Automobile, electronic rearview mirror with cleaning function and control method of electronic rearview mirror

Abstract

The invention discloses an automobile, an electronic rearview mirror with a cleaning function and a control method of the electronic rearview mirror, and relates to the technical field of electronic rearview mirrors. The control method comprises the following steps: acquiring an image frame shot by the electronic rearview mirror and vehicle driving state data; and judging the current illumination scene according to the illumination condition and the gain parameter, pre-judging a tunnel or basement entrance, and generating a scene flag bit. And carrying out region-of-interest identification processing and brightness normalization processing on the image frame. And calculating a multi-dimensional characteristic index based on the preprocessed data and the vehicle driving state data. And performing interference locking judgment. If the vehicle is not locked, the fogging state or the dirty state is distinguished according to the scene flag bit and the multi-dimensional feature index, and a control instruction is generated in combination with the vehicle driving state. And executing a corresponding heating demisting or blowing cleaning action according to the control instruction. After the action is executed, the image is collected again, the image quality score and improvement amount are calculated, whether the action is stopped or not is determined according to the acceptance check result, and the execution dosage is adjusted or the reference parameters are updated.

Description

Technical Field

[0001] This invention relates to the field of electronic rearview mirror technology, and more specifically, to an electronic rearview mirror for automobiles with a cleaning function and a control method thereof. Background Technology

[0002] Electronic rearview mirror systems, as an important auxiliary device for safe driving, capture the external environment through cameras and display the images on an in-vehicle screen. This provides drivers with a wider field of vision and effectively reduces blind spots inherent in traditional optical rearview mirrors, leading to their increasingly widespread application in commercial and passenger vehicles. With the development of automotive intelligence, drivers are placing higher demands on the reliability of electronic rearview mirror systems in various harsh environmental conditions such as rain, snow, and sandstorms. Maintaining a clear field of vision has become a crucial prerequisite for ensuring driving safety.

[0003] In practical use, the camera lens of electronic rearview mirrors is easily affected by the external environment. In rainy or snowy weather, it can be covered by moisture and frost, while in windy or sandy conditions, dust can easily accumulate. These contaminants can cause a serious decline in image quality or even complete failure, preventing drivers from obtaining clear information about the external environment in a timely manner and significantly increasing driving risks. To address this issue, some existing technologies have incorporated cleaning functions for electronic rearview mirrors, but the triggering methods for these cleaning operations have significant limitations.

[0004] Current electronic rearview mirror cleaning functions lack the ability to actively recognize image clarity. They cannot automatically detect whether the lens is contaminated or whether the image quality has deteriorated, and the cleaning operation can only be triggered manually by the driver. This reliance on manual judgment and operation not only fails to respond promptly to lens contamination, resulting in unresolved blurred vision, but also distracts the driver, causing additional inconvenience and safety hazards. Especially in complex road conditions, the lag and operational risks of manually triggering cleaning are even more pronounced, making it difficult to meet the vehicle's need for a consistently clear field of vision under various operating conditions. Summary of the Invention

[0005] The present invention provides an automobile, an electronic rearview mirror with a cleaning function, and a control method thereof to improve at least one of the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides a control method for an electronic rearview mirror with a cleaning function, comprising steps S1 to S6.

[0007] S1. Acquire image frames captured by the electronic rearview mirror, as well as vehicle driving status data.

[0008] S2. Determine the current lighting scene based on the lighting conditions and gain parameters, predict the tunnel or basement entrance, and generate scene markers.

[0009] S3. Perform region of interest (ROI) identification and brightness normalization on the image frame.

[0010] S4. Based on the preprocessed data and vehicle driving status data, calculate multi-dimensional feature indicators. The indicators include at least: contrast and sharpness reflecting image details, halo index reflecting nighttime scattering, dew point margin reflecting the risk of physical fogging, stable spot characteristics reflecting lens dirt, noise intensity reflecting image quality interference, and a unified image quality score.

[0011] S5. Interference locking is determined based on noise intensity. If not locked, the fogging or dirt conditions are distinguished based on scene flags and multi-dimensional feature indicators, and control commands are generated in conjunction with the vehicle's driving status.

[0012] S6. Execute the corresponding heating defogging or air blowing cleaning actions according to the control instructions. After the actions are executed, re-acquire images, calculate the image quality score and improvement amount, and decide whether to stop the action, adjust the execution dosage, or update the reference baseline parameters based on the acceptance results.

[0013] Secondly, the present invention provides an electronic rearview mirror with a cleaning function, comprising a camera with a heating function, and a cleaning component configured to blow air toward the lens of the camera.

[0014] The electronic rearview mirror is adapted to perform a control method for an electronic rearview mirror with a cleaning function as described in any paragraph of the first aspect.

[0015] Thirdly, the present invention provides an automobile equipped with an electronic rearview mirror with a cleaning function as described in the second aspect.

[0016] By adopting the above technical solution, the present invention can achieve the following technical effects: The electronic rearview mirror control method of this invention achieves intelligent identification and precise control of lens contamination and fogging by integrating multi-dimensional image feature analysis and vehicle driving status data. This method can dynamically adjust the detection strategy based on lighting conditions, noise intensity, and scene prediction, effectively distinguishing between lens dirt, fogging, and transient interference, and avoiding false triggering caused by sudden changes in ambient light or sensor noise.

