Automatic windscreen wiper control method and system

Abstract

The invention discloses an automatic windscreen wiper control method and system, and relates to the technical field of windscreen wiper control, and the method comprises the steps: obtaining foreign matter information on the surface of a windscreen, and determining the position of a foreign matter in a vehicle coordinate system; acquiring sight information of a driver, and further determining an effective sight range; projecting the position of the foreign matter to a sight line coordinate system of a driver, and obtaining the shielding proportion of the effective sight line range shielded by the foreign matter; and the current driving state is obtained, and the windscreen wiper is controlled to work according to the current driving state, the foreign matter information and the shielding proportion and a preset triggering rule. According to the automatic windscreen wiper control method and system, the foreign matter information, the shielding proportion that the foreign matter information shields the effective sight line range and the current driving state are comprehensively considered, the windscreen wiper is controlled more accurately, intelligently and efficiently, and the technical problem that in the related technology, the windscreen wiper is started by mistake or is not started in time is solved.

Description

Technical Field

[0001] The present application relates to the technical field of wiper control, and in particular to an automatic wiper control method and system. Background Art

[0002] At present, wipers are an indispensable function for every car, especially in rainy environments. Wipers are closely related to the life safety of drivers and passengers. Although some vehicles are equipped with automatic wiper functions, how to use wipers efficiently and turn on wipers intelligently is even more important.

[0003] In the related art, a driving recorder is used to determine whether a target object appears on the windshield based on driving images, and then determine whether to control the wipers to operate.

[0004] However, since the above recognition process only relies on the camera and does not take into account the actual driving situation, the automatic opening of the wiper is not smart enough and is prone to false start or untimely start. Summary of the Invention

[0005] The present application provides an automatic wiper control method and system, which can solve the technical problem of incorrect or untimely wiper activation in the prior art.

[0006] In a first aspect, the present application provides an automatic wiper control method, the method comprising: Obtain information about foreign objects on the windshield surface and determine the position of the foreign objects in the vehicle coordinate system; Obtain the driver's line of sight information and then determine the effective line of sight range; Projecting the position of the foreign object onto the driver's line of sight coordinate system to obtain the occlusion ratio of the effective line of sight that is blocked by the foreign object; The current driving state is obtained, and according to the above current driving state, foreign object information and occlusion ratio, the wiper operation is controlled according to the preset triggering rules.

[0007] In conjunction with the first aspect, in one embodiment, obtaining foreign object information on the windshield surface specifically includes: Acquire optical data and extract a first time series feature through an optical sensor installed on the windshield; the optical sensor is provided with a light shield; Acquire capacitance data through a capacitance sensor installed on the inside of the windshield and extract a second time series feature; collecting image data of the windshield surface through a first camera and extracting image visual features; Fusing the first temporal feature, the second temporal feature, and the image visual feature to obtain a multimodal feature; The above multimodal features are input into the trained occlusion detection model to output the foreign object information on the windshield surface.

[0008] In conjunction with the first aspect, in one embodiment, determining the position of the foreign object in the vehicle coordinate system specifically includes: Define the vehicle coordinate system with the driver's eye point as the origin; Obtaining pixel coordinates of the foreign object in the image, and converting the pixel coordinates into camera rays in a camera coordinate system using the intrinsic parameters of the first camera; Based on the glass plane equation, the intersection of the ray and the windshield is solved to obtain the 3D coordinates of the foreign object in the camera coordinate system; The 3D coordinates are converted into the vehicle coordinate system by using the external parameters of the first camera to obtain the three-dimensional coordinates of the foreign object.

[0009] In combination with the first aspect, in one embodiment, after obtaining the three-dimensional coordinate system of the foreign object, the method further includes: The three-dimensional coordinates of the foreign matter are corrected using the optical data and capacitance data.

[0010] In conjunction with the first aspect, in one embodiment, obtaining the driver's line of sight information and then determining the effective line of sight range specifically includes: Acquire the driver's sight line information through a second camera, and determine the sight line direction based on the sight line information; Taking the above-mentioned sight direction as the reference axis, the sight range of the positive and negative first angle in the horizontal direction and the positive and negative second angle in the vertical direction is taken as the effective sight range; the above-mentioned first angle is greater than or equal to the second angle.

[0011] In conjunction with the first aspect, in one embodiment, based on the current driving state, foreign object information, and occlusion ratio, controlling the wiper operation according to preset triggering rules specifically includes: According to the above current driving state, foreign object information and occlusion ratio, based on preset trigger rules of different priorities, the trigger rule with the highest priority currently satisfied is determined, and the wiper operation is controlled according to the trigger rule.

[0012] In conjunction with the first aspect, in one embodiment, before determining the trigger rule with the highest priority currently satisfied, the method further includes: Input the above current driving state into the pre-trained intention matching model and output the driver's intention; The trigger ratio threshold and the windshield cleaning area are determined based on the above-mentioned driver intention.

