor component 191 and the fill light B and its measurement and control unit 17 that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor. At the same time, the second detector sensor part 162 is used in a short period to detect the change of the distance and magnetic field of the target in the berth. Combined with the real-time change information provided by the first processor 10201, the two internal and external detectors are mixed to summarize various situations, and the changes that meet the preset conditions are immediately complemented and verified. Comprehensive processing is given. If it is determined that a car is entering or leaving, the image sensor component 191 is immediately used to detect the change of the distance and magnetic field of the target in the berth. Parking space image data is continuously collected at certain intervals until the vehicle completely stops or leaves. After image collection is completed, the second processor 10202 powers off relevant components, screens and packages the required selected image data according to the distance of the target and the law of magnetic field changes, and submits them to the background for license plate cloud recognition, or immediately implements license plate recognition on the spot (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, after the data is reported, the background will use another set of candidate algorithms to further implement license plate cloud recognition to select the best one), controls the 4G wireless communication module 1211 and its transceiver antenna 1221 to complete the reporting of all business data including the selected image data, and then enters a low-power sleep state.

[0069] Embodiment 4: Image acquisition and processing of dual-camera mixed inspection on the ground in a parking space This embodiment aims at a series of problems such as irregular parking visual blind spots, image defocusing or incomplete license plates derived from the fact that most existing products in the industry use a single camera to collect parking images in parking spaces. Through the reasonable layout of dual cameras and the efficient linkage mechanism between devices, the accuracy of vehicle detection and the success rate of license plate capture and recognition in complex scenes are guaranteed, and power is further saved. The installation and deployment of the equipment has almost no impact on the appearance of the city and does not hinder the walking of people on the road. This solution can be used for the upgrade and transformation of existing license plate cloud recognition cameras or parking space-type roadside equipment. Its main working principle is:

[0070] A-4: Set the average center of the front and rear license plates of the straight parking space: Fig.11 As shown, based on a certain number of representative parking examples on site, the three-dimensional coordinates of the center of the front license plate in case 1 and case 3 where the front of the parking vehicle does not exceed the front line are measured and recorded, and the three-dimensional coordinates of the center of the rear license plate in case 2 where the front of the parking vehicle exceeds the front line are measured and recorded; after sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is obtained by counting the front and rear license plates respectively, and this is used as the average center of the front and rear license plates;

[0071] B-4: A ground-mounted dual-camera mixed inspection image acquisition and processing device is installed on the curb wall of the front berth (i.e., berth 001 in this embodiment) outside the front line of the berth (specifically berth 002 in this embodiment) where the front line and the curb line intersect vertically. Fig.12 , which at least includes: a processor 1020 and an image sensor assembly 19 including a first image sensor component 191, a second image sensor component 192 and a two-choice switching unit connected thereto by wire, a first fill light B1 and its measurement and control unit 171, a second fill light B2 and its measurement and control unit 172, a detector sensor unit 16 and an online data communication interface 121; the processor 1020 is composed of a first processor 10201 connected to the detector sensor unit 16 through a sensor interface to perform comprehensive analysis and processing of whether there is a car or not, and a second processor 10202 connected to the image sensor assembly 19 and the fill light and its measurement and control unit 17 through a video terminal and a measurement and control terminal; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensor unit 16 is composed of a first uniaxial directional geomagnetic M1 1611, a second uniaxial directional geomagnetic M2 1612 and the first radar ranging module R1 1621, the second radar ranging module R2 1622; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor 10202 to collect the image; the online data communication interface 121 is a combination of a 2.4GHz wireless transceiver 1211 and its transceiver antenna 1212, a 4G wireless communication module 1211 and its transceiver antenna 1221; combined Fig.13As shown, the first image sensor component 191 for photographing the front license plate of the present parking space and the second fill light B2 used in pair with the second image sensor component 192 are located at the far end of the image acquisition and processing device, the second image sensor component 192 for photographing the rear license plate of the front parking space and the first fill light B1 used in pair with the first image sensor component 191 are located at the near end of the image acquisition and processing device, and the built-in detector sensor part 16 for detecting the change of the entry and exit status of the present parking space and the front parking space is located at the near end of the image acquisition and processing device; the processor 10201 triggers the linkage wireless signal transmission and reception with the preset associated parking space image acquisition and processing device through the 2.4GHz wireless transceiver 1212 and its transceiver antenna 1222, and the processor 10202 triggers the linkage wireless signal transmission and reception with the preset associated parking space image acquisition and processing device through the 4G wireless communication module 1211 and its transceiver antenna 1221 Send relevant information of the present berth to the back-end system and cloud platform; the center axis of the lens J1 of the first image sensor component 191, the center axis of the light beam of the paired first fill light B1, the sensitivity direction of the first uniaxial directional geomagnetic M1, and the center axis of the beam of the first radar ranging module R1 must all point to the spherical space within the average center radius of 0.6 meters of the license plate in front of the present berth; the center axis of the lens J2 of the second image sensor component 192, the center axis of the light beam of the paired second fill light B2, and the center axis of the beam of the second radar ranging module R2 must all point to the spherical space within the average center radius of 0.6 meters of the license plate behind the front berth, and the sensitivity direction of the second uniaxial directional geomagnetic M2 is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic M1;