[0017] The control method combines physical dew point margin calculation with closed-loop verification of image quality scoring. It can proactively activate anti-fogging in scenarios prone to condensation, such as tunnels and underground parking garages, and evaluate the improvement effect in real time after cleaning or defogging, adaptively adjusting execution parameters to maintain clear visibility under various operating conditions. Furthermore, by introducing vehicle speed gating, frequency-sweeping pulse blowing, and a reference parameter self-learning mechanism, the effectiveness of cleaning actions and the adaptability and reliability of the control method are further improved, significantly reducing the need for manual intervention and enhancing driving safety. Attached Figure Description

[0018] Figure 1 This is a front view of the electronic rearview mirror.

[0019] Figure 2 This is a front view of the electronic rearview mirror with the outer casing components removed.

[0020] Figure 3 It is an axonometric view of the electronic rearview mirror from the first perspective.

[0021] Figure 4 It is an axonometric view of the electronic rearview mirror from a second perspective.

[0022] Figure 5 This is a rear view of the electronic rearview mirror with the fixed components hidden.

[0023] Figure 6 This is a rear view of the electronic rearview mirror with the housing assembly hidden.

[0024] Figure 7 It is an isometric view of the camera module from a first-person perspective.

[0025] Figure 8 It is an isometric view of the camera component from a second perspective.

[0026] Figure 9 This is a schematic diagram of the structure of an electronic rearview mirror.

[0027] Figure 10 This is a simplified logic diagram of the control method.

[0028] Figure 11 This is a logic diagram of the control method.

[0029] Markings in the diagram: 1-Mounting hole, 2-Fixing component, 3-Supporting component, 4-Solenoid valve, 5-Housing component, 6-Blowing component, 7-First camera, 8-Shooting groove, 9-Second plane, 10-Second camera, 11-Blowing hole, 12-Lens bracket, 13-Camera assembly, 14-Display screen Detailed Implementation The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings in the embodiments of the present invention.

[0030] Example 1, please refer to Figures 1 to 9A first embodiment of the present invention provides an electronic rearview mirror with a cleaning function, comprising a housing assembly suitable for mounting on a vehicle body, a camera assembly 13 having a first camera 7, and a cleaning assembly configured to blow air toward the lens of the first camera 7. The camera assembly 13 and the cleaning assembly are coupled to the housing assembly. The cleaning assembly includes an air blowing element 6 adapted to blow air toward the lens, and a solenoid valve 4 coupled between the air blowing element 6 and an air source. Preferably, the electronic rearview mirror further includes a control module and a display screen 14. The control module is electrically connected to the camera assembly 13, the display screen 14, and the cleaning assembly.

[0031] like Figures 1 to 5 As shown, the housing assembly includes a fixing member 2, a support member 3 and an outer shell member 5 engaged with the fixing member 2. A receiving cavity is formed between the fixing member 2 and the outer shell member 5 to accommodate the support member 3, the camera assembly 13, and the cleaning assembly. The fixing member 2 is provided with a mounting hole 1, and the support member 3 is adapted to be engaged with the vehicle body through the mounting hole 1.

[0032] like Figure 3 As shown, the housing assembly is provided with a shooting recess 8. The shooting recess 8 is constructed with a first plane facing the rear of the vehicle body and a second plane 9 close to the vehicle body. The first camera 7 is located on the first plane. The air blowing element 6 is located on the second plane 9.

[0033] like Figure 2 , Figure 4 , Figure 5 and Figure 6 As shown, the camera assembly 13 also includes a second camera 10. The first camera 7 is configured to capture images of the rear side of the vehicle body. The second camera 10 is configured to capture images below the first camera 7. Preferably, the included angle A between the shooting axes of the first camera 7 and the second camera 10 is 105° to 120°.

[0034] like Figures 6 to 8 As shown, the camera assembly 13 includes a lens bracket 12. The first camera 7 is attached to the lens bracket 12. The lens bracket 12 is attached to the housing assembly. The air blower 6 is attached to the lens bracket 12 and / or the housing assembly.

[0035] like Figure 7 As shown, the air blowing component 6 is provided with an air blowing hole 11 facing the lens. Preferably, one end of the air blowing component 6 is constructed as a columnar structure. The air blowing hole 11 is disposed on the side of the columnar structure. The axis of the columnar structure is parallel to the shooting axis of the first camera 7.

[0036] Preferably, the camera has a heating function to heat and defog.

[0037] Example 2: The present invention provides a car equipped with an electronic rearview mirror with a cleaning function as described in Example 1.

[0038] Example 3: The present invention provides a control method for an electronic rearview mirror with a cleaning function, which includes steps S1 to S6.

[0039] S1. Acquire image frames captured by the electronic rearview mirror, as well as vehicle driving status data. Preferably, the vehicle driving status data includes: illumination value, image sensor gain, and a marker indicating that the vehicle is about to enter a tunnel or underground parking garage.