[0013] In conjunction with the first aspect, in one embodiment, the trigger rules include, in descending order of priority, a security trigger rule and an intent matching trigger rule; The above safety trigger rule is: when the current speed of the vehicle is greater than the first speed threshold and the occlusion ratio is greater than the preset maximum ratio, the wipers are controlled to spray water and operate at high speed; The above intention matching trigger rule is: when the above occlusion ratio is greater than the above trigger ratio threshold, the wiper is controlled to work according to the above cleaning area and foreign object information; the above maximum ratio is greater than the trigger ratio threshold.

[0014] In combination with the first aspect, in one embodiment, when controlling the wiper to operate according to the intention matching trigger rule, the method further includes: Based on a preset mapping table of wiper operating speed reduction percentages corresponding to light intensity and vehicle speed, the wiper operating speed reduction percentage under the current light intensity and current vehicle speed is obtained to update the wiper operating speed.

[0015] In a second aspect, the present application provides an automatic wiper control system, the system comprising: An acquisition unit, which is used to acquire information about foreign objects on the windshield surface and determine the position of the foreign objects in the vehicle coordinate system; and acquire information about the driver's line of sight to determine the effective line of sight range; A decision control unit is used to project the position of the above-mentioned foreign object into the driver's line of sight coordinate system, obtain the obstruction ratio of the above-mentioned effective line of sight blocked by the foreign object; and control the wiper operation according to preset triggering rules based on the above-mentioned current driving state, foreign object information and obstruction ratio.

[0016] The beneficial effects of the technical solution provided by this application include: By obtaining foreign object information on the windshield surface and determining the position of the foreign object in the vehicle coordinate system; obtaining the driver's line of sight information, and then determining the effective line of sight range; and projecting the position of the above foreign object to the driver's line of sight coordinate system, obtaining the occlusion ratio of the above effective line of sight blocked by the foreign object; then obtaining the current driving state, and controlling the wiper operation according to preset triggering rules based on the above current driving state, foreign object information and occlusion ratio; comprehensively considering the foreign object information, the occlusion ratio of the effective line of sight blocked by the foreign object information and the current driving state, a more accurate, intelligent and efficient control of the wiper is achieved, and the technical problem of the wiper starting in error or untimely start-up in the related technology is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of an embodiment of the automatic wiper control method of the present application; Figure 2 This is a schematic diagram of the process of obtaining foreign object information on the windshield surface in an embodiment of the present application; Figure 3This is a schematic diagram of the architecture of an embodiment of the automatic wiper control system of the present application; Figure 4 This is a schematic diagram of the architecture of another embodiment of the automatic wiper control system of the present application. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0019] In a first aspect, an embodiment of the present application provides an automatic wiper control method.

[0020] In one embodiment, referring to Figure 1 , Figure 1 This is a flow chart of an embodiment of the automatic wiper control method of the present application. The automatic wiper control method includes: S1. Obtaining information about a foreign object on the windshield surface and determining the location of the foreign object in the vehicle coordinate system; S2. Obtain the driver's line of sight information and then determine the effective line of sight range; S3 projecting the position of the foreign object onto the driver's line of sight coordinate system to obtain the effective range of vision blocked by foreign objects occluded ratio; S4. Obtain the current driving state, and control the wipers according to preset triggering rules based on the current driving state, foreign object information, and occlusion ratio.

[0021] The order of the above steps S1 and S2 can be swapped.

[0022] In this embodiment, by obtaining foreign object information on the windshield surface and determining the position of the foreign object in the vehicle coordinate system; obtaining the driver's line of sight information and then determining the effective line of sight range; and projecting the position of the above foreign object onto the driver's line of sight coordinate system, obtaining the blockage ratio of the above effective line of sight blocked by the foreign object; then obtaining the current driving state, and controlling the wiper operation according to preset triggering rules based on the above current driving state, foreign object information and blockage ratio; comprehensively considering the foreign object information, the blockage ratio of the effective line of sight blocked by the foreign object information and the current driving state, achieving more accurate, intelligent and efficient control of the wiper, and solving the technical problem of incorrect or untimely start-up of the wiper in the related art.

[0023] Building on the above embodiment, in this embodiment, multiple optical sensors are installed on the vehicle's front windshield to cover the entire windshield. Light shields are also provided for the optical sensors to effectively reduce interference from external light. Capacitive sensors, preferably a capacitive sensor array, are installed on the inside of the windshield to cover the entire windshield area. A high-resolution lens is used to capture image information from the windshield.

[0024] Furthermore, in the above step S1, obtaining foreign matter information on the windshield surface specifically includes: Optical data is obtained and a first time series feature is extracted through an optical sensor installed on the windshield; a light shield is provided on the above optical sensor; capacitance data is obtained and a second time series feature is extracted through a capacitive sensor installed on the inside of the windshield; at the same time, image data of the surface of the above windshield is collected through a first camera and image visual features are extracted.

[0025] Then, the first time series feature, the second time series feature and the image visual feature are fused to obtain a multimodal feature.

[0026] Finally, the above multimodal features are input into the trained occlusion detection model to output the foreign object information on the windshield surface.

