Road surface recognition method and device and vehicle

Abstract

The invention discloses a road surface recognition method and device and a vehicle, and relates to the technical field of vehicle control. Comprising the steps that under the condition that a vehicle meets mirage recognition triggering conditions, a feature parameter set used for recognizing mirage is determined; matching each feature parameter contained in the feature parameter set with a corresponding preset mirage feature, so as to determine the number of feature parameters matched with the preset mirage feature through optical feature abnormality and spatial displacement abnormality existing in the road surface; according to the number of the characteristic parameters, determining a confidence coefficient used for reflecting whether mirage exists on the road surface or not; and according to the confidence level corresponding to the confidence coefficient, the vehicle is controlled to execute mirage prompt operation, so that the recognition accuracy of the pavement mirage can be improved, and unnecessary vehicle braking is avoided.

Description

Technical Field

[0001] This disclosure relates to the field of vehicle control technology, and in particular to a road surface recognition method, device and vehicle. Background Technology

[0002] With the rapid development of intelligent driving technology, the reliability and accuracy of vehicle environmental perception systems under complex weather and road conditions have increasingly become a focus of industry attention. Among these, road mirages, or the phenomenon of mirages, have become a significant issue affecting driving decisions because they can interfere with driver vision on hot roads. In hot weather, the earth's surface heats up dramatically due to solar radiation, forming an inversion layer structure where the near-surface air temperature is high and the upper air temperature is low. This causes light refraction due to changes in air density with altitude. When the angle of incidence exceeds a critical angle, total internal reflection occurs, reflecting skylight into the driver's line of sight, creating an optical illusion similar to a water surface. This phenomenon can easily lead to driver misjudgment, causing unnecessary braking or evasive maneuvers, reducing traffic efficiency and potentially increasing the risk of rear-end collisions.

[0003] Traditional methods for identifying road surface mirages mostly rely on algorithms such as visual detection and millimeter-wave radar feature detection. However, because mirages are optically very similar to real water surfaces, the false positive rate is high. Additionally, mirage identification can be performed using static historical meteorological data, but this data is based solely on experience and cannot respond to real-time dynamic changes in the road surface.

[0004] Therefore, how to improve the accuracy of road surface mirage recognition and avoid unnecessary vehicle braking has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of the above problems, this disclosure provides a road surface recognition method, device, and vehicle to overcome or at least partially solve the above problems. The technical solution is as follows: If the vehicle meets the mirage recognition triggering conditions, determine the set of feature parameters used to identify the mirage; Each feature parameter in the feature parameter set is matched with its corresponding preset mirage feature to determine the number of feature parameters that match the preset mirage feature by using optical feature anomalies and spatial displacement anomalies on the road surface. Based on the number of the feature parameters, determine the confidence level used to reflect whether a mirage exists on the road surface; Based on the confidence level corresponding to the confidence level, the vehicle is controlled to perform a mirage prompt operation.

[0006] By determining whether a vehicle meets the mirage recognition trigger conditions and enters a mirage-prone environment, a multi-dimensional feature parameter set is constructed and matched with preset mirage features for verification. Based on the number of matching feature parameters, a probabilistic confidence output is generated, avoiding misjudgments caused by traditional single sensors or visual features. This significantly improves the accuracy of recognizing real water surfaces and mirage optical phenomena in complex environments, preventing unnecessary vehicle braking. The feature parameter set extraction is performed in real time, replacing traditional static historical meteorological data, ensuring that mirage recognition is synchronized with actual road conditions. Furthermore, a hierarchical prompting strategy based on confidence level classification allows drivers to intuitively understand driving risks, avoiding unnecessary interference in low-risk scenarios and providing clear operational guidance in high-risk situations, significantly improving the efficiency and safety of driving decisions.

[0007] Optionally, a set of feature parameters for identifying mirages is determined, specifically including: Collect multi-dimensional road surface sensing data; wherein, the multi-dimensional road surface sensing data includes at least road surface images and spectral data; Determine the characteristic displacement deviation of the water surface region to be identified in multiple road surface images due to dynamic displacement; The characteristic displacement deviation and the spectral data are used as feature parameters for identifying mirages and added to the feature parameter set.

[0008] Spectral data can capture the differences in optical reflection characteristics between mirages and real water surfaces, while dynamic displacement deviation distinguishes the dynamic characteristics of static mirages from those of real water surfaces by analyzing the displacement changes of the area to be identified during the approach of a vehicle. By fusing optical features and dynamic displacement features, multi-parameter cross-validation can effectively reduce the risk of misjudgment caused by optical similarity in traditional single vision detection, thereby significantly reducing the false alarm rate in actual driving scenarios, avoiding unnecessary vehicle braking, and improving driving safety and efficiency.