[0072] C-4: Under normal circumstances, the second processor 10202 is in a low-power sleep state, and the image sensor component 19 is in a power-off shutdown state; the first processor 10201 is awakened by its long-cycle timer and uses the first and second radar ranging modules R1 and R2 to scan and detect the distance changes of the current berth and the previous berth targets respectively, and is awakened by its short-cycle timer and uses the first and second uniaxial directional geomagnetic M1 and M2 to scan and detect the magnetic field changes and uses the 2.4 GHz wireless transceiver 1212 and its transceiver antenna 12 22 receives the trigger linkage information sent to it by the preset front and rear parking image acquisition and processing equipment in a wireless manner, mixes multiple internal and external detectors to summarize various situations of the associated parking spaces, and immediately complements and confirms the changes that meet the preset conditions. If there is a new state change in the parking space that triggers the linkage information, or if a "request for assistance in capturing the rear license plate" trigger linkage information is received from the front parking space, the second processor 10202 is awakened in a wired manner and relevant information is output to it; if there is a state change in the front parking space that triggers the linkage information, the 2 .4GHz wireless transceiver 1212 and its transceiver antenna 1222 report to the image acquisition and processing equipment of the front berth in a wireless manner; after all output matters are completed, it enters a low-power sleep state; D-4: If the second processor 10202 receives a new state change trigger linkage information of this berth after being awakened, it will power on the first image sensor component 191, the two-choice switching unit 190 and the first fill light B1 and its measurement and control unit 171 that perform ambient light metering to switch the glass slide in advance or light up the fill light when the light is poor, and select the first image sensor component 191 to collect the image data of this berth; if the "request for assistance in capturing the rear license plate" trigger linkage information sent by the front berth is received, the second image sensor component 191 will be powered on. The image sensor component 192, the two-choice switching unit 190, and the second fill light B2 and its measurement and control unit 172 that perform ambient light metering to switch the glass slide in advance or light up the fill light when the light is poor, select the second image sensor component 192 to collect the front parking image data; after the image acquisition is completed, the second processor 10202 powers off the relevant components, packages and submits the parking image data to the background to implement license plate cloud recognition, or immediately implements license plate recognition on the spot (when cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best), controls the 4G wireless communication module 1211 and its transceiver antenna 1221 to complete the reporting of all business data including the parking image data, and then enters a low-power sleep state.

[0073] Embodiment 5: Indoor non-character parking space ceiling-mounted multi-camera external inspection image acquisition and processing. This embodiment aims at the various problems of existing indoor parking lot video detection and recognition products, such as high misjudgment rate, high power consumption, and non-support for ETC wireless communication and recognition, which occur when one camera manages 3-4 parking spaces. An external detector is used to realize on-site real-time detection and trigger linkage between devices. The one-to-one correspondence between parking spaces and cameras realizes accurate matching of license plates and parking spaces. At the same time, if there is a business need, the national standard ETC 5.8GHz transceiver and wake-up device can be turned on and the electronically controlled pan-tilt head can be mobilized to direct the transceiver antenna to the relevant parking space and wirelessly interact with the ETC electronic tag of the parked vehicle. This multi-pronged approach ensures the stability and reliability of parking space management and the development of various new businesses including shared parking, parking space reservation, reverse car search, and new energy vehicle charging. This design is a supplement and improvement to the original technical solution and can be used for upgrading and transformation of existing license plate cloud recognition cameras or parking space-type roadside equipment. Its main working principle is:

[0074] A-5: Set the average center of the front license plate of the indoor non-character parking space: Fig.14 As shown, first install a parking limiter at a suitable place before the parking end boundary line of the parking space (as shown in parking space 001), so that the distance between the front or rear of the parked vehicle and the front boundary line of the parking space is less than 1 meter. Then, after the installation is completed and put into use, based on a certain number of representative parking examples on site, measure and record the three-dimensional coordinates of the center of the license plate of the parked vehicle near the front boundary line of the parking space; after the sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is statistically obtained, which is used as the average center of the front license plate;