[0040] First, acquire the raw image frames from the camera. and maintain a record containing the past A historical frame buffer queue is used for subsequent temporal median filtering analysis. Simultaneously, multimodal data, including ambient temperature, is synchronously acquired via the CAN bus or onboard sensors. relative humidity Module temperature Speed Current image sensor gain and light sensor values .

[0041] S2. Determine the current lighting scene based on illumination conditions and gain parameters, and predict the tunnel or basement entrance, generating scene markers. Before performing specific image quality analysis, the current macro scene is first determined, which determines the branching direction of subsequent algorithms.

[0042] First, a night / low-light gating is performed. When the light value is lower than the night light threshold or the image sensor gain is higher than the preset gain threshold, the night mode flag is set. In night mode, subsequent fog detection will use the halo index as the primary criterion.

[0043] .

[0044] In the formula This is the night mode indicator. This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise. For a moment Illumination value. This represents the nighttime light threshold. For a moment Image sensor gain. This is the gain threshold. For logical "OR".

[0045] In this embodiment, image sensor gain is used to determine nighttime conditions. In other embodiments, an equivalent technical feature, ISO value, can be used, which is also within the scope of protection of this invention. In nighttime mode, since high gain noise can interfere with edge detection, subsequent fogging determination will no longer primarily rely on the sharpness gradient, but will instead rely on the "halo index".

[0046] Then, a tunnel or underground parking garage prediction gating system is implemented. When an intruder is about to enter a tunnel or underground parking garage, or when the illumination value changes abruptly within an adjacent cycle, a tunnel prediction flag is set. When the tunnel prediction flag is set, an active anti-fog maintenance mode is activated, and heating is turned on in advance to increase the dew point margin.

[0047] .

[0048] In the formula It serves as a warning sign for tunnels or underground parking garages. This is a sign indicating that you are about to enter a tunnel or underground parking garage. For a moment Illumination value. This represents the threshold for sudden changes in light intensity. This is an indicator function. For a moment Illumination value. For logical "OR".

[0049] Tunnel Prediction Marker At that time, regardless of the current image quality, the system will prioritize entering the "active anti-fog maintenance mode" and turn on the heating in advance to increase the dew point margin and prevent condensation when entering the tunnel.

[0050] S3. Perform region of interest (ROI) identification and brightness normalization on the image frames. Through ROI selection and image preprocessing, subsequent metrics are made to reflect the "lens status" as much as possible, rather than false degradation caused by AE / gain fluctuations, localized strong light, or informationless areas.

[0051] The preferred method for identifying regions of interest is as follows.

[0052] When the lighting scene is during the day, exclude overexposed areas and areas where the gradient magnitude or texture energy is below the threshold to obtain the region of interest.

[0053] When the lighting scene is at night, identify pixels whose brightness exceeds the threshold, use them as the center of the point light source, and include the surrounding annular area in the region of interest.

[0054] Subsequent indicators are all in (e.g.: (etc.) should be limited to This is done on the screen to reduce false triggering caused by "non-lens degradation factors".

[0055] Region of Interest Mask The generation logic is as follows (those skilled in the art can select and combine the following features according to actual needs, and this invention does not specifically limit this). 1. Exclude highly saturated areas (overexposed areas caused by headlights / direct sunlight). Specifically, traverse each pixel within the candidate area. If its brightness value is greater than the saturation threshold, remove the pixel from the region of interest mask. Overexposed pixels (such as headlights, direct sunlight, reflections) have lost texture details, and their gradient information is invalid. These areas will introduce noise when calculating contrast or sharpness, and their changes are independent of the lens state. 2. Exclude sky / solid color / low texture areas, prioritizing road / vehicle areas (beneficial for sharpness / contrast stability). Specifically, within the candidate area, calculate the gradient variance or entropy of each small local block (such as 8x8 pixels). If the variance is lower than the texture threshold, the block is considered to lack texture and is removed from the region of interest. Combining the color model (the sky is usually blue or gray) and image position (the sky is located at the top of the image), directly exclude large areas at the top. Areas like the sky and solid-color walls inherently lack texture, resulting in naturally low and fluctuating sharpness. Including them makes image quality scores overly sensitive to scene changes rather than lens smudges / fog. Night mode can retain areas containing point light sources to calculate halo metrics. .

[0056] To eliminate the interference of automatic exposure (AE) fluctuations on image quality evaluation, this embodiment performs brightness normalization processing on the acquired images. The specific brightness normalization processing is as follows.

[0057] .

[0058] In the formula Represents the normalized pixel The value of . Indicates the original input frame in pixels The brightness value. Indicates the region of interest Internal to the original input frame Find the mean. To prevent numbers from being divided by zero.

[0059] For a moment The region of interest (pixel set / mask). This embodiment uses brightness values ​​for calculation; in other embodiments, grayscale values ​​can also be used for calculation, and such equivalent technical solutions are also within the protection scope of this invention.