[0009] Optionally, the characteristic displacement deviation can be the optical flow displacement ratio or the characteristic point displacement deviation. This determines the characteristic displacement deviation of the water surface region to be identified in multiple road surface images due to dynamic displacement, specifically including: The displacement deviation of each feature point is determined based on the standard deviation between the displacement vectors of each feature point. Determine whether the optical flow vectors corresponding to the regions to be identified in multiple road surface images exhibit random motion; If so, based on the vehicle's driving parameters, predict the expected displacement between the vehicle and the water surface area to be identified; The optical flow displacement ratio corresponding to the optical flow vector is determined based on the actual displacement generated by the optical flow vector and the expected displacement.

[0010] The optical flow displacement ratio effectively distinguishes the dynamic characteristics of mirages from real water surfaces by comparing the expected displacement of the vehicle with the actual displacement of the optical flow vector. At the same time, by analyzing the standard deviation of the displacement vector of feature points, the degree of disorder of the movement of the area to be identified is quantified. By identifying mirages through the motion characteristics of optical flow vectors and feature points, the instability of single visual features is overcome, and misjudgments caused by static scenes such as road surface reflection and water surface reflection are greatly reduced. It can distinguish between real wading areas and mirage illusions in real time under high-speed driving conditions, further reducing the risk of misjudgment.

[0011] Optionally, each feature parameter included in the feature parameter set is matched with its corresponding preset mirage feature, specifically including: The characteristic displacement deviation is matched with its corresponding theoretical characteristic displacement value to determine whether the characteristic displacement deviation is greater than a preset multiple of the theoretical characteristic displacement value. The key bands in the spectral data used to identify mirages are determined, and the key bands are matched with the key bands in the water surface spectral data to determine whether the difference between the two is greater than a preset difference.

[0012] Feature displacement deviation matching can accurately identify abnormal displacements caused by the lack of physical support in mirages, while spectral data matching further verifies the consistency of optical features by analyzing the differences in reflection characteristics of key bands. The combination of the two achieves dual verification of dynamic displacement and optical characteristics, effectively reducing the risk of misjudgment caused by the similarity of a single feature.

[0013] Optionally, the key bands are matched with key bands in the water surface spectral data, specifically including: The key bands are matched with key bands in the water surface spectral data to determine the band difference degree corresponding to the key bands; The difference between the spectral data and the water surface spectral data is obtained by weighted summation based on the basic weights corresponding to each key band.

[0014] By matching key bands with key bands in water surface spectral data, it is possible to accurately focus on core bands that are sensitive to spectral differences, avoiding interference from irrelevant bands. At the same time, the band difference is weighted and summed according to the basic weight of each key band. By comprehensively considering the importance of different bands to the spectral characteristics of water surface, different basic weights are assigned to key bands, which can more accurately assess the overall difference between spectral data and water surface spectral data, effectively improving the accuracy of spectral difference assessment.

[0015] Optionally, before weighted summing the band differences according to the band weights corresponding to each key band, the method further includes: Determine the area type and environmental data where the vehicle is located; Based on the region type and / or the environmental data, the basic weights corresponding to the specified key bands in the key bands are adjusted.

[0016] The spectral characteristics of water surfaces vary significantly under different regions and environmental conditions. By adjusting the basic weights corresponding to key bands based on region type and environmental data, the spectral matching can better reflect the actual optical characteristics of the current water surface. This effectively improves the correlation between the key band difference and the actual scene, avoids the matching deviation of fixed weights in complex environments, and ensures more accurate assessment of the key band difference.

[0017] Optionally, before determining the feature parameter set used to identify mirages based on the collected multi-dimensional road surface perception data, the method further includes: In response to a trigger signal sent by the user, determine that the vehicle meets the mirage recognition trigger conditions; Alternatively, collect environmental data and data on the behavior of vehicles ahead; Based on the environmental data and the preceding vehicle's behavior data, determine whether the vehicle meets the mirage recognition trigger conditions.

[0018] Based on a dual-trigger mechanism of user-initiated triggering and automatic environmental perception, users can not only actively send trigger signals based on real-time observation to quickly start the mirage recognition process and avoid risk omissions due to system delays, but also collect environmental data and the behavior data of vehicles ahead to comprehensively analyze and determine the mirage recognition trigger conditions, more accurately capture the characteristics of mirage appearance and start the mirage recognition process, which is more flexible than a single trigger mode.