[0075] B-5: A ceiling-mounted multi-camera external inspection image acquisition and processing device is installed on the ceiling in the center of the 6 parking spaces, with each lens corresponding to the parking spaces. Fig.15As shown, it at least includes: a 9th processor 10209 and an image sensor assembly 19 including 1st to 6th image sensor components 191-196 and a six-choice switching unit 190 connected thereto by wire, a fill light B and its measurement and control unit 17, and a 9th online data communication interface 129; in addition, it is necessary to further set up external 1st to 6th detectors with the same structure near the front boundary lines of the 6 berths. Now, taking the 1st detector as an example, it at least includes: a 1st processor 10201 and a 1st detector sensor part 161 connected thereto by wire, and a 1st online data communication interface 121; the detector can add a digital code and / or a QR code on its product surface so that it is set at 0.4 meters outside the berth at the center of the front boundary line of the berth parking movement with the appearance of a berth identification plate; the 1st detector sensor part 161 is composed of the 1_1st uniaxial directional geomagnetic M1 16111, the 1_2nd uniaxial directional geomagnetic M2 16112 and the first radar ranging module R1 1612, the second to sixth detectors are similar and will not be expanded here; the processor is a SOC transceiver integrated single chip microcomputer, a microprocessor, a RISC-V core microcontroller, an ARM or DSP processor; the image sensor components 191-196 are the general term for a series of components including the lenses J1-J6 that output the optical images captured by the lenses J1-J6 through their respective image sensors and their supporting circuits as the signals required for the ninth processor to collect images; the online data communication interface 129 includes a TCP / IP network interface 1292, and 2 .4GHz wireless transceiver 1291 and its transceiver antenna 12912; the online data communication interface 121-126 is a combination of 2.4GHz wireless transceivers 1211-1261 and their corresponding transceiver antennas 12112-12612; the ninth processor 10209 sends the relevant information of the berth in this area to the back-end system and the cloud platform through the TCP / IP network interface 1292; the central axis of the lens J1-J6 of the 1st to 6th image sensor components 191-196, the sensitivity direction of the 1_1, 2-1...6-1 single-axis directional geomagnetic M1, M3...M11, and the central axis of the beam of the 1st to 6th radar ranging modules R1-R6 must point to the spherical space within the average center radius of 0.6 meters of the front license plate of the 1st to 6th parking spaces respectively; the 1_2, 2-2...6-2 single-axis directional geomagnetic M2, The sensitivity directions of M4...M12 are respectively perpendicular or parallel to the sensitivity directions of the 1-1, 2-1...6-1 uniaxial directional geomagnetic M1, M3...M11;

[0076] C-5: Under normal circumstances, the image sensor assembly 19 is in a power-off shutdown state; the 9th processor 10209 listens and receives the berth state change trigger linkage information sent by the external nth detector (n is an integer from 1 to 6, the same below) through the 2.4GHz wireless transceiver 1291 and its transceiver antenna 12912 in a short period; the nth processor 1020n uses the nth radar ranging module to scan and detect the distance changes of the target in the berth after being awakened by its long period timer, and uses the n_1th and n_2th single-axis directional geomagnetic scans to detect magnetic field changes after being awakened by its short period timer, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing is carried out, and if there is a new berth state change triggering linkage information, the relevant information is reported to the 9th processor 10209 through the nth online data communication interface, and after the reporting is completed, it enters a low-power sleep state; D-5: If the 9th processor receives the linkage information triggered by the new state change of the nth berth, it powers on the nth image sensor component 19n, the six-choice switching unit 190, and the fill light B and its measurement and control unit 17 for performing ambient light measurement to switch the slide in advance or light up the fill light when the light is not good, and selects the nth image sensor component 19n to collect the berth image data; after the image acquisition is completed, the ETC or RFID functional components (such as Figure 14-15 As shown, the image acquisition and processing equipment in this embodiment further includes a microwave radio frequency communicator 11, which has a national standard ETC 5.8GHz transceiver and wake-up device 111 and a transceiver antenna 112 built in. At the same time, there is also an ETC vehicle-mounted electronic tag on the vehicle) and wireless communication business processing is required, the national standard ETC 5.8GHz transceiver and wake-up device 111 can be turned on and the electric control pan-tilt head can be mobilized to direct the transceiver antenna 112 to the relevant berth and interact with the ETC electronic tag of the parked vehicle to complete the wireless communication interaction. Then the ninth processor 10209 powers off the relevant components, packages and submits the image data to the background to implement license plate cloud recognition, or immediately implements license plate on-site recognition (cloud recognition or on-site recognition has no results, or the results come out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best), and controls the TCP / IP network interface 1292 to complete the reporting of all business data including the image data.

