An in-loop test system for the performance of intelligent vehicle cameras in rainy weather

Abstract

The present invention provides an in-loop test system for the performance of an intelligent vehicle camera in rainy weather, including a real rainfall simulation box, a display, an industrial control computer, and a camera under test. The in-loop test process of the present invention is based on the real rainfall simulation box, combined with virtual traffic simulation. The virtual traffic simulation provides complex road conditions and dynamic and static traffic participant information for the camera under test. The real rainfall simulation box is responsible for simulating the blur effect caused by raindrops in the camera's field of view. By fusing physical rainfall simulation and virtual simulation image information, it provides image information in a highly realistic rainfall environment for the camera under test for closed-loop testing, and can test the camera hardware, camera image processing algorithm, camera ECU, and ADAS algorithm.

Description

Technical Field

[0001] The present invention relates to an in-loop test system for camera performance, and particularly to an in-loop test system for the camera performance of intelligent vehicles in rainy weather. Background Art

[0002] The rainy environment poses great challenges to the camera sensors of intelligent vehicles. It is mainly manifested that in rainy weather, the intelligent vehicle cameras cannot complete accurate target recognition, target tracking and other perception functions. The functional boundary refers to the step change of the intelligent vehicle camera function caused by slightly changing the specific index value while other scene indexes remain unchanged. However, the specific functional boundary needs to be fully tested to be determined. There are three sources of image noise generated by rain, namely, hierarchical superposition blur caused by different rainfall amounts, light scattering effect blur caused by raindrops of different sizes, and image blur caused by the synthesis of the movement of the vehicle and the rain. Most of the existing test methods for the functions of intelligent vehicles in rainy scenarios adopt closed-field tests, that is, within a specific closed-field condition, a rainy scenario for vehicle-level tests is constructed by using real physical rainfall simulation equipment and dynamic and static traffic participants such as dummies and dummy vehicles. This test method has the disadvantages of extremely high test costs, low test efficiency, and poor controllability of test scene elements. There is an urgent need for a micro-reality rainfall simulation platform with high fidelity, strong controllability and low cost in the laboratory to test the camera performance of intelligent vehicles in rainy weather. Summary of the Invention

[0003] To solve the above technical problems, the present invention provides an in-loop test system for the camera performance of intelligent vehicles in rainy weather, including a real rainfall simulation box, a display, an industrial control computer and a camera under test. The display and the camera under test are respectively arranged on the left and right sides of the real rainfall simulation box. The real rainfall simulation box is a transparent box body, and at least one side of the box body facing the camera under test is a vehicle-grade glass. A plurality of main water pipes are arranged side by side above the real rainfall simulation box. Both ends of each main water pipe are respectively pivotally connected to the box body of the real rainfall simulation box through bearings. A stepping motor is further arranged in the real rainfall simulation box. The output shaft of the stepping motor is connected with a main sprocket. A driven sprocket is respectively arranged at one end of each main water pipe. The main sprocket and the driven sprocket are connected by a chain drive.

[0004] The lower part of each main water pipe is evenly provided with 6 kinds of nozzles with different flow specifications; they include fine rain nozzles, light rain nozzles, moderate rain nozzles, heavy rain nozzles, rainstorm nozzles and heavy rainstorm nozzles respectively. The coverage area of a single nozzle ensures the complete and uniform coverage of the image rainfall blurring effect within the camera's perspective. The 6 kinds of nozzles with different flow specifications are used to simulate 6 kinds of rainfall with different intensities. The rainfall intensities are divided into fine rain, light rain, moderate rain, heavy rain, rainstorm and heavy rainstorm. When all the nozzles of the same specification are opened, the corresponding rainfall amounts need to meet the following requirements: in the case of no leakage and external natural factors within 1 hour, fine rain ≤ 5mm, light rain 5 - 10mm, moderate rain 10 - 25mm, heavy rain 26 - 40mm, rainstorm 50 - 80mm, heavy rainstorm 90 - 120mm.