[0027] In this embodiment, by fusing the information collected by the optical sensor, the capacitive sensor and the first camera and using the trained occlusion detection model, the foreign object information on the windshield surface can be obtained, the category of the foreign object can be confirmed, and the subsequent control of the wiper working mode can be facilitated.

[0028] Optionally, optical sensors can capture: changes in light intensity (e.g., raindrops scatter light, causing a decrease in the light intensity received by the sensor; stains have different reflectivity than glass, causing differences in light intensity); and the angle of light reflection or transmission (e.g., transparent raindrops change the angle of light refraction, while non-transparent stains block light, causing sudden changes in angle). Furthermore, if multispectral optical sensors are used, spectral distribution can also be captured, as different objects absorb and reflect light differently, such as the different spectral characteristics of bird droppings and raindrops.

[0029] Capacitive sensors collect: the absolute values ​​of capacitance in different areas on the inside of the windshield to reflect whether the area is covered by an object and the dielectric properties of the object; the amount or rate of change of capacitance to reflect the dynamic changes of the object, such as the gradual increase in capacitance as raindrops accumulate and the fluctuation of capacitance as raindrops slide; spatial distribution data, through an array layout, can locate the position of objects on the glass, such as a "sudden change in capacitance in the upper left corner", indicating that there is an object in the upper left corner of the glass.

[0030] Optionally, as shown in Table 1 below, the optical data and capacitance data are continuous time series signals, and the “law of signal change over time” is mined through time series feature extraction to reflect the dynamic characteristics of the object.

[0031] Among them, the following feature types are obtained by temporal feature extraction: Trend characteristics: Long-term trend of light intensity / capacitance (e.g., continuous increase → dirt accumulation; continuous decrease → raindrops are scraped off) Fluctuation characteristics: signal period, frequency, and amplitude (e.g., raindrops vibrate with vehicle speed or wind speed, causing high-frequency fluctuations in light intensity or capacitance; bird droppings are static, with weak fluctuations); Correlation features: Correlation between sensor data at different times (e.g., if the capacitance value rises 1 second after the optical signal drops, the two synchronously verify the presence of an object). Mutation characteristics: the moment and amplitude of the signal’s mutation (e.g., a sudden drop in light intensity → instantaneous obstruction by large particles of stain).

[0032] Therefore, the "original electrical signal" is converted into a time series feature vector (such as rate of change, duration, and periodicity) that describes the dynamic behavior of the object, providing a basis for subsequent judgment of "whether the object is a dynamic raindrop or a static stain."

[0033] Optionally, the first camera captures static or dynamic images of the glass, and mines static characteristics such as "object shape, texture, position" through visual feature extraction, which complements the temporal features.

[0034] Among them, the IMS camera uses visual feature extraction to output feature types: Geometric characteristics: object shape (raindrop ≈ ellipse, bird droppings ≈ irregular shape), size (area, diameter), position (coordinates on the glass, such as (x=10cm, y=20cm)); Texture characteristics: surface roughness (raindrops are smooth → fine texture; bird droppings are rough → messy texture), edge clarity (raindrop edges are fuzzy; stain edges are clear); Motion characteristics (if it is an image sequence): the object's movement trajectory and speed (e.g., a raindrop slides to the right with the speed of the car, while bird droppings remain stationary); Color characteristics: the color difference between the object and the background (for example, bird droppings are yellowish-brown, and raindrops are transparent → close to the color of glass, but can be indirectly reflected through differences in light intensity).

[0035] Therefore, the "image pixels" are converted into visual feature vectors that describe the static shape of the object, providing a basis for subsequent judgments such as "whether the object is a raindrop or a stain" and "whether it is within the driver's line of sight."

[0036] Table 1

[0037] like Figure 2 As shown, optionally, the obtaining of foreign matter information on the windshield surface specifically includes: First, the optical data of the optical sensor and the capacitance data of the capacitive sensor are respectively subjected to time series feature extraction, and the visual feature extraction is performed on the camera image data; Then, the temporal features and visual features are fused to obtain multimodal features.

[0038] Finally, the fused multimodal features are processed by the attention mechanism, and the processed features are input into the trained occlusion detection model to decide and output the foreign object information on the windshield surface.

[0039] Optionally, if the foreign matter is determined to be raindrops, the wipers are automatically turned on for scraping; if the foreign matter is determined to be stains such as bird droppings, water is automatically sprayed and the wipers are turned on for scraping.

[0040] Specifically, multiple optical sensors can be evenly distributed across the windshield. Using principles such as light reflection, refraction, or scattering, they determine the presence of an object based on its reflected light, transmitted light, and angle. Furthermore, light shields can be installed around the optical sensors to reduce interference from external light.

[0041] At the same time, a capacitive sensor array can be installed on the inside of the front windshield to cover the entire front windshield area to make up for the insufficient detection of local areas by optical sensors. The changes in capacitance values ​​can also be filtered through software algorithms to reduce interference from environmental factors such as humidity.