[0077] It is worth mentioning that for extended business needs such as reserved parking, which requires the installation of parking locks and the completion of video license plate recognition and / or ETC vehicle-mounted electronic tag wireless recognition outside the parking space in advance to further obtain unlocking authorization, reserved parking detection points can be set up, such as Fig.14As shown, on the basis of the present embodiment, it is only necessary to add external 7th and 8th detectors to the reserved parking detection point, and correspondingly install 7th and 8th image sensor components in the ceiling-mounted multi-camera external inspection image acquisition and processing equipment, and expand the six-to-one switching unit to an eight-to-one switching unit, thereby realizing equipment reuse and new business development, which will not be repeated here.

[0078] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the claims of the present invention.

Claims

1. A method for collecting and processing images of a non-character parking space pole-type dual-camera internal inspection, suitable for a license plate cloud recognition camera or a parking space type roadside device, characterized in that: The steps include: A-1: A parking limiter is set on the ground at 1.1-1.6 meters away from the parking end boundary line of the berth to standardize the parking posture of the vehicle, and a pole-type dual-camera internal inspection image acquisition and processing device is set at a suitable place around the center of the end boundary line, which at least includes: a processor and an image sensor component including the first and second image sensor components and a two-choice switching unit connected to it by wire, and the first and second fill lights and their measurement and control units, detector sensing parts and online data communication interfaces; the processor is composed of a first processor connected to the detector sensing part for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor component for intercommunication; the first and second processors are SOC transceiver integrated microcontrollers, microprocessors, RISC-V core microcontrollers, ARM or DSP processor; the detector sensing part is composed of the first and second uniaxial directional geomagnetic and radar ranging modules; the image sensor component is a general term for a series of components including the lens that output the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor to collect the image; the online data communication interface is one or more of the 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna The combination of the following items: the first image sensor component on the pole for photographing low-position license plates has a lens height of 0.48±0.18 meters from the berth ground, the second image sensor component for photographing high-position license plates has a lens height of 0.78±0.18 meters from the berth ground, the first fill light with a depression angle of 20°±7° used in conjunction with the first image sensor component is 0.78±0.3 meters from the berth ground, and the second fill light used in conjunction with the second image sensor component is 0.68±0.2 meters from the berth ground for horizontal forward fill light; the built-in detector sensor part for detecting changes in the state of vehicles entering and exiting the berth is 0.68±0.2 meters from the berth ground m; the processor sends the relevant information of the berth to the back-end system and the cloud platform through the online data communication interface; the lens center axis of the first and second image sensor components, the sensitivity direction of the first uniaxial directional geomagnetic, and the beam center axis of the radar ranging module are basically parallel to the center axis of the left and right boundaries of the berth with a deviation of ≤10°, and the sensitivity direction of the second uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic; B-1: The second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the radar ranging module to scan and detect changes in the distance of targets in the berth, and after being awakened by its short-cycle timer, uses the first and second single-axis directional geomagnetic scans to detect changes in the magnetic field, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing is performed, and if there is a new berth status change that triggers linkage information, the second processor is awakened in a wired manner and relevant information is output to it, and after the output is completed, it enters a low-power sleep state; C-1: If the second processor receives a new state change trigger linkage information of the berth after being awakened, the first and second image sensor components, the two-choice switching unit, and the first and second fill lights and their measurement and control units that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor are powered on in sequence. First, the first image sensor component is selected to collect the low-level image data of the berth, and then the second image sensor component is selected to collect the high-level image data of the berth; after the image acquisition is completed, the components with ETC or RFID functions and the need to perform wireless communication business processing can be further executed according to the existing business process, and then the second processor powers off the relevant components, packages and submits the image data to the background to implement license plate cloud recognition, or implements license plate recognition on the spot immediately. If there is no result in cloud recognition or on-site recognition, or the result comes out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best, and manipulates the online data communication interface to complete the reporting of all business data including the image data, and then enters a low-power sleep state.