[0005] A water tank is provided below the real rainfall simulation box. A water pump is provided inside the water tank. The water pump is connected to the water delivery pipe through a filter. The water delivery pipe is connected to a water distributor. The water distributor is provided with several water distribution ports. Each water distribution port is connected to the corresponding main water pipe inside the real rainfall simulation box through a water supply pipe;

[0006] The industrial control computer is respectively connected to the display, the camera under test, the stepping motor and the water pump.

[0007] The end of the water supply pipe is connected to the main water pipe through a hose.

[0008] The bottom of the real rainfall simulation box is provided with a drain port and a corresponding rainwater recovery channel.

[0009] The closed-loop test process of the present invention is based on the real rainfall simulation box and combines virtual traffic simulation. The virtual traffic simulation provides complex road conditions and dynamic and static traffic participant information for the camera under test. The real rainfall simulation box is responsible for simulating the blurring effect caused by raindrops in the camera's perspective. By fusing physical rainfall simulation and virtual simulation image information, it provides image information in a highly realistic rainfall environment for the camera under test for closed-loop testing. The specific process is as follows:

[0010] The virtual simulation software outputs the image information of the camera's perspective and the vehicle's motion speed at the current moment. Among them, the vehicle motion speed information will be synthesized with the expected simulated rainfall speed in the real rainfall simulation box to calculate the spraying angle of the nozzles on the main water pipe. The nozzles corresponding to the rainfall intensity on the main water pipe are opened, the water pump is started, and the rotation of the nozzles on the main water pipe is controlled by the frequency of the input electrical pulse signal of the stepping motor to simulate the rainfall state corresponding to the vehicle speed, forming the blurred noise of the real raindrop image. In addition, the image perspective information of the measured camera provides the image information of the virtual traffic participants constructed by the virtual simulation software through the high-definition display. The refresh rate of the display should be higher than that of the measured camera to ensure the real-time nature of the image information received by the camera. The measured camera receives the above two kinds of fused image information, and provides the target physical parameters required by the ADAS system through the measured target detection algorithm and the motion state estimation algorithm. The ADAS algorithm combines the input target physical parameters for planning and outputs the lateral and longitudinal control information to the virtual simulation software to update the image information of the camera's perspective in the real-time virtual simulation scenario, completing the closed-loop test.

[0011] The method for controlling the spraying angle of the nozzles is as follows:

[0012] During the falling process of the raindrops, they are under the combined action of gravity, pressure difference resistance, and air buoyancy to reach the force balance. The pressure difference resistance f received by the raindrops during the falling process is:

[0013]

[0014] Among them, ρ is the fluid density, S is the horizontal projection area, C d is the air resistance coefficient. The motion equation under the combined action of the above three forces is:

[0015]

[0016] Among them, ρ 水滴 is the rainwater density, r is the raindrop radius. Let dv / dt = 0, and the raindrop speed v rain can finally be obtained as:

[0017]

[0018] Through the above formula, the vertical falling speed of raindrops corresponding to different raindrop radii is obtained, and through the motion synthesis with the horizontal vehicle motion speed v veicle output by the simulation software at this moment, the raindrop spraying angle Φ and the speed v sim of the sprayed raindrops are determined. The spraying angle of the nozzles on the main water pipe is adjusted by the stepping motor respectively, and the water pump water volume is controlled to construct the real rainfall simulation box.

[0019]

[0020]

[0021] Among them, v veicle : represents the longitudinal speed of the vehicle in the test scenario; v rain : represents the raindrop speed in the test scenario;

[0022] In the present invention, a stepper motor is used as the specific implementation device for controlling the raindrop direction, and the step angle of the stepper motor is:

[0023]

[0024] In the formula, K is the energization coefficient. When the number of phases is equal to the number of beats, K = 1; otherwise, K = 2; m is the number of stator phases; z is the number of rotor teeth. If the frequency of the input electrical pulse signal of the stepper motor is F, the rotational speed of the stepper motor is:

[0025]

[0026] The rotational speed of the stepper motor represents the response speed of the platform to rainfall. The higher the frequency of the electrical pulse signal, the more the actual platform can respond to the vehicle speed change and then simulate the rainfall meteorology in real time.