[0019] Optionally, based on the environmental data and the preceding vehicle behavior data, it is determined whether the vehicle meets the mirage recognition trigger conditions, specifically including: Based on the environmental data, determine whether the current environment is capable of producing a mirage; If so, based on the preceding vehicle behavior data, determine whether a preset number of preceding vehicles have decelerated. If so, based on the behavior data of the target vehicle closest to the vehicle, and if it is determined that the target vehicle does not splash water, the vehicle is determined to meet the mirage recognition trigger condition.

[0020] By assessing the likelihood of mirage formation using environmental data, invalid scenarios that are unlikely to generate mirages are filtered out, avoiding false triggering of non-mirror scenarios and reducing system resource consumption. Secondly, by combining the analysis of the deceleration behavior of the preceding vehicle group, the consistency of group behavior is used to enhance the credibility of the mirage's existence. By excluding the splashing behavior of the nearest target vehicle and eliminating interference from the real water surface, it is ensured that the triggering condition is only for optical mirages and not actual water bodies, avoiding the risk of misjudgment based on a single environmental parameter. Furthermore, the behavior of the preceding vehicle improves the rationality of the triggering decision, providing a more reliable triggering basis for subsequent feature parameter analysis.

[0021] A road surface recognition device, applied to a vehicle control system, the device comprising: The recognition module is used to determine the set of feature parameters for recognizing mirages when the vehicle meets the mirage recognition trigger conditions; The matching module is used to match each feature parameter contained in the feature parameter set with its corresponding preset mirage feature, so as to determine the number of feature parameters that match the preset mirage feature by means of optical feature anomalies and spatial displacement anomalies existing on the road surface. The confidence level determination module is used to determine the confidence level reflecting whether a mirage exists on the road surface based on the number of the feature parameters. The control module is used to control the vehicle to perform a mirage prompt operation based on the confidence level corresponding to the confidence level.

[0022] A vehicle comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: If the vehicle meets the mirage recognition triggering conditions, determine the set of feature parameters used to identify the mirage; Each feature parameter in the feature parameter set is matched with its corresponding preset mirage feature to determine the number of feature parameters that match the preset mirage feature by using optical feature anomalies and spatial displacement anomalies on the road surface. Based on the number of the feature parameters, determine the confidence level used to reflect whether a mirage exists on the road surface; Based on the confidence level corresponding to the confidence level, the vehicle is controlled to perform a mirage prompt operation.

[0023] By employing the aforementioned technical solutions, the road surface recognition method, device, and vehicle disclosed herein, when determining that a vehicle has entered an environment prone to mirages, construct a multi-dimensional feature parameter set and match it with preset mirage features for verification. Based on the number of matching feature parameters, a probabilistic confidence output is generated, avoiding misjudgments caused by traditional single sensors or visual features. This significantly improves the recognition accuracy of real water surfaces and mirage optical phenomena in complex environments, preventing unnecessary vehicle braking. The feature parameter set extraction is performed in real time, replacing traditional static historical meteorological data with a real-time collected feature parameter set, ensuring that mirage recognition is synchronized with the actual road surface condition. Furthermore, a hierarchical prompting strategy based on confidence level classification allows drivers to intuitively understand driving risks, avoiding unnecessary interference prompts in low-risk scenarios and providing clear operational guidance in high-risk situations, significantly improving the efficiency and safety of driving decisions.

[0024] The above description is merely an overview of the technical solution disclosed herein. In order to better understand the technical means of this disclosure and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this disclosure more apparent and understandable, specific embodiments of this disclosure are described below. Attached Figure Description

[0025] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this disclosure. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic flowchart of a road surface recognition method provided in an embodiment of this application is shown; Figure 2 This paper shows a schematic diagram of the structure of a road surface recognition device provided in an embodiment of this application; Figure 3 A schematic diagram of the structure of a vehicle provided in an embodiment of this application is shown. Detailed Implementation

[0026] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0027] With the rapid development of intelligent driving technology, the reliability and accuracy of vehicle environmental perception systems under complex weather and road conditions have increasingly become a focus of industry attention. Among these, road mirages, or the phenomenon of mirages, have become a significant issue affecting driving decisions because they can interfere with driver vision on hot roads. In hot weather, the earth's surface heats up dramatically due to solar radiation, forming an inversion layer structure where the near-surface air temperature is high and the upper air temperature is low. This causes light refraction due to changes in air density with altitude. When the angle of incidence exceeds a critical angle, total internal reflection occurs, reflecting skylight into the driver's line of sight, creating an optical illusion similar to a water surface. This phenomenon can easily lead to driver misjudgment, causing unnecessary braking or evasive maneuvers, reducing traffic efficiency and potentially increasing the risk of rear-end collisions.