[0042] In this embodiment, two cameras are installed, both of which are IMS (Intelligent Monitoring System) high-definition cameras. The first camera is used to capture images of the windshield. An image recognition algorithm is then used to analyze and identify objects in the images. This image recognition algorithm learns from the image features of objects such as raindrops and stains, builds a model to determine the presence of these objects in the input image, and analyzes their position, size, and shape. The second camera captures images of the driver's eyes for processing and analysis, specifically to calculate information such as eye movement angle and gaze direction.

[0043] Furthermore, in one embodiment, in the above step S1, determining the position of the foreign object in the vehicle coordinate system specifically includes: First, the vehicle coordinate system is defined with the driver's eye point as the origin.

[0044] Secondly, the pixel coordinates (u, v) of the foreign object in the image are obtained, and the pixel coordinates (u, v) are converted into camera rays in the camera coordinate system through the intrinsic parameters of the first camera, that is, the direction pointing from the first camera to the foreign object; wherein the image is the image of the windshield surface captured by the first camera.

[0045] Then, based on the glass plane equation, the intersection of the camera ray and the windshield is solved to obtain the 3D coordinates of the above foreign matter in the camera coordinate system.

[0046] Finally, the 3D coordinates are converted to the vehicle coordinate system using the external parameters of the first camera to obtain the three-dimensional coordinates (x, y, z) of the foreign object.

[0047] Optionally, the calibration board can be used to obtain the intrinsic parameters (such as focal length, principal point, etc.) and extrinsic parameters (position and posture of the camera in the vehicle coordinate system) of the first camera for conversion between 2D pixels and 3D coordinates.

[0048] Optionally, in this embodiment, after obtaining the three-dimensional coordinate system of the foreign object, the method further includes: The three-dimensional coordinates of the foreign matter are corrected using the optical data and capacitance data.

[0049] In this embodiment, the capacitive sensor array outputs the glass area where the foreign object is located, the optical sensor outputs the trigger position, and compares them with the results of the first camera. The error can be corrected through the filtering algorithm to improve the accuracy of the three-dimensional coordinates of the foreign object.

[0050] Furthermore, in one embodiment, in the above step S2, obtaining the driver's line of sight information and then determining the effective line of sight range specifically includes: First, the driver's line of sight information is obtained through the second camera, and the line of sight direction is determined based on the above line of sight information; Then, taking the above-mentioned sight direction as the reference axis, the sight range of the positive and negative first angle in the horizontal direction and the positive and negative second angle in the vertical direction is taken as the effective sight range; the above-mentioned first angle is greater than or equal to the second angle.

[0051] In this embodiment, the effective viewing range is defined based on the driver's viewing direction, thereby achieving an intelligent control effect that accurately matches the driver's actual attention focus.

[0052] Optionally, the first angle is 30°, and the second angle is 20°, that is, with the line of sight direction as the reference axis, the line of sight range of plus or minus 30° in the horizontal direction and plus or minus degrees in the vertical direction is used as the effective line of sight range.

[0053] Furthermore, in one embodiment, in the above step S4, the wiper operation is controlled according to a preset triggering rule based on the above current driving state, foreign object information, and occlusion ratio, specifically including: According to the above current driving state, foreign object information and occlusion ratio, based on preset trigger rules of different priorities, the trigger rule with the highest priority currently satisfied is determined, and the wiper operation is controlled according to the trigger rule.

[0054] In this embodiment, by dynamically determining the trigger rule with the highest priority that is currently satisfied, it is ensured that when multiple trigger rules conflict, the trigger rule with the highest priority can be executed according to the rule priority.

[0055] Furthermore, in this embodiment, before determining the trigger rule with the highest priority currently satisfied, the following steps are also included: First, the above current driving state is input into a pre-trained intention matching model, and the driver's intention is output.

[0056] Then, based on the above driver intention, the trigger ratio threshold and the cleaning area of ​​the windshield are determined.

[0057] Optionally, the cleaning area of ​​the windshield includes a central area and a side window area.

[0058] In this embodiment, by obtaining the driver's intention, the trigger ratio threshold and the cleaning area of ​​the windshield can be adjusted, thereby achieving the effect of effectively combining the driver's intention with precise wiper control.

[0059] Preferably, in this embodiment, the above trigger rules include security trigger rules and intention matching trigger rules in descending order of priority, ie, high-priority security trigger rules and low-priority intention matching trigger rules.

[0060] The above safety trigger rule is: when the current speed of the vehicle is greater than the first speed threshold and the occlusion ratio is greater than the preset maximum ratio, the wipers are controlled to spray water and operate at high speed; The above intention matching trigger rule is: when the above occlusion ratio is greater than the above trigger ratio threshold, the wiper is controlled to work according to the above cleaning area and foreign object information; the above maximum ratio is greater than the trigger ratio threshold.