[0027] Before the experiment, first determine the rainfall intensity to be simulated, and select the corresponding specification nozzles on the main water pipe to open in advance. At the same time, the type selection of the water pump should meet the working pressure standard range of the nozzles. To ensure the normal use of the water pump, the vacuum degree at the water pump suction port cannot be too large, otherwise when the absolute pressure is lower than the air separation pressure of water, the air dissolved in the water will be separated to form bubbles, resulting in cavitation phenomenon, causing vibration and noise. Therefore, it is necessary to limit the vacuum degree at the water pump suction port to be less than 0.03 MPa, which is limited by restricting the pumping height of the water pump. Usually, the pumping height should not be greater than 0.5 m. Pipe filters should be installed in the water inlet and return channels of the water tank for use by the rainfall equipment.

[0028] The beneficial effects of the present invention:

[0029] The present invention synthesizes the vehicle speed and the raindrop falling speed output by the simulation software to achieve the image blurring effect, and then determines the direction of the raindrop nozzle to ensure the fidelity of the micro physical rainfall simulation method. At the same time, the blurring of the actual camera view is not only caused by the synthesis of raindrop size and movement, but also by the image blurring caused by different rainfall amounts. Therefore, the present invention designs different models of nozzles to be arranged on the same water pipe to ensure the construction of scenarios that meet the test requirements of different rainfall amounts. The innovation of the present invention lies in constructing a test platform by coupling virtual and real methods, embedding the camera under test into the test platform in the form of a hardware entity, and fully considering the distribution characteristics of rainy days and the movement characteristics of raindrops in the platform and synthesizing them with vehicle movement to simulate a real rainfall environment. In the present invention, the road site information and traffic information are virtual scenarios constructed by the simulation software. Therefore, various road features such as crossroads, T-shaped intersections, and on- and off-ramps can be constructed, and different dynamic and static traffic participants such as bicycles, pedestrians, and motorcycles can also be simulated, which has better controllability compared to the test elements of real vehicle tests in closed sites. The micro physical rainfall simulation method is used in the construction of the platform, and its construction cost is significantly lower than that of the wide-area physical rainfall simulation method in closed sites. The present invention can test camera hardware, camera image processing algorithms, camera ECU, and ADAS algorithms. Description of the Drawings

[0030] Figure 1 Schematic diagram of the overall structure of the in-loop test system of the present invention;

[0031] Figure 2 Schematic diagram of the structure of the real rainfall simulation box of the present invention;

[0032] Figure 3 Schematic diagram of the process of the in-loop test system of the present invention;

[0033] Figure 4 Schematic diagram of the rainfall simulation angle at different vehicle speeds and different rainfall intensities of the present invention;

[0034] Figure 5 Schematic diagram of the image information of the view angle of the camera under test of the present invention;

[0035] Figure 6 Schematic diagram of the in-loop test process of the present invention.

[0036] 1. Real rainfall simulation box 2. Display 3. Industrial computer 4. Camera under test 5. Automotive-grade glass 6. Main water pipe 7. Stepper motor 9. Main sprocket 10. Driven sprocket 11. Chain 12. Nozzle 13. Water tank 14. Water pump 15. Filter 16. Water delivery pipe 17. Water distributor 18. Water supply pipe 19. Hose. Detailed Implementation Manner

[0037] like Figure 1-2 As shown, the present invention provides an in-loop test system for the performance of an intelligent automobile camera in rainy weather, comprising a real rainfall simulation box 1, a display 2, an industrial computer 3 and a camera to be tested 4, wherein the display 2 and the camera to be tested 4 are respectively arranged on the left and right sides of the real rainfall simulation box 1, and the real rainfall simulation box 1 is a transparent box, and at least one side of the box facing the camera to be tested is a car-grade glass 5; a plurality of main water pipes 6 are arranged side by side on the upper part of the real rainfall simulation box 1, and both ends of each main water pipe 6 are respectively pivotally connected to the inner wall of the box of the real rainfall simulation box 1 through ball bearings; a stepper motor 7 is also arranged in the real rainfall simulation box 1, and the output shaft of the stepper motor 7 is connected to a main sprocket 9, and one end of each main water pipe 6 is respectively provided with a driven sprocket 10, and the main sprocket 9 and the driven sprocket 10 are connected through a chain 11 for transmission;