[0060] A brightness-normalized image is used to suppress the effects of exposure fluctuations. This normalization process makes the subsequently calculated sharpness and contrast metrics insensitive to overall drift in automatic exposure and gain. Through this process, the calculated metrics only reflect changes in the image's texture structure and are insensitive to overall drift in ambient light intensity. Furthermore, a smoothed and denoised reference frame is generated based on the brightness-normalized image. This is used for subsequent noise intensity estimation. S4. Based on the preprocessed data and vehicle driving status data, calculate multi-dimensional feature indicators. The indicators include at least: contrast and sharpness reflecting image details, halo index reflecting nighttime scattering, dew point margin reflecting the risk of physical fogging, stable spot characteristics reflecting lens dirt, noise intensity reflecting image quality interference, and a unified image quality score.

[0061] The contrast ratio is calculated using the RMS contrast ratio formula, as follows.

[0062] .

[0063] .

[0064] In the formula For a moment RMS contrast ratio. This is for normalized contrast. For a moment The region of interest. The mean brightness within the region of interest. This is a brightness-normalized image. :bucket Reference contrast. Index the scene bucket. Represents the normalized pixel The value of . To prevent numbers from being divided by zero.

[0065] The resolution index is calculated using the variance of the Laplacian operator, as follows.

[0066] .

[0067] .

[0068] In the formula For a moment Clarity metrics. To normalize the sharpness. The variance operator is defined for the region of interest. For the Laplace operator. The reference sharpness for scene bucket b. Index scene buckets (such as daytime, nighttime, etc.). This is a brightness-normalized image.

[0069] At night, fog can cause noticeable scattered halos around streetlights or vehicle headlights. This characteristic is quantified by calculating the energy ratio between the core of a point light source and its surrounding annular region. This indicator... It is a key criterion for identifying fog or condensation at night. In this embodiment, the nighttime halo index is obtained by extracting a set of point light sources and calculating the halo energy ratio for each point light source.

[0070] First, the point light source set is extracted: Highlighted connected components are extracted from the region of interest at night to obtain the point light source set. Then, based on each point light source... Calculate the halo energy ratio. Finally, summarize the overall halo index for the night (e.g., the mean).

[0071] .

[0072] .

[0073] In the formula Point light source The halo index. For pixel position index. For Center, radius arrive The ring-shaped region. For Center, radius The core area. Index the point light source (connected components / bright spot targets). Represents the normalized pixel The value of . Index for the current time. To prevent numbers from being divided by zero. This is a radius parameter used to achieve either a fixed or adaptive configuration. For a moment The overall halo index. Represents a set All Operator for finding the mean. For a moment A collection of point light sources.

[0074] Halo Index Used to identify enhanced scattering caused by fogging or condensation in nighttime scenes. Fogging / condensation enhances scattering around point light sources, typically leading to... A significant increase has been observed, therefore nighttime fog formation should rely more on... .

[0075] Preferably, the multidimensional feature index also includes the daytime fog index and the nighttime fusion fog index.

[0076] .

[0077] .

[0078] In the formula This is the daytime fog index. This represents the weighting coefficient for the first day. This is the weighting coefficient for the second day. This is the nighttime fog index. This is the weighting coefficient for the first night. This is the weighting coefficient for the second night.

[0079] This embodiment is based on ambient temperature. and relative humidity Calculating dew point temperature using the Magnus formula And further calculate the dew point margin. .when When the value is close to or less than 0, the risk of physical fogging is extremely high.

[0080] .

[0081] .

[0082] .

[0083] In the formula These are intermediate variables in the Magnus formula. It is the first Magnus constant. It is the second Magnus constant. The ambient temperature. It is the natural logarithm. Relative humidity (in %). This is the dew point temperature. This refers to the module temperature. This refers to the dew point margin.

[0084] Fog index combined with dew point margin Used for comprehensive judgment of fogging conditions. Dew point margin indicates how far the lens is from condensation (fogging) with a safe temperature margin. and Choose a fixed value based on actual needs.

[0085] To distinguish between lens surface stains (stationary) and road background (moving with the vehicle), this embodiment uses temporal median filtering to extract a background reference and calculates the difference between the current frame and the background reference. Based on the magnitude of this difference, stable spot features reflecting lens dirt are generated.

[0086] First, calculate the time median plot. Next, calculate the magnitude of the difference. Finally, based on the magnitude of the difference... Generate speckle confidence The larger the value, the more likely there is a fixed stain or water stain.

[0087] .

[0088] .

[0089] In the formula For a moment Median over time in pixels The value of . This is a median operation (taking the median of a sequence). for Pixels after time normalization value Represents the normalized pixel The value of . The difference range.

[0090] Specifically, regarding the magnitude of the difference Thresholding, connected component and morphological filtering, and cross-time stability statistics are used to obtain the speckle confidence. .based on Confidence of generated spots The higher the value, the more likely it is that there are abnormal areas in the image that do not move over time, which is very likely lens dirt. The larger the stain, the more it resembles a "fixed stain / stable water stain," and the more it should be treated using the "air blowing" method.

[0091] To avoid noise caused by high ISO sensitivity at night being misjudged as dirt, the noise intensity after ISO compensation was first calculated. Used for noise intensity calculations to avoid mistaking nighttime noise for "clear details".

[0092] .

[0093] .