[0061] In this embodiment, if the highest-priority trigger rule currently satisfied is a safety trigger rule, i.e., the vehicle's current speed is greater than a first speed threshold and the obstruction ratio is greater than a preset maximum ratio, the wipers are controlled to spray water and operate at high speed. If the highest-priority trigger rule currently satisfied is an intent-match trigger rule, i.e., the obstruction ratio is greater than the trigger ratio threshold, indicating that the safety trigger rule is not satisfied, the wipers are controlled to operate based on the cleaning area and foreign object information.

[0062] In this embodiment, by combining optical sensors, capacitive sensors, cameras with image recognition algorithms, intention recognition algorithms, etc., the wipers will automatically turn on when there are raindrops or stains on the windshield that affect the driver's line of sight. At the same time, it will automatically choose whether to spray water according to the type of stains, thereby controlling the wipers more accurately, intelligently and efficiently.

[0063] Preferably, when controlling the operation of the wiper according to the intention matching trigger rule, it also includes optimizing the operating parameters of the wiper according to the current light intensity and the current vehicle speed.

[0064] Specifically, based on the current light intensity and vehicle speed, the wiper operating parameters are optimized, including: Based on a preset mapping table of wiper operating speed reduction percentages corresponding to light intensity and vehicle speed, the wiper operating speed reduction percentage under the current light intensity and current vehicle speed is obtained to update the wiper operating speed.

[0065] In this embodiment, the current percentage of wiper speed reduction is determined by the current light intensity and the current vehicle speed, accurately matching the driving scene and effectively improving driving comfort and energy efficiency.

[0066] Optionally, before determining the triggering rule, it is necessary to first determine whether there are foreign objects such as raindrops or stains on the windshield; secondly, determine whether the foreign objects affect the driver's driving intention, which can be achieved in the following ways: 1. The second camera captures the driver's eye position, pupil direction and other line of sight information in real time, and calculates the three-dimensional direction of the line of sight (such as the horizontal and vertical angles of the line of sight).

[0067] Among them, characteristics can be obtained through hardware, movement can be confirmed by algorithm, and accuracy can be improved through calibration.

[0068] The hardware may include an infrared camera and a near-infrared light, both of which are deployed in front of the driver's seat. The infrared light highlights the pupil (dark spot) and corneal reflection (CR, high light spot) to resist interference from ambient light.

[0069] The algorithm works as follows: The pupil center is found through threshold segmentation and circular fitting, and the center of the CR is located through highlight detection. The relative position of the pupil and CR is calculated. Then, using facial key points (such as the corners of the eyes and the tip of the nose) and the PnP algorithm, the yaw (left-right rotation) and pitch (up-down rotation) angles of the head are estimated.

[0070] The relative rotation is: the pupil-CR vector is mapped to the rotation angle of the eyeball relative to the head, and the absolute direction is: final sight angle = head posture angle + relative eyeball rotation angle, outputting horizontal and vertical angles.

[0071] The calibration process is as follows: the driver looks at known 3D markers on the windshield to correct individual differences.

[0072] 2. Define the driver's effective field of view (e.g., a conical area centered on the line of sight with 30 degrees to the left and 20 degrees up and down). This range can be adjusted according to the actual scenario.

[0073] 3. Determine the three-dimensional position (x, y, z) of foreign objects (such as raindrops and stains) in the vehicle coordinate system through camera images.

[0074] 4. Projection matching judgment 1) Project the object’s 3D position into the driver’s line of sight coordinate system to determine whether it falls within the pre-set line of sight angle.

[0075] 2) If the projection angle of the foreign object is within the angular tolerance range of the line of sight (such as ±30 degrees horizontally and ±20 degrees vertically), it is judged as "within the line of sight", otherwise it is "out of range".

[0076] Optionally, when the second camera collects the driver's line of sight information, if the driver's line of sight stays at a fixed position on the front windshield for longer than a threshold, it is more likely to be judged that a foreign object is interfering with the driver.

[0077] Optionally, in one embodiment, determining whether an object exists includes: 1. Multimodal feature fusion through bidirectional LSTM + attention mechanism: Feature-level fusion: Features extracted from optical sensors, such as the light intensity change rate, capacitance sensor capacitance fluctuation frequency, and camera object texture roughness, are combined to form multimodal features.

[0078] The system dynamically learns the associations between sensors through a bidirectional LSTM + attention mechanism. For example, when the optical signal decreases and the capacitance value increases simultaneously (correlation > 0.8), the optical signal's weight decreases, while the capacitance signal's weight increases. At night, when the ambient light intensity is less than 5 lux, the weight of the camera's texture features is automatically reduced (to avoid low-light interference), while the weight of the optical or capacitance signal is increased. The increase in the optical or capacitance signal's weight is equal to the decrease in the camera's texture weight.

[0079] 2. Occlusion detection model for object existence judgment: Classifier design: Use random forest or lightweight CNN, input multimodal feature fusion, output the probability of object existence (such as P=0.92 indicates the presence of raindrops), and confirm the foreign object information with the highest probability of existence and output it.

[0080] Anti-interference mechanism: Through timing consistency check (for example, an object is confirmed to exist only after being detected in three consecutive frames), short-term interference (such as insects flying by) is filtered out.