[0038] The lower part of each main water pipe 6 is provided with 6 nozzles 12 with different flow specifications at equal intervals, including drizzle nozzles, light rain nozzles, moderate rain nozzles, heavy rain nozzles, rainstorm nozzles and heavy rainstorm nozzles respectively. The coverage area of a single nozzle 12 is circular with a diameter of about 4 to 5m to ensure complete and uniform coverage of the image rainfall blur effect within the camera viewing angle. The 6 nozzles 12 with different flow specifications are used to simulate 6 different intensities of rainfall. The rainfall intensity is divided into drizzle, light rain, moderate rain, heavy rain, rainstorm and heavy rainstorm. The corresponding rainfall when all the nozzles 12 of the same specification are turned on must meet the following requirements: within 1 hour, in the absence of leakage and external natural factors, drizzle ≤5mm, light rain 5 to 10mm, moderate rain 10 to 25mm, heavy rain 26 to 40mm, rainstorm 50 to 80mm, and heavy rainstorm 90 to 120mm.

[0039] A water tank 13 is provided below the real rainfall simulation box 1. A water pump 14 is provided in the water tank 13. The water pump 14 is connected to a water pipe 16 through a filter 15. The water pipe 16 is connected to a water distributor 17. The water distributor 17 is provided with a plurality of water outlets. Each water outlet is connected to a corresponding main water pipe 6 in the real rainfall simulation box 1 through a water supply pipe 18 through a hose 19.

[0040] The industrial computer 3 is connected to the display 2, the camera 4 under test, the stepping motor 7 and the water pump 14 respectively.

[0041] A drainage outlet and a corresponding rainwater recovery channel are provided at the bottom of the realistic rainfall simulation box 1 for collecting rainwater, and the recovery channel should not be located within the camera's field of view.

[0042] The described stepper motor 7 is used to control the rainfall direction of the main water pipe 6 to meet the rainfall blur effect of the perspective of the measured camera 4; the main sprocket 9 is coaxially connected to the output shaft of the stepper motor 7 and cooperates with the chain 11 to meet the rotational motion requirements; the chain 11 cooperates with the sprockets to meet the power transmission requirements; the driven sprocket 10 is fixed to each of the parallel main water pipes 6, and the main water pipe 6 is driven by the chain 11 to rotate; the types of the nozzles 12 are selected according to the test requirements, and different nozzle 12 specifications determine different rainfall amounts; the water divider 17 divides the rainwater provided by the water pump 14 into the parallel main water pipes 6; the main water pipe 6 is used to carry rainwater, one end is closed, and the other end is connected to the hose 19 as the water inlet end; since the main water pipe 6 needs to rotate during the test, the hose 19 is used for water supply, and the other end of the hose 19 is connected to the water supply pipe 18; the ball bearings are used to connect the main water pipe 6 to the overall support end of the platform, playing a role in support and rotation; the water pump 14 is used to provide water pressure for the rainfall simulation device; the filter 15 filters the water in the water tank 13 to prevent particulate matter from blocking the water pipes or nozzles; the water tank 13 is used to store the water source required for the test experiment.

[0043] As Figure 3 shown, based on the real rainfall simulation box 1 and combined with virtual traffic simulation, the present invention proposes a micro physical rainfall simulation method. The virtual traffic simulation provides complex road conditions and dynamic and static traffic participant information for the measured camera, and the real rainfall simulation box is responsible for simulating the blur effect caused by raindrops from the perspective of the camera.