[0094] In the formula For a moment Noise intensity estimation. Region of Interest Calculate the variance. Region of interest. This is a brightness-normalized image. Reference image for smoothing and denoising. It is a light filtering operator (any implementation such as mean / Gaussian / two-sided).

[0095] Then ISO compensation is performed (to avoid misjudging high noise at night as degradation).

[0096] .

[0097] In the formula This represents the noise intensity after ISO compensation. This is the operator for finding the maximum value. For a moment Noise intensity estimation. This is the noise-ISO coefficient (obtained through calibration). For a moment Image sensor gain. The image sensor gain is used as a benchmark.

[0098] Then calculate the exposure jump index.

[0099] .

[0100] In the formula To increase the exposure of indicators. This is the mean operator for the region of interest. For a moment Region of interest (also available in implementation) However, consistency must be maintained. for Brightness-normalized image at any given time. This is a brightness-normalized image.

[0101] Finally, a unified image quality score is calculated by combining the above indicators. This serves as the basis for closed-loop acceptance. The unified image quality score is calculated as follows.

[0102]

[0103] In the formula For a moment Image quality rating (the higher the better). To normalize the sharpness. This is for normalized contrast. This is the overexposure interference index. This represents the noise intensity after ISO compensation. for The weighting coefficients. for The weighting coefficients. for The weighting coefficients. for The weighting coefficients.

[0104] Overexposure interference index It is constructed as the percentage of saturated pixels or the percentage of highlight area within the region of interest. Its core function is to identify image quality degradation caused by factors other than lens issues, thus avoiding accidental triggering of cleaning / heating.

[0105] S5. Interference locking is determined based on noise intensity. If not locked, the fogging or dirt conditions are distinguished based on scene flags and multi-dimensional feature indicators, and control commands are generated in conjunction with the vehicle's driving status. Preferably, S5 specifically includes steps S51 to S54.

[0106] S51, Overexposure Interference Index Noise intensity Or exposure jump When the threshold is exceeded, it is determined to be an interference state, and the system enters an interference lock state, suppressing cleaning and heating actions and only outputting selected frames. If the system is not locked, it distinguishes between fogging and dirt states based on scene flags and multi-dimensional feature indicators, and generates control commands in conjunction with the vehicle's driving status.

[0107] .

[0108] In the formula The higher the confidence level of the disturbance (the more likely it is to be an "external disturbance / system fluctuation"). This is the overexposure interference index. The glare threshold. This represents the noise intensity after ISO compensation. This is the noise threshold. To increase the exposure of indicators. This is the exposure jump threshold. for The interference weighting coefficient. for The interference weighting coefficient. for The interference weighting coefficient. This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise.

[0109] like It enters a locked state, suppressing invalid actions (no heating / no blowing or reduced triggering), primarily performing frame selection output and prompts to avoid erroneous actions related to "non-lens issues." Interference threshold.

[0110] S52, if tunnel prediction signs Set to perform anti-fog action, maintaining the module temperature above the dew point temperature by at least one dew point margin.

[0111] S53, Based on confidence level of daytime fog Confidence level of nighttime fog and dew point margin Determine if fog has formed. If fog is detected, execute the defogging action.

[0112] Confidence level for daytime fog: .

[0113] Confidence level for nighttime fogging: .

[0114] In the formula Confidence level for daytime fog. Confidence level for nighttime fogging. The fogging weight is used to determine the fogging confidence level. The fogging weight for dew point margin. This is the daytime fog index. This is the nighttime fog index. This is the threshold for daytime fog index. The threshold for nighttime fogging. This refers to the dew point margin. This is the dew point margin threshold (it is generally easier to fog if the dew point is below this threshold). This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise.

[0115] S54, Based on dirt and smudge confidence Determine if the vehicle is dirty. If the dirt determination is successful and the vehicle speed is... If the speed exceeds the minimum effective vehicle speed threshold, an air blowing cleaning action will be performed.

[0116] The soiling confidence level is calculated as follows.

[0117] .

[0118] .

[0119] In the formula For the level of confidence in dirtiness. To normalize the sharpness. This is the threshold for sharpness degradation (below this value, it looks more like dirt / occlusion). For the confidence level of the spots. This is the overexposure interference index. This represents the noise intensity after ISO compensation. for The dirt weight coefficient. for The dirt weight coefficient. for The dirt weight coefficient. for The dirt weight coefficient. This is a speed limit control sign. The speed is the vehicle speed. This is the minimum effective vehicle speed threshold. This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise.

[0120] Specifically, speed gating prevents air blowing from being triggered at low speeds or when stationary.

[0121] S6. Execute the corresponding heating defogging or air blowing cleaning actions according to the control instructions. After the actions are executed, re-acquire images, calculate the image quality score and improvement amount, and decide whether to stop the action, adjust the execution dosage, or update the reference baseline parameters based on the acceptance results.

[0122] Preferably, the anti-fogging and defogging actions employ a closed-loop proportional and hysteresis control strategy to adjust the heating power. This is to keep the dew point margin stable above the safe threshold.

[0123] .

[0124] In the formula, time is... Heating power (can be duty cycle or gear value). This is the maximum heating power (used for rapid heating). Minimum sustaining power (used to maintain temperature stability). This is the safety dew point margin threshold (e.g., 3°C). The hysteresis bandwidth (e.g., 1°C) is used to prevent frequent switching near the threshold. This refers to the dew point margin.