2. A method for collecting and processing images of a vertical pole type dual-camera mixed inspection of a straight parking space, suitable for a license plate cloud recognition camera or a parking space type roadside device, characterized in that: The steps include: A-2: Set up straight parking spaces for standardized parking and the average center of the rear license plate of irregular parking: Based on a certain number of representative parking examples on site, measure and record the three-dimensional coordinates of the center of the rear license plate when the parked vehicle does not exceed the rear line, and measure and record the three-dimensional coordinates of the center of the rear license plate when the parked vehicle exceeds the rear line; After the sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is calculated according to the license plate after standard parking and the license plate after irregular parking, and this is used as the average center of the license plate after standard parking and irregular parking; B-2: A pole-type dual-camera mixed inspection image acquisition and processing device is set at a suitable place on the curb platform at a distance of 1±0.25 meters from the front line of the berth and 0.35±0.20 meters from the curb edge, which at least includes: a processor and an image sensor assembly including the first and second image sensor components and a two-choice switching unit connected to it by wire, a fill light and its measurement and control unit, a detector sensor part and an online data communication interface; the processor is composed of a first processor connected to the detector sensor part for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor assembly for intercommunication; the first and second processors are SOC transceiver integrated microcontrollers, microprocessors, RISC-V core microcontrollers, ARM or DSP processor; the detector sensing part is composed of the first and second uniaxial directional geomagnetic and the first and second radar ranging modules; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor to collect the image; the online data communication interface is a combination of one or more of a 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna; the first image sensor component used to shoot the license plate of the front parking space after the standard parking has a lens height of 0.77±0.2 meters from the parking space ground, and the second image sensor component used to shoot the license plate of the front parking space after the irregular parking has a lens height of 0.87±0.2 meters from the parking space ground m, the fill light used in conjunction with the first and second image sensor components is 0.87±0.2 m above the ground of the berth, and the first and second uniaxial directional geomagnetic and first and second radar ranging modules of the built-in detector sensor used to detect the change of the entry and exit status of the vehicle in the current berth and the front berth are 0.85±0.2 m above the ground of the berth; the processor receives the trigger linkage wireless signal sent to it by the preset associated berth image acquisition and processing device through the online data communication interface, and sends the relevant information of the current berth to the preset associated berth image acquisition and processing device and / or the back-end system and cloud platform; the central axis of the lens of the first image sensor component and the central axis of the beam of the matching fill light, and the central axis of the beam of the first radar ranging module must all point to the spherical space within 0.6 m of the average center radius of the license plate after the standard parking of the front berth; the central axis of the lens of the second image sensor component must point to the spherical space within 0.6 meters of spherical space; the sensitivity direction of the first uniaxial directional geomagnetic and the beam center axis of the second radar ranging module are basically horizontal and perpendicular to the left and right sides of the vehicle body, with an allowable deviation of ≤15°, and the sensitivity direction of the second uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic;. C-2: The second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the first and second radar ranging modules to scan and detect the distance changes of the targets in the front berth and the current berth respectively; after being awakened by its short-cycle timer, the first processor uses the first and second single-axis directional geomagnetic scanning to detect the magnetic field changes and uses the online data communication interface to receive the trigger linkage information sent to it by the preset front and rear berth image acquisition and processing equipment in a wireless manner, and the mixed internal and external multiple detectors summarize the various conditions of the associated berths, and give instant complementary verification and comprehensive processing to the changes that meet the preset conditions. If there is a new state change of the current berth that triggers the linkage information, the ETC or RFID functional components that need to perform wireless communication business processing can be further executed according to the existing business process, and the relevant business information is reported to the image acquisition and processing equipment of the rear berth in a wireless manner through the online data communication interface; if there is a new state change of the front berth that triggers the linkage information, the second processor is awakened in a wired manner and the relevant information is output to it; after all output matters are completed, it enters a low-power sleep state; D-2: If the second processor receives a new state change trigger linkage information of the front berth after being awakened, the first and second image sensor components, the two-choice switching unit, and the fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor are powered on, and the first image sensor component is selected to collect the front berth image data 1; then the second image sensor component is selected to collect the front berth image data 2; after the image acquisition is completed, the second processor powers off the relevant components, packages and submits the image data 1 and 2 to the background to implement license plate cloud recognition, or implements license plate recognition on the spot immediately. If there is no result in cloud recognition or on-site recognition, or the result comes out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best, and manipulates the online data communication interface to complete the reporting of all business data including the image data 1 and / or image data 2, and then enters a low-power sleep state.