[0044] Since different raindrop sizes correspond to different raindrop falling speeds in the real scenario, the size of the pre-simulated raindrops is determined before the test experiment, and then the falling speed is determined according to fluid mechanics to determine the motion characteristics of the raindrops. Since in the actual rainy-day vehicle driving process, the image blur of the camera perspective is caused by the synthesis of the vehicle moving forward and the raindrops moving vertically downward, therefore, the present invention proposes to synthesize the vehicle speed and the raindrop falling speed output by the simulation software to meet the image blur effect, and then determine the raindrop nozzle direction to ensure the fidelity of the micro physical rainfall simulation method. At the same time, the blur of the actual camera perspective is not only caused by the synthesis of raindrop size and motion, but also there is image blur caused by different rainfall amounts, and this blur is mainly manifested in the rainfall amounts formed by different nozzles in the vehicle forward direction. The present invention designs different types of nozzles arranged on the same water pipe to ensure the construction of the scenario that meets the test requirements of different rainfall amounts in the real rainfall simulation box 1.

[0045] The method for controlling the spraying angle of the nozzle 12 is as follows:

[0046] During the falling process of the raindrop, it is under the combined action of gravity, pressure difference resistance and air buoyancy to reach the force balance. The pressure difference resistance f received by the raindrop during the falling process is:

[0047]

[0048] Among them, ρ is the fluid density, S is the horizontal projection area, and C d is the air resistance coefficient. The motion equation under the combined action of the above three forces is:

[0049]

[0050] Among them, ρ 水滴 is the rainwater density, r is the raindrop radius. Let dv / dt = 0, and the raindrop velocity v rain can be finally obtained as:

[0051]

[0052] Through the above formula, the vertical falling velocity of raindrops corresponding to different raindrop radii is obtained, and through the vehicle motion velocity v veicle in the horizontal direction output by the simulation software at this moment, the motion synthesis is carried out to determine the raindrop ejection angle Φ and the velocity v sim of the ejected raindrops. The ejection angle of the nozzle 12 of the main water pipe 6 is adjusted by the stepper motor 7 respectively, and the water pump volume is controlled by the water pump 14 to construct a realistic rainfall simulation box, as Figure 1 shown in the overall structural schematic diagram of the in-loop test system of the present invention;

[0053] Figure 2 shown in the structural schematic diagram of the realistic rainfall simulation box of the present invention;

[0054] Figure 3 shown in the process schematic diagram of the in-loop test system of the present invention;

[0055] As shown in the figure, Figure 4 the above three figures (a)-(c) represent different raindrop ejection angles corresponding to different rainfall intensities at the same vehicle speed, Figure 4 and the following three figures (d)-(f) represent the raindrop ejection angles corresponding to different vehicle speeds under the same rainfall intensity.

[0056]

[0057]

[0058] Among them, v veicle represents the longitudinal vehicle speed in the test scenario; v rain represents the raindrop velocity in the test scenario;

[0059] In the present invention, the stepping motor 7 is used as the specific implementation device for controlling the direction of raindrops. The stepping motor 7 is an electromechanical actuator that converts an electrical pulse signal into a corresponding angular displacement or linear displacement. The core principle of the stepping motor's operation is that magnetic flux always passes along the path with the least magnetic resistance. The angular displacement of the stepping motor 7 is proportional to the number of input electrical pulses, and the rotational speed is proportional to the frequency of the input electrical pulses. That is, by controlling the number, frequency of the input electrical pulses, and the energization mode of the stator windings, the angular displacement, rotation angle, and rotation direction of the rotor of the stepping motor 7 can be controlled. Each time power is supplied, that is, each operation step, the rotor takes one step. When each phase winding is energized in turn, the rotor rotates by one tooth pitch. Therefore, the step angle is:

[0060]

[0061] In the formula, K is the energization coefficient. When the number of phases is equal to the number of beats, K = 1; otherwise, K = 2; m is the number of stator phases; z is the number of rotor teeth. If the frequency of the input electrical pulse signal of the stepping motor 7 is F, then the rotational speed of the stepping motor 7 is:

[0062]

[0063] The rotational speed of the stepping motor 7 represents the response speed of the platform to rainfall. The higher the frequency of the electrical pulse signal, the more the actual platform can respond to the vehicle speed change and then simulate the rainfall meteorology in real time.