[0125] Specifically, if Below At maximum power Heating rapidly raises the module temperature. If Entering the hysteresis interval Heating power varies Linear decrease achieves smooth adjustment. If Reaching or exceeding Switch to minimum sustaining power It operates energy-efficiently and maintains its anti-fog effect. When the tunnel prediction markers are cleared... and Stop the anti-fog action when necessary.

[0126] This control law is activated in advance before the tunnel / basement entrance to achieve preventive temperature control, effectively reducing the risk of condensation caused by drastic changes in temperature and humidity, and forming a gradient temperature control strategy with the subsequent "defogging action".

[0127] The defogging process employs a closed-loop acceptance test, and the image quality improvement meets the requirements. and Or reach the maximum heating time Stop heating when the time is right.

[0128] .

[0129] In the formula To improve image quality. For a moment Image quality rating. To activate the strong defogging time Image quality rating. The threshold for achieving the required image quality. This refers to the dew point margin. This is the safety dew point margin threshold.

[0130] During air cleaning, to handle stubborn water droplets or semi-dry mud, the air cleaning action employs a variable frequency sweep pulse and performs closed-loop dosage adjustment. Preferably, the air pulse interval... With the number of times Decreasing.

[0131] .

[0132] In the formula For the first Short blow intervals. The minimum interval. This is the initial interval. Index for the number of short blows. This represents the interval step size.

[0133] When the confidence level of the spots Decrease to the spot threshold Below, or at any time Image quality rating Restored to the image quality threshold Stop now. Otherwise, increase the dosage and try again.

[0134] Specifically, the process doesn't end immediately after the action is performed; instead, it enters the acceptance phase. Taking blowing air as an example, the improvement in image quality score before and after blowing air is compared. If the improvement amount is not met, the next air delivery dose will be automatically increased. The operation is as follows.

[0135] First, wait for the preset time after blowing air. Then, frame images are acquired for acceptance and frame selection.

[0136] Then calculate the first... The amount of improvement required for the first air blowing test.

[0137]

[0138] In the formula For the first The amount of image quality improvement per breath. Rate the image quality after blowing air. The image quality score before blowing air. Index of the number of blowing attempts.

[0139] If it fails and Then increase the dosage. This represents the maximum number of breaths.

[0140]

[0141] In the formula For the first One-time inhalation dose. This is the maximum blowing dose. For the first One-time inhalation dose. These are the parameter tuning coefficients. This represents the minimum improvement threshold.

[0142] If it happens multiple times consecutively (e.g., 3 times or more) Cleaning Ineffective: Reports a "Camera Obstruction / Cleaning Failed" error, prompting manual cleaning or switching to an alternative view, and stops the operation to protect the hardware.

[0143] The electronic rearview mirror control method of this invention achieves intelligent identification and precise control of lens contamination and fogging by integrating multi-dimensional image feature analysis and vehicle driving status data. This method can dynamically adjust the detection strategy based on lighting conditions, noise intensity, and scene prediction, effectively distinguishing between lens dirt, fogging, and transient interference, and avoiding false triggering caused by sudden changes in ambient light or sensor noise.

[0144] The control method combines physical dew point margin calculation with closed-loop verification of image quality scoring. It can proactively activate anti-fogging in scenarios prone to condensation, such as tunnels and underground parking garages, and evaluate the improvement effect in real time after cleaning or defogging, adaptively adjusting execution parameters to maintain clear visibility under various operating conditions. Furthermore, by introducing vehicle speed gating, frequency-sweeping pulse blowing, and a reference parameter self-learning mechanism, the effectiveness of cleaning actions and the adaptability and reliability of the control method are further improved, significantly reducing the need for manual intervention and enhancing driving safety.

[0145] The control method for the electronic rearview mirror also includes step S7. S7: Select the clearest frame and output it to ensure continuous usability of the output image. Preferred steps S7 include S71 to S73.

[0146] S71. Select the optimal frame index in the frame acquisition window after the action.

[0147] .

[0148] In the formula The optimal frame index (specifically: the frame number that maximizes the image quality score). The independent variable that makes the expression within the parentheses reach its maximum value. . The frame number within the sampling window. For the first Frame (corresponding time) The image quality rating. For the first The time index corresponding to the frame.

[0149] S72, outputs the clearest frame.

[0150] .

[0151] In the formula The clearest frame output. For a moment The input frame (or the corresponding buffered frame). Optimal index The corresponding moment.

[0152] S73. If all consecutive windows are unavailable, a degradation trigger will be activated.

[0153] .

[0154] In the formula This is the operator for finding the maximum value. This is the set of scores for each frame within the sampling window. This is the lowest available image quality threshold.

[0155] The control method for electronic rearview mirrors also includes a self-learning calibration step S8. When determining a normal, non-foggy, and non-interference-free state, the reference sharpness is dynamically updated. Specifically, the buckets are updated to improve robustness across different scenarios. This self-learning process ensures that the control method can adapt to lens aging or long-term environmental changes, avoiding false alarms caused by reference drift due to equipment aging.