3. A method for collecting and processing images of non-character parking spaces with a single-camera mixed inspection, suitable for license plate cloud recognition cameras, characterized in that: The steps include: A-3: A ground-mounted single-camera mixed-inspection image acquisition and processing device is set at a suitable place around the center of the boundary line at the end of the parking space, which at least includes: a processor and an image sensor component and a fill light and its measurement and control unit connected to it by wire, the first and second detector sensing parts and an online data communication interface; the processor is composed of a first processor connected to the first detector sensing part for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the second detector sensing part and the image sensor component; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the first detector sensing part is composed of the first and second single-axis directional geomagnetic or integrated AMR three-axis geomagnetic, and the first radar ranging module; the second detector sensing part is composed of the third and fourth The image sensor component is composed of a single-axis directional geomagnetic and a second radar ranging module; the image sensor component is a general term for a series of components including a lens that outputs the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor to collect the image; the online data communication interface is a combination of one or more of a 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna; the image sensor component for photographing the license plate has a lens located at the center of the terminal boundary line, and its left or right side is a supporting fill light with an elevation angle of 20°±7°; the external first detector sensor part and the first processor for detecting changes in the entry and exit status of vehicles in the parking space can add digital codes and / or QR codes on the surface of the product so that it can be installed at the center of the boundary line of the first end of the parking space with the appearance of a parking space identification plate, 0.4 outside the parking space. ±0.4 meters, and the built-in second detector sensor and the second processor are located in the image acquisition and processing equipment at the center of the terminal boundary line; the second processor sends the relevant information of the berth to the back-end system and the cloud platform through the online data communication interface; the lens center axis of the image sensor component, the sensitivity direction of the third uniaxial directional geomagnetic and the beam center axis of the second radar ranging module must all point to the direction of the vehicle entering and exiting and form an elevation angle of 20°±7° with the ground; the sensitivity direction of the first uniaxial directional geomagnetic and the beam center axis of the first radar ranging module are all facing the sky and are basically perpendicular to the ground with a deviation of ≤10°; the sensitivity directions of the second and fourth uniaxial directional geomagnetic are respectively perpendicular or parallel to the sensitivity directions of the first and third uniaxial directional geomagnetic; B-3: The second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the first radar ranging module to scan and detect changes in the distance of targets outside the berth, and after being awakened by its short-cycle timer, uses the first and second single-axis directional geomagnetic scans to detect changes in the magnetic field, and immediately complements and confirms the changes that meet the preset conditions. If there is a new change in the state outside the berth that triggers linkage information, the second processor is awakened in a wired manner and relevant information is output to it. After the output is completed, it enters a low-power sleep state; C-3: If the second processor receives a new state change trigger linkage information outside the berth after being awakened, it will power on and pre-start the image sensor component and the fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light in poor light conditions. At the same time, the second detector sensor part is used in a short period to detect the change of the distance and magnetic field of the target in the berth. Combined with the real-time change information provided by the first processor, the two internal and external detectors are mixed to summarize various situations, and the changes that meet the preset conditions are immediately complemented and verified. Comprehensive processing is given. If it is determined that a car is entering or leaving, the image sensor component is immediately used to detect the change of the distance and magnetic field of the target in the berth. Parking space image data is collected continuously at intervals until the vehicle completely stops or leaves; after image collection is completed, the second processor powers off relevant components, screens and packages the required selected image data according to the distance of the target and the law of magnetic field changes, and submits them to the background to implement license plate cloud recognition, or to implement license plate recognition on the spot immediately. When cloud recognition or on-site recognition has no results, or when the results come out but the character confidence is insufficient, after the data is reported, the background will use another set of candidate algorithms to further implement license plate cloud recognition to select the best, and control the online data communication interface to complete the reporting of all business data including the selected image data, and then enter a low-power sleep state.