[0064] At the same time, the selection of the water pump 14 should meet the working pressure standard range of the nozzle 12. To ensure the normal use of the water pump 14, the vacuum degree at the water intake of the water pump 14 cannot be too large. Otherwise, when the absolute pressure is lower than the air separation pressure of water, the air dissolved in the water will be separated to form bubbles, resulting in cavitation phenomenon, causing vibration and noise. Therefore, it is necessary to limit the vacuum degree at the water intake of the water pump 14 to be less than 0.03 MPa, which is limited by restricting the pumping height of the water pump 14. Usually, the pumping height should not be greater than 0.5 m. Pipe filters should be installed in the inlet and return channels of the water tank 1 for use by the rainfall equipment.

[0065] Specifically, the in-loop test process is as Figure 6 shown. In the existing in-loop test platform for cameras, the simulation of the rainfall environment is mostly directly provided by virtual simulation software, and its simulation degree is poor, unable to ensure that the camera obtains the image noise in the real rainfall environment. The present invention fuses physical rainfall simulation and virtual simulation image information to provide the tested camera with image information in a high-fidelity rainfall environment for closed-loop testing. The specific process is as follows:

[0066] The virtual simulation software outputs the image information of the camera's perspective and the vehicle's motion speed at the current moment. Among them, the vehicle motion speed information will be synthesized with the expected simulated rainfall speed in the real rainfall simulation box 1 to calculate the spraying angle of the nozzle 12 of the main water pipe 6, the pumping pressure of the water pump 14, and the number of opened main water pipes 6, and control the rotation of the nozzle 12 of the main water pipe 6 through the frequency of the input electrical pulse signal of the stepping motor 7 to simulate the rainfall state corresponding to the vehicle speed, forming the blurred noise of the real raindrop image. In addition, the image perspective information of the measured camera 4 provides the image information of the virtual traffic participants constructed by the virtual simulation software through the high-definition display 2. The refresh rate of the display 2 should be higher than that of the measured camera 4 to ensure the real-time nature of the image information received by the camera. The measured camera 4 receives the above two kinds of fused image information, as Figure 5 shown, and provides the target physical parameters required by the ADAS system, such as the target distance and target speed, etc., through the measured target detection algorithm and motion state estimation algorithm. The ADAS algorithm plans in combination with the input target physical parameters and outputs the lateral and longitudinal control information to the virtual simulation software to update the image information of the virtual simulation scene camera perspective in real time, completing the closed-loop test. The present invention can test the camera hardware, the camera image processing algorithm, the camera ECU, and the ADAS algorithm.