[0156] When the following conditions are met: "no fog, no strong interference, no unstable spots, and stable image quality", the bucket... Update reference clarity.

[0157] .

[0158] In the formula The reference sharpness for scene bucket b. For assignment / update symbols. The learning rate (between 0 and 1). Weights to retain the old values. This represents the current clarity metric.

[0159] Update the reference noise level.

[0160] .

[0161] In the formula For bucket The reference noise intensity. This represents the noise intensity after ISO compensation.

[0162] Through the above implementation methods, this embodiment constructs a complete closed loop with perception, decision-making, execution, and evolution capabilities, effectively solving the pain points of traditional electronic rearview mirror cleaning functions that are "unintelligent, inaccurate, and lacking in results".

[0163] Obviously, the embodiments described above are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0164] In the several embodiments provided in this invention, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus and method embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0165] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0166] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks. It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0167] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0168] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0169] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0170] The use of "first" and "second" in the embodiments is merely to distinguish similar objects and does not represent a specific ordering of objects. It is understood that "first" and "second" can be interchanged in a specific order or sequence where permitted. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0171] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A control method for an electronic rearview mirror with a cleaning function, characterized in that, Include: Acquire image frames captured by the electronic rearview mirror, as well as vehicle driving status data; The current lighting scene is determined based on the lighting conditions and gain parameters, and the entrance to the tunnel or basement is predicted, generating scene markers. Perform region of interest (ROI) identification and brightness normalization on the image frames; Based on the preprocessed data and vehicle driving status data, multidimensional feature indicators are calculated. The metrics include at least: contrast and sharpness reflecting image details, halo index reflecting nighttime scattering, dew point margin reflecting the risk of physical fogging, stable spot characteristics reflecting lens dirt, noise intensity reflecting image quality interference, and uniform image quality score. Interference locking is determined based on noise intensity; if not locked, fogging or dirt conditions are distinguished based on scene flags and multi-dimensional feature indicators, and control commands are generated in combination with vehicle driving status. The corresponding heating and defogging or air cleaning actions are executed according to the control instructions; after the actions are executed, the images are re-acquired to calculate the image quality score and the amount of improvement, and the decision is made on whether to stop the action, adjust the execution dose or update the reference baseline parameters based on the acceptance results.

2. The control method for an electronic rearview mirror with cleaning function according to claim 1, characterized in that, The vehicle driving status data includes: illumination value, image sensor gain, and an indicator that the vehicle is about to enter a tunnel or underground parking garage; When the illumination value is lower than the nighttime illumination threshold or the image sensor gain is higher than the preset gain threshold, the nighttime mode flag is set; in nighttime mode, the halo index will be used as the main criterion for subsequent fogging detection. ； In the formula This is the night mode indicator; This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise. For a moment Illumination value; The nighttime light threshold; For a moment Image sensor gain; This is the gain threshold; For logical "OR"; When the tunnel or underground parking garage is about to be entered, or when the light value changes abruptly within an adjacent cycle, the tunnel prediction flag is set. When the tunnel prediction flag is set, the active anti-fog maintenance mode is entered, and heating is turned on in advance to increase the dew point margin. ； In the formula For predicting tunnels or underground parking garages; A sign indicating that you are about to enter a tunnel or underground parking garage; For a moment Illumination value; This represents the threshold for sudden changes in light intensity.

3. The control method for an electronic rearview mirror with cleaning function according to claim 1, characterized in that, The process for identifying regions of interest is as follows; When the lighting scene is daytime, exclude overexposed areas and areas where the gradient magnitude or texture energy is below the threshold to obtain the region of interest; When the lighting scene is at night, identify pixels whose brightness exceeds the threshold, use them as the center of the point light source, and include the surrounding ring area in the region of interest. The specific steps for brightness normalization are as follows; ； In the formula Represents the normalized pixel The value; Indicates the original input frame in pixels The brightness value; Indicates the region of interest Internal to the original input frame Find the mean; To prevent numbers from being divided by zero.

4. The control method for an electronic rearview mirror with cleaning function according to claim 1, characterized in that, The contrast ratio is calculated as follows; ； ； In the formula For a moment RMS contrast ratio; To normalize the contrast; For a moment Region of interest; The mean brightness within the region of interest; A brightness-normalized image; :bucket Reference contrast; Index the scene bucket; Represents the normalized pixel The value; To prevent numbers from being divided by zero; The resolution index is calculated as follows; ； ； In the formula For a moment Clarity metrics; To normalize sharpness; The variance operator is defined within the region of interest. For the Laplace operator; The reference sharpness for scene bucket b; The nighttime halo index is calculated by extracting a set of point light sources and calculating the halo energy ratio for each point light source. ； ； In the formula Point light source Halo index; For pixel position index; For Center, radius arrive The annular region; For Center, radius The core area; For a moment The overall halo index; Represents a set All Operators for finding the mean; For a moment A set of point light sources; The multidimensional feature index also includes the daytime fog index and the nighttime combined fog index; ； ； In the formula The daytime fog index; The weighting coefficient for the first daytime period; The weighting coefficient for the second daytime; The nighttime fog index; The first night's weighting coefficient; This is the weighting coefficient for the second night.