4. A method for collecting and processing images of a ground-mounted dual-camera mixed inspection of a parking space, suitable for a license plate cloud recognition camera or a parking space type roadside device, characterized in that: The steps include: A-4: Set the average center of the front and rear license plates in a straight parking space: Based on a certain number of representative parking examples on site, measure and record the 3D coordinates of the center of the front license plate when the front of the parking vehicle does not exceed the front line, and measure and record the 3D coordinates of the center of the rear license plate when the front of the parking vehicle exceeds the front line; after sample collection is completed, the average value of the relevant samples in the 3D coordinate system of the parking space is calculated according to the front and rear license plates, and this is used as the average center of the front and rear license plates; B-4: A ground-mounted dual-camera mixed inspection image acquisition and processing device is installed on the front parking space curb wall outside the front line where the front line of the parking space and the curb line intersect vertically, and the device at least comprises: a processor and an image sensor assembly including the first and second image sensor components and a two-choice switching unit connected to the processor by wire, and the first and second fill lights and their measurement and control units, a detector sensing unit and an online data communication interface; the processor is composed of a first processor connected to the detector sensing unit for comprehensive analysis and processing of the presence or absence of a vehicle, and a second processor connected to the image sensor assembly for comprehensive analysis and processing of the presence or absence of a vehicle. The processors are interconnected and matched; the first and second processors are SOC transceiver integrated single-chip microcomputers, microprocessors, RISC-V core microcontrollers, ARM or DSP processors; the detector sensing part is composed of the first and second single-axis directional geomagnetic and the first and second radar ranging modules; the image sensor component is a general term for a series of components including the lens that output the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the second processor to collect the image; the online data communication interface is a 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and its transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / a combination of one or more of 6G, WiFi, Bluetooth, ZigBee wireless communication modules and their transceiver antennas; a first image sensor component for photographing the front license plate of the present parking space and a second fill light used in pair with the second image sensor component are located at the far end of the image acquisition and processing device; a second image sensor component for photographing the rear license plate of the front parking space and a first fill light used in pair with the first image sensor component are located at the near end of the image acquisition and processing device; a built-in detector sensor for detecting changes in the entry and exit status of vehicles in the present parking space and the front parking space is located at the near end of the image acquisition and processing device; the processor receives a trigger linkage wireless signal sent to it by a preset associated parking space image acquisition and processing device through the online data communication interface. No., send the relevant information of this berth to the preset associated berth image acquisition and processing equipment and / or back-end system, cloud platform; the central axis of the lens of the first image sensor component and the central axis of the light beam of the paired first fill light, the sensitivity direction of the first uniaxial directional geomagnetic, and the central axis of the beam of the first radar ranging module must all point to the spherical space within the average center radius of 0.6 meters of the front license plate of the berth; the central axis of the lens of the second image sensor component and the central axis of the light beam of the paired second fill light, and the central axis of the beam of the second radar ranging module must all point to the spherical space within the average center radius of 0.6 meters of the rear license plate of the front berth, and at the same time, the sensitivity direction of the second uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the first uniaxial directional geomagnetic; C-4: The second processor is in a low-power sleep state, and the image sensor component is in a power-off shutdown state; after being awakened by its long-cycle timer, the first processor uses the first and second radar ranging modules to respectively scan and detect changes in the distance of the targets in the current berth and the front berth; after being awakened by its short-cycle timer, the first processor uses the first and second single-axis directional geomagnetic scanning to detect changes in the magnetic field and uses the online data communication interface to receive the trigger linkage information sent to it by the preset front and rear berth image acquisition and processing equipment in a wireless manner, and uses multiple internal and external detectors to summarize various conditions of the associated berths, and immediately complements and confirms the changes that meet the preset conditions. Comprehensive processing, if there is a new state change in the current berth to trigger linkage information, or if a "request for assistance in capturing the rear license plate" trigger linkage information is received from the front berth, the second processor is awakened in a wired manner and relevant information is output to it; if there is a state change in the front berth to trigger linkage information, it is reported to the image acquisition and processing equipment of the front berth in a wireless manner through the online data communication interface; after all output matters are completed, it enters a low-power sleep state; D-4: If the second processor receives a new state change trigger linkage information of the current berth after being awakened, the first image sensor component, the two-choice switching unit, and the first fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor, and select the first image sensor component to collect image data of the current berth; if the "request for assistance in capturing the rear license plate" trigger linkage information sent by the front berth is received, the second image sensor component, the two-choice switching unit, and the second fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor, and select the second image sensor component to collect the front berth image data. Parking space image data; after image acquisition is completed, those with ETC or RFID functional components and requiring wireless communication business processing can be further executed according to the existing business process, and then the second processor powers off the relevant components, packages and submits the parking space image data to the background to implement license plate cloud recognition, or implements license plate recognition on the spot immediately. When cloud recognition or on-site recognition has no results, or the results come out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best, and manipulates the online data communication interface to complete the reporting of all business data including the parking space image data, and then enters a low-power sleep state.