[0081] Optionally, in one embodiment, determining the degree of line of sight obstruction specifically includes: If the effective visual range is blocked by foreign objects by more than 50%, the obstruction level is medium; If the effective visual range is blocked by foreign objects by more than 70% and the current vehicle speed is greater than the first speed threshold, emergency cleaning is triggered.

[0082] Optionally, in one embodiment, the driving intention recognition specifically includes: The current driving state is used as input features, and the current driving state includes: Vehicle status: current speed, steering wheel angle, brake pedal signal; Environmental data: navigation weather (such as heavy rain warnings), rainfall levels; Driver behavior: gaze dwell time, blink frequency (fatigue detection).

[0083] The algorithm architecture of the intent matching model uses a trained LSTM + hidden Markov model to output the intent probability distribution (such as an 85% probability of high-speed driving and a 10% probability of parking) to determine the driver's intent.

[0084] Optionally, intent-based scenario classification includes: High-speed scenarios: Trigger sensitivity is improved, with a trigger ratio threshold of 15%. This means interference is confirmed based on the driver's intention, and the wipers are activated when the occlusion ratio is greater than 15%. Congested scenario: The trigger threshold is increased, and the trigger ratio threshold is 30%, that is, the interference is confirmed in combination with the driver's intention, and it can only be started when the blocking ratio is greater than 30% to avoid frequent actions.

[0085] In this embodiment, the trigger rules include security trigger rules and intent matching trigger rules in descending order of priority.

[0086] Safety first: When the occlusion ratio is greater than 70% and the current vehicle speed is greater than 60km / h, the high-speed wipers and water spray are immediately activated; Intention matching: If the driver intends to change lanes or turn, the windshield cleaning area is the side window area of ​​the windshield (located by the capacitive sensor); if the driver intends to go straight, the windshield cleaning area is the center area of ​​the windshield (located by the capacitive sensor).

[0087] For comfort optimization: At night or in a tunnel, the wiper speed is reduced by 30% to reduce light reflection interference.

[0088] When the security trigger rule and the intent match trigger rule conflict, the priority is security > intent. For example: If severe obstruction and light rain are detected at the same time, emergency cleaning is performed first and weather parameters are ignored.

[0089] The method of this embodiment automatically triggers the wipers to clean when there is rain, stains, obstructions or other foreign matter on the vehicle's windshield if the triggering rules are met, thereby achieving more accurate, intelligent and efficient control of the wipers.

[0090] In a second aspect, an embodiment of the present application also provides an automatic wiper control system.

[0091] In one embodiment, referring to Figure 3 , Figure 3 FIG1 is a schematic diagram of the architecture of an embodiment of an automatic wiper control system of the present application. The automatic wiper control system includes an acquisition unit and a decision control unit.

[0092] The acquisition unit is used to acquire information about foreign objects on the windshield surface and determine the position of the foreign objects in the vehicle coordinate system; and to acquire information about the driver's line of sight and thereby determine the effective line of sight range.

[0093] The above-mentioned decision control unit is used to project the position of the above-mentioned foreign object into the driver's line of sight coordinate system, obtain the obstruction ratio of the above-mentioned effective line of sight blocked by the foreign object; and control the wiper operation according to the preset trigger rules based on the above-mentioned current driving state, foreign object information and obstruction ratio.

[0094] Furthermore, in one embodiment, the acquisition unit is further configured to: Acquire optical data and extract a first time series feature through an optical sensor installed on the windshield; the optical sensor is provided with a light shield; Acquire capacitance data through a capacitance sensor installed on the inside of the windshield and extract a second time series feature; collecting image data of the windshield surface through a first camera and extracting image visual features; Fusing the first temporal feature, the second temporal feature, and the image visual feature to obtain a multimodal feature; The above multimodal features are input into the trained occlusion detection model to output the foreign object information on the windshield surface.

[0095] Furthermore, in one embodiment, the acquisition unit is further configured to: Define the vehicle coordinate system with the driver's eye point as the origin; Obtaining pixel coordinates of the foreign object in the image, and converting the pixel coordinates into camera rays in a camera coordinate system using the intrinsic parameters of the first camera; Based on the glass plane equation, the intersection of the ray and the windshield is solved to obtain the 3D coordinates of the foreign object in the camera coordinate system; The 3D coordinates are converted into the vehicle coordinate system by using the external parameters of the first camera to obtain the three-dimensional coordinates of the foreign object.

[0096] Furthermore, in one embodiment, the acquisition unit is further configured to: The three-dimensional coordinates of the foreign matter are corrected using the optical data and capacitance data.

[0097] Furthermore, in one embodiment, the acquisition unit is further configured to: Acquire the driver's sight line information through a second camera, and determine the sight line direction based on the sight line information; Taking the above-mentioned sight direction as the reference axis, the sight range of the positive and negative first angle in the horizontal direction and the positive and negative second angle in the vertical direction is taken as the effective sight range; the above-mentioned first angle is greater than or equal to the second angle.