Claims

1. An in-loop test system for the performance of an intelligent vehicle camera in rainy weather, characterized in that: It includes a real rainfall simulation box, a display, an industrial control computer, and a camera under test. The display and the camera under test are respectively arranged on the left and right sides of the real rainfall simulation box. At least one side of the real rainfall simulation box facing the camera under test is automotive-grade glass; several main water pipes are arranged side by side above the real rainfall simulation box. Both ends of each main water pipe are pivotally connected to the box body of the real rainfall simulation box through bearings; a stepping motor is also arranged inside the real rainfall simulation box. The output shaft of the stepping motor is connected with a main sprocket. One end of each main water pipe is respectively provided with a driven sprocket. The main sprocket and the driven sprocket are connected by a chain drive; six different flow rate specifications of nozzles are equally spaced on the lower part of each main water pipe; they respectively include a fine rain nozzle, a light rain nozzle, a moderate rain nozzle, a heavy rain nozzle, a rainstorm nozzle, and a heavy rainstorm nozzle. The six different flow rate specifications of nozzles are used to simulate six different intensities of rainfall. The rainfall intensities are divided into fine rain, light rain, moderate rain, heavy rain, rainstorm, and heavy rainstorm. The rainfall amounts corresponding to all the nozzles of the same specification being fully opened should meet the following: in the case of no leakage and external natural factors within 1 hour, fine rain ≤ 5 mm, light rain 5 - 10 mm, moderate rain 10 - 25 mm, heavy rain 26 - 40 mm, rainstorm 50 - 80 mm, heavy rainstorm 90 - 120 mm; A water tank is arranged below the real rainfall simulation box. A water pump is arranged inside the water tank. The water pump is connected to a water delivery pipe, and the water delivery pipe is connected to a water distributor. The water distributor is provided with several water distribution ports. Each water distribution port is connected to the corresponding main water pipe inside the real rainfall simulation box through a water supply pipe; The industrial control computer is respectively connected to the display, the camera under test, the stepping motor, and the water pump; The specific process of the closed-loop test of the camera under test is as follows: The virtual simulation software outputs the image information of the camera view and the movement speed of the vehicle at the current moment. Among them, the vehicle movement speed information will be synthesized with the expected simulated rainfall speed in the real rainfall simulation box to calculate the spraying angle of the nozzles of the main water pipe. Open the nozzles corresponding to the rainfall intensity on the main water pipe, start the water pump, and control the rotation of the nozzles of the main water pipe through the frequency of the input electrical pulse signal of the stepping motor to simulate the rainfall state corresponding to the vehicle speed, forming real raindrop image blur noise; in addition, the image view information of the camera under test provides the image information of the virtual traffic participants constructed by the virtual simulation software through the high-definition display. The refresh rate of the display should be higher than the refresh rate of the camera under test to ensure the real-time nature of the image information received by the camera; the camera under test receives the above two fused image information, provides the target physical parameters required by the ADAS system through the target detection algorithm and the motion state estimation algorithm of the object under test. The ADAS algorithm combines the input target physical parameters for planning and outputs the lateral and longitudinal control information to the virtual simulation software to update the image information of the camera view of the real-time virtual simulation scene, completing the closed-loop test.

2. The in-loop test system for the performance of an intelligent vehicle camera facing rainy weather according to claim 1, characterized in that: The end of the water supply pipe is connected to the main water pipe through a hose; a drain port and a rainwater recovery channel are arranged at the bottom of the real rainfall simulation box.

3. The in-loop test system for the performance of an intelligent vehicle camera in rainy weather according to claim 1 or 2, characterized in that: The method for controlling the spraying angle of the nozzles is as follows: During the falling process of raindrops, the raindrops reach the force balance under the combined action of gravity, differential pressure resistance, and air buoyancy. The differential pressure resistance f suffered by raindrops during the falling process is as follows: where ρ is the fluid density, S is the horizontal projected area, and C s is the air resistance coefficient. The equation of motion under the combined action of the above three forces is as follows: where ρ 水滴 is the rainwater density, r is the raindrop radius. Let dv / dt = 0, and the raindrop velocity v rain is given by: Through the above formula, the vertical falling speed of raindrops corresponding to different raindrop radii is obtained, and through the horizontal vehicle movement speed v output by the simulation software at this moment veicle the motion synthesis is carried out to determine the raindrop spraying angle Φ and the speed v of the sprayed raindrops sim , respectively adjust the spraying angle of the nozzle of the main water pipe through the stepper motor, control the water pump volume through the water pump, and construct a realistic rainfall simulation box; where v veicle represents the longitudinal speed of the vehicle in the test scenario; v rain represents the raindrop speed in the test scenario; As a specific implementation device for controlling the direction of raindrops, the step angle of the stepper motor is: In the formula, K is the energization coefficient. When the number of phases is equal to the number of beats, K = 1; otherwise, K = 2; m is the number of stator phases; z is the number of rotor teeth. If the frequency of the input electrical pulse signal of the stepper motor is F, the rotational speed of the stepper motor is: The rotational speed of the stepper motor represents the response speed of the platform to rainfall. The higher the frequency of the electrical pulse signal, the more capable the actual platform is of responding to the vehicle speed change and then simulating the rainfall meteorology in real time.

4. The in-loop test system for the performance of an intelligent vehicle camera in rainy weather according to claim 1 or 2, characterized in that: The method for controlling the pumping pressure of the water pump is as follows: Limit the vacuum degree at the water intake of the water pump to be less than 0.03 MPa; limit it by restricting the pumping height of the water pump, and the pumping height is not greater than 0.5 m.

Images