5. A method for collecting and processing images of indoor non-letter parking spaces with ceiling-mounted multi-cameras, applicable to license plate cloud recognition cameras or parking space-type roadside equipment, characterized in that: The steps include: A-5: Setting the average center of the front license plate of the indoor non-character parking space: First, install a parking limiter at a suitable place before the parking end boundary line of the parking space, so that the distance between the front or rear of the parked vehicle and the front boundary line of the parking space is less than 1 meter. Then, after the installation is completed and put into use, based on a certain number of representative parking examples on site, measure and record the three-dimensional coordinates of the center of the license plate of the parked vehicle near the front boundary line of the parking space; After the sample collection is completed, the average value of the relevant samples in the three-dimensional coordinate system of the parking space is obtained and used as the average center of the front license plate; B-5: A ceiling-mounted multi-camera external inspection image acquisition and processing device with lenses corresponding to the parking spaces one by one is set on the ceiling in the center of the n parking space area, 2≤n≤8, which at least includes: a 9th processor and an image sensor assembly including the 1st to nth image sensor components and an n-to-one switching unit connected to it by wire, a fill light and its measurement and control unit, and a 9th online data communication interface; in addition, n external detectors need to be further set near the front boundary line of the n parking spaces, which at least include: an nth processor and an nth detector sensor part connected to it by wire, and an nth online data communication interface. The detector can add digital codes and / or QR codes on its product surface so that it can be set at 0.4±0.4 meters outside the parking space at the center of the front boundary line of the parking space with the appearance of a parking space identification plate; the nth detector sensor part is composed of the n_1st, n_2nd uniaxial directional geomagnetic and n_3rd uniaxial directional geomagnetic sensors. The radar ranging module is composed of: the processor is a SOC transceiver integrated single-chip microcomputer, a microprocessor, a RISC-V core microcontroller, an ARM or a DSP processor; the image sensor component is a general term for a series of components including a lens, which outputs the optical image captured by the lens through the image sensor and its supporting circuit as the signal required for the 9th processor to collect the image; the online data communication interface includes a wired and / or wireless data communication interface, the wired data communication interface includes a UART, RS232 / RS485 / RS422, USB or TCP / IP network interface, and the wireless data communication interface is a 433, 470-510MHz, 2.4GHz frequency band micro-power wireless transceiver and a combination of one or more of the transceiver antenna, NB-IOT, Lora / LoraWAN / CLAA, 4G / 5G / 6G, WiFi, Bluetooth, ZigBee wireless communication module and its transceiver antenna; the 9th processor sends the relevant information of the parking spaces in this area to the back-end system and the cloud platform through the 9th online data communication interface; the lens center axis of the nth image sensor component, the sensitivity direction of the n_1th uniaxial directional geomagnetic, and the beam center axis of the nth radar ranging module must all point to the spherical space within 0.6 meters of the average center radius of the license plate at the front end of the nth parking space; the sensitivity direction of the n_2th uniaxial directional geomagnetic is perpendicular or parallel to the sensitivity direction of the n_1th uniaxial directional geomagnetic; C-5: The image sensor assembly is in a power-off shutdown state; the 9th processor listens to and receives the berth status change trigger linkage information sent by the n external detectors through the 9th online data communication interface in a short period; the nth processor uses the nth radar ranging module to scan and detect the distance changes of the target in the berth after being awakened by its long period timer, and uses the n_1th and n_2th single-axis directional geomagnetic scans to detect magnetic field changes after being awakened by its short period timer, and performs instant complementary verification and comprehensive processing on the changes that meet the preset conditions. If there is a new berth status change triggering linkage information, the relevant information is reported to the 9th processor through the nth online data communication interface, and after the reporting is completed, it enters a low-power sleep state; D-5: If the 9th processor receives a new state change trigger linkage information for the nth berth, it powers on the nth image sensor component, the n-to-one switching unit, and the fill light and its measurement and control unit that perform ambient light measurement to switch the glass slide in advance or light up the fill light when the light is poor, and selects the nth image sensor component to collect berth image data; after the image acquisition is completed, the components with ETC or RFID functions and need to perform wireless communication business processing can be further executed according to the existing business process, and then the 9th processor powers off the relevant components, packages and submits the image data to the background to implement license plate cloud recognition, or implements license plate recognition on the spot immediately. If there is no result from cloud recognition or on-site recognition, or the result comes out but the character confidence is insufficient, after the data is reported, the background uses another set of candidate algorithms to further implement license plate cloud recognition to select the best and control the 9th online data communication interface to complete the reporting of all business data including the image data.

6. The image acquisition and processing method according to any one of claims 1 to 5, characterized in that: The image sensor component includes: a polarizing film that also serves as a lens viewfinder and can suppress license plate reflection, a fixed aperture fixed focal length lens with an aperture of F0.95-F2.8 and a focal length of 1.9mm-16mm that matches the use environment, a lens mount that includes a neutral gray density dimming film and a high-transmittance white glass sheet and can be controlled to switch between the two through a glass sheet switching end, a 400nm-650 or 700nm visible light passing infrared light cutoff filter attached to the lens mount, a GalaxyCore GC0403 1 / 3 inch 768 x 576 resolution CMOS image sensor, and an automatic exposure and automatic white balance camera controller chip connected to the image sensor through an internal connection end, and the external connection end of the chip can be used as a signal source end of the image sensor component to connect to the switching unit or the processor; the camera controller chip can be preset as needed to crop the output JPG picture with a resolution of 640×480 in a centered, left-biased or right-biased manner based on the original frame to further save data traffic and bandwidth.

7. The image acquisition and processing method according to claim 6, characterized in that: The camera controller chip is SN9C2788 from SONIX Technology; the switching unit is the low-resistance analog switch chip CH440 / 442 / 443 / 444 / 445 / 448 from Nanjing Qinheng Microelectronics.

Images