[0098] Furthermore, in one embodiment, the decision control unit is further configured to: According to the above current driving state, foreign object information and occlusion ratio, based on preset trigger rules of different priorities, the trigger rule with the highest priority currently satisfied is determined, and the wiper operation is controlled according to the trigger rule.

[0099] Furthermore, in one embodiment, the decision control unit is further configured to: Input the above current driving state into the pre-trained intention matching model and output the driver's intention; The trigger ratio threshold and the windshield cleaning area are determined based on the above-mentioned driver intention.

[0100] Furthermore, in one embodiment, the trigger rules include, in descending order of priority, a security trigger rule and an intent matching trigger rule; The above safety trigger rule is: when the current speed of the vehicle is greater than the first speed threshold and the occlusion ratio is greater than the preset maximum ratio, the wipers are controlled to spray water and operate at high speed; The above intention matching trigger rule is: when the above occlusion ratio is greater than the above trigger ratio threshold, the wiper is controlled to work according to the above cleaning area and foreign object information; the above maximum ratio is greater than the trigger ratio threshold.

[0101] Furthermore, in one embodiment, the decision control unit is further configured to: Based on a preset mapping table of wiper operating speed reduction percentages corresponding to light intensity and vehicle speed, the wiper operating speed reduction percentage under the current light intensity and current vehicle speed is obtained to update the wiper operating speed.

[0102] Alternatively, as Figure 4 As shown, the above system further includes a sensor layer, which includes an optical sensor, a capacitive sensor and two IMS cameras.

[0103] The acquisition unit includes a data acquisition layer and a data fusion layer.

[0104] The data acquisition layer includes a signal conversion module and a data preliminary processing module, which are used to collect data from various sensors. Specifically, for sensors, it collects the electrical signals they output and converts them into digital signals; for cameras, it collects image data.

[0105] The data fusion layer includes a feature extraction module and a fusion model module. These modules use data-level fusion to directly fuse data from different sensors. For example, they combine light intensity changes detected by an optical sensor, capacitance changes detected by a capacitive sensor, and camera image data to form a multidimensional data vector. Machine learning algorithms (such as neural networks) are then used to train and analyze this fused data to establish an object occlusion detection model.

[0106] Specifically, a bidirectional LSTM is used to extract and process data from the optical and capacitive sensors. MobileNetV3 is used to extract visual features from the image data. The extracted feature data is then fused through channel stacking or dot multiplication. The fused data is then fed into a self-attention mechanism, and the final output data is determined.

[0107] The decision control unit comprises a decision layer and a control layer. The decision layer includes an analysis and decision module, which comprehensively determines the current triggering rules to determine whether to trigger the wipers and generates a decision instruction that is sent to the control layer. The control layer also includes a device control module and a driver reminder module, which control the wipers according to the decision instruction and provide reminders to the driver.

[0108] Specifically, the workflow of the system in this embodiment includes: A1. The system uses optical sensors, capacitive sensors, and a first camera to collect real-time information about foreign objects on the windshield surface. A second camera tracks the driver's gaze, including movement and direction, and simultaneously captures the current driving status. A2. The data acquisition layer converts various sensor signals into digital signals and performs preliminary processing on line of sight information in preparation for subsequent analysis.

[0109] A3. The data fusion layer extracts features from the optical sensor, capacitive sensor, and primary camera data, as well as intent recognition-related data (i.e., the current driving state), and then normalizes the data to ensure uniform data scale. A4. A feature-level fusion algorithm is used to correlate and combine the data features from the optical sensor, capacitive sensor, and the first camera to form a multimodal feature. A5. The decision layer uses multimodal features to determine whether there is a foreign object on the windshield. It then uses the focus and direction of the gaze determined by the eye-tracking algorithm to determine the percentage of the foreign object obstructing the view. It then uses the intention recognition algorithm to determine the driver's intent and determines whether to trigger the wipers according to established triggering rules. A6. The control layer starts and stops the wipers, adjusts speed, and switches wiper modes based on instructions from the decision layer. A7. The system provides driver feedback on windshield conditions and wiper operating status via in-vehicle displays and voice prompts. During wiper operation, the system continuously monitors sensor data and changes in driving intent to dynamically optimize wiper operation.

[0110] Among them, the functional implementation of each module in the above-mentioned automatic wiper control system corresponds to the various steps in the above-mentioned automatic wiper control method embodiment, and its functions and implementation processes will not be repeated here one by one.

[0111] It should be noted that the serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0112] The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices. The terms "first", "second" and "third" are used to distinguish different objects, etc., and do not represent a sequence, nor do they limit the "first", "second" and "third" to different types.

[0113] In the description of the embodiments of this application, the words "exemplary," "for example," or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary," "for example," or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "for example," or "for example" is intended to present the relevant concepts in a concrete manner.

[0114] In the description of the embodiments of the present application, unless otherwise specified, “ / ” means or, for example, A / B can mean A or B; “and / or” in the text is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, “multiple” refers to two or more than two.