5. The control method for an electronic rearview mirror with cleaning function according to claim 1, characterized in that, Dew point margin and dew point temperature are calculated based on the Magnus formula. ； ； ； In the formula These are intermediate variables in the Magnus formula; It is the first Magnus constant; It is the second Magnus constant; Ambient temperature; It is the natural logarithm; Relative humidity; This refers to the dew point temperature. For module temperature; For dew point margin; The background reference is extracted using temporal median filtering, and the difference between the current frame and the background reference is calculated; then, stable spot features reflecting lens dirt are generated based on the magnitude of the difference. ； ； In the formula For a moment Median over time in pixels The value; This is for median operations; for Pixels after time normalization value Represents the normalized pixel The value; The difference range.

6. The control method for an electronic rearview mirror with cleaning function according to claim 1, characterized in that, The noise intensity, which reflects image quality interference, is calculated as follows; ； ； In the formula This represents the noise intensity after ISO compensation. For finding the maximum value operator; Region of Interest Calculate the variance. Region of interest; A brightness-normalized image; A reference image for smoothing and denoising; Noise-ISO coefficient; For a moment Image sensor gain; The image sensor gain is used as a reference. To increase the exposure threshold; For the mean operator of the region of interest; For a moment Region of interest; for Brightness-normalized image at any given time; The unified image quality score is calculated as follows; In the formula For a moment Image quality rating; To normalize sharpness; To normalize the contrast; The overexposure interference index; This represents the noise intensity after ISO compensation. for Weighting coefficients; for Weighting coefficients; for Weighting coefficients; for The weighting coefficients.

7. A control method for an electronic rearview mirror with a cleaning function according to any one of claims 1 to 5, characterized in that, Interference locking is determined based on noise intensity; If not locked, the system distinguishes between fogging and dirt conditions based on scene flags and multi-dimensional feature indicators, and generates control commands in conjunction with the vehicle's driving status, specifically including: When overexposure interference index Noise intensity Or exposure jump When the threshold is exceeded, it is determined to be an interference state and enters the interference lock state, only performing frame selection output; If tunnel prediction signs Set the module to perform anti-fog action, maintaining the module temperature at least one dew point margin above the dew point temperature. Based on confidence level of daytime fog Confidence level of nighttime fog and dew point margin Determine if fog has formed; if fog is detected, execute the defogging action. Based on dirt confidence Determine if the vehicle is dirty; if the dirt determination is valid and the vehicle speed is... If the speed exceeds the minimum effective vehicle speed threshold, then an air blowing cleaning action will be performed; Confidence level for daytime fog: ; Confidence level for nighttime fogging: ; In the formula Confidence level for daytime fogging; Confidence level for nighttime fogging; The fogging weight is used to determine the fogging confidence level. Fogging weight for dew point margin; The daytime fog index; The nighttime fog index; The threshold for daytime fog index; The threshold for nighttime fog formation; For dew point margin; This is the dew point margin threshold; This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise. The soiling confidence level is calculated as follows; ； ； In the formula For the level of dirtiness confidence; To normalize sharpness; The resolution degradation threshold; For the confidence level of the spots; The overexposure interference index; This represents the noise intensity after ISO compensation. for The dirt weight coefficient; for The dirt weight coefficient; for The dirt weight coefficient; for The dirt weight coefficient; For vehicle speed control signs; For vehicle speed; This is the minimum effective vehicle speed threshold.

8. The control method for an electronic rearview mirror with cleaning function according to claim 7, characterized in that, The anti-fogging and defogging actions employ a closed-loop proportional and hysteresis control strategy to adjust the heating power. To keep the dew point margin stable above the safety threshold; The defogging process employs a closed-loop acceptance test, and the image quality improvement meets the requirements. and Or reach the maximum heating time Stop heating when the time is right; In the formula To improve image quality; For a moment Image quality rating; To activate the strong defogging time Image quality rating; The threshold for image quality to meet the standard; For dew point margin; This is the safety dew point margin threshold. The air-blowing cleaning action uses a variable frequency sweep pulse and performs closed-loop dose adjustment; Blowing pulse interval With the number of times Decreasing; ； If the image quality improvement after a single breath If the target is not reached and the maximum number of breaths is not exceeded, increase the dosage for the next breath. : ； ； In the formula For the first Short interval between blows; Minimum interval; This is the initial interval; Index for short blow counts; This is the interval step size; For the first One-time air dose; This is the maximum blowing dose; For the first One-time air dose; These are the parameter tuning coefficients; The minimum improvement threshold; For the first The amount of image quality improvement per breath; Rate the image quality after blowing air; The image quality score before blowing air.

9. An electronic rearview mirror with a cleaning function, characterized in that, It includes a camera with a heating function, and a cleaning component configured to blow air toward the lens of the camera; The electronic rearview mirror is adapted to perform a control method for an electronic rearview mirror with a cleaning function as described in any one of claims 1 to 8.

10. A car, characterized in that, The device is equipped with an electronic rearview mirror with a cleaning function as described in claim 9.

Images