[0115] In some processes described in the embodiments of the present application, multiple operations or steps are included that appear in a specific order. However, it should be understood that these operations or steps may not be performed in the order in which they appear in the embodiments of the present application or may be performed in parallel. The sequence numbers of the operations are only used to distinguish between different operations, and the sequence numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations or steps may be performed in sequence or in parallel, and these operations or steps may be combined.

[0116] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, or the part that contributes to the existing technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above and includes a number of instructions for enabling a terminal device to execute the methods described in each embodiment of this application.

[0117] The above are only preferred embodiments of the present application and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for controlling an automatic wiper, characterized in that: The method comprises: Obtain information about foreign objects on the windshield surface and determine the position of the foreign objects in the vehicle coordinate system; Obtain the driver's line of sight information and then determine the effective line of sight range; Projecting the position of the foreign object onto the driver's sight line coordinate system to obtain an occlusion ratio of the effective sight line range blocked by the foreign object; The current driving state is obtained, and according to the current driving state, foreign object information and occlusion ratio, the wiper is controlled to work according to preset triggering rules.

2. The automatic wiper control method according to claim 1, wherein: Obtain information about foreign objects on the windshield surface, including: Acquiring optical data and extracting a first time series feature through an optical sensor installed on the windshield; the optical sensor is provided with a light shield; Acquire capacitance data through a capacitance sensor installed on the inside of the windshield and extract a second time series feature; collecting image data of the windshield surface by a first camera and extracting image visual features; Fusing the first time series feature, the second time series feature, and the image visual feature to obtain a multimodal feature; The multimodal features are input into a trained occlusion detection model to output foreign object information on the windshield surface.

3. The automatic wiper control method according to claim 2, wherein: Determine the position of the foreign object in the vehicle coordinate system, specifically including: Define the vehicle coordinate system with the driver's eye point as the origin; Obtaining pixel coordinates of the foreign object in the image, and converting the pixel coordinates into camera rays in a camera coordinate system using the intrinsic parameters of the first camera; Based on the glass plane equation, the intersection of the ray and the windshield is solved to obtain the 3D coordinates of the foreign object in the camera coordinate system; The 3D coordinates are converted into a vehicle coordinate system using the external parameters of the first camera to obtain the three-dimensional coordinates of the foreign object.

4. The automatic wiper control method according to claim 3, wherein: After obtaining the three-dimensional coordinate system of the foreign object, the method further includes: The three-dimensional coordinates of the foreign object are corrected using the optical data and the capacitance data.

5. The automatic wiper control method according to claim 1, wherein: Obtain the driver's line of sight information and then determine the effective line of sight range, including: Acquiring the driver's sight line information through a second camera, and determining the sight line direction based on the sight line information; Taking the sight direction as the reference axis, the sight range of the positive and negative first angle in the horizontal direction and the positive and negative second angle in the vertical direction is taken as the effective sight range; the first angle is greater than or equal to the second angle.

6. The automatic wiper control method according to claim 1, wherein: According to the current driving state, foreign object information and occlusion ratio, the wiper is controlled to operate according to a preset triggering rule, specifically including: According to the current driving state, foreign object information and occlusion ratio, based on preset trigger rules of different priorities, the trigger rule with the highest priority currently satisfied is determined, and the wiper operation is controlled according to the trigger rule.

7. The automatic wiper control method according to claim 6, wherein: Before determining the highest priority trigger rule currently satisfied, it also includes: Inputting the current driving state into a pre-trained intention matching model and outputting the driver's intention; A trigger ratio threshold and a cleaning area of ​​the windshield are determined according to the driver's intention.

8. The automatic wiper control method according to claim 7, wherein: The trigger rules include security trigger rules and intent matching trigger rules in descending order of priority; The safety trigger rule is: when the current speed of the vehicle is greater than a first speed threshold and the occlusion ratio is greater than a preset maximum ratio, the wipers are controlled to spray water and operate at high speed; The intention matching trigger rule is: when the occlusion ratio is greater than the trigger ratio threshold, the wiper is controlled to work according to the cleaning area and foreign object information; the maximum ratio is greater than the trigger ratio threshold.

9. The automatic wiper control method according to claim 8, wherein: When controlling the wiper operation according to the intent matching trigger rule, it also includes: Based on a preset mapping table of wiper operating speed reduction percentages corresponding to light intensity and vehicle speed, the wiper operating speed reduction percentage under the current light intensity and current vehicle speed is obtained to update the wiper operating speed.

10. An automatic wiper control system, characterized in that: The system comprises: An acquisition unit, which is used to acquire information about foreign objects on the windshield surface and determine the position of the foreign objects in the vehicle coordinate system; and acquire information about the driver's line of sight to determine the effective line of sight range; A decision control unit is used to project the position of the foreign object into the driver's line of sight coordinate system, obtain the occlusion ratio of the effective line of sight blocked by the foreign object; and control the wiper operation according to preset triggering rules based on the current driving state, foreign object information and occlusion ratio.

Images