[0028] Traditional methods for identifying road surface mirages mostly rely on algorithms such as visual detection and millimeter-wave radar feature detection. However, because mirages are optically very similar to real water surfaces, the false positive rate is high. Additionally, mirage identification can be performed using static historical meteorological data, but this data is based solely on experience and cannot respond to real-time dynamic changes in the road surface.

[0029] Therefore, how to improve the accuracy of road surface mirage recognition and avoid unnecessary vehicle braking has become a technical problem that urgently needs to be solved by those skilled in the art.

[0030] Based on the above application scenarios, in order to solve the technical problem of how to improve the accuracy of road surface mirage recognition and avoid unnecessary vehicle braking, this application provides a road surface recognition method, such as... Figure 1 As shown, Figure 1 This is a schematic flowchart illustrating a road surface recognition method provided in an embodiment of this application. The method can be applied to vehicle control systems and includes: S101: If the vehicle meets the mirage recognition triggering conditions, determine the feature parameter set used to identify the mirage.

[0031] A mirage is a virtual image formed by the refraction of light due to differences in air density. A road mirage (also known as a surface mirage) is a virtual image formed by light rays bending downwards. This virtual image is located below the actual object, appearing as a reflection from water or a sinking object. In intelligent driving scenarios, road mirages are a key area for identification. During vehicle operation, false puddles or deformed road surfaces created by road mirages can easily mislead drivers, causing them to misperceive the situation. For drivers with a habit of braking, encountering a mirage can lead to unnecessary deceleration or braking, resulting in unnecessary energy consumption, reduced road efficiency, and potential rear-end collisions. Considering the impact of road mirages on driving behavior, the following examples primarily focus on identifying road mirages that are more likely to occur in everyday travel scenarios, such as water surfaces. If the vehicle is traveling on non-urban roads, such as high-altitude mountain roads or desert highways with poor road conditions, obstacle mirages also need to be identified to avoid safety risks caused by misoperations such as lane changes. Therefore, road mirages may mislead drivers into thinking there is standing water ahead. Thus, during real-time vehicle operation, accurately identifying areas on the road that appear to be water, determining whether they are real water or mirages caused by high-temperature refraction, and providing drivers with certain prompts can effectively prevent drivers from making unnecessary braking decisions due to misidentification of water surfaces.

[0032] However, mirages only form under specific environmental conditions. Generally, mirages occur under conditions of high temperature and strong sunlight. Only when these environmental conditions work together can abnormal refraction or total reflection of light occur, ultimately forming a false road surface mirage. Mirage recognition requires the vehicle control system to collect and analyze data from multiple sensors. If real-time mirage recognition of the road surface is initiated when it is impossible for a mirage to form, it will cause the system to consume a large amount of computing resources, increasing system power consumption. Therefore, the prerequisite for mirage recognition of the road surface is to determine that the vehicle meets the mirage recognition trigger conditions. The mirage recognition trigger conditions refer to the conditions used to determine whether the mirage recognition process needs to be initiated. In this embodiment, two triggering mechanisms are provided: user-initiated triggering and vehicle-initiated triggering. The vehicle control system will only initiate the subsequent mirage recognition process if it detects that the vehicle meets the condition. Only then will it call the vehicle's sensors to collect the corresponding data and perform subsequent analysis. This way, the recognition process is only initiated in scenarios where mirages are frequent, which can concentrate computing resources on high-risk periods, thereby reducing the overall power consumption and computing power of the system and adapting to the resource-constrained characteristics of in-vehicle embedded devices.

[0033] After initiating the mirage recognition process, multi-modal sensors mounted on the vehicle need to collect multi-dimensional road surface perception data in real time to determine if a mirage appears in front of the vehicle. If the distance between the vehicle and the suspected water surface area is far, the vehicle does not need to brake or collect data. Only when the vehicle gradually approaches the suspected water surface area will the driver potentially brake due to misidentification of water. At this time, multi-dimensional road surface perception data needs to be collected for mirage recognition. During data collection, multiple data points are collected continuously according to the vehicle's spatial position. For example, when the vehicle is 10 meters away from the suspected water surface area, multi-dimensional road surface perception data is collected, with a specified displacement interval of 0.5 meters. Then, multi-dimensional road surface perception data is collected every time the distance between the vehicle and the suspected water surface area decreases by 0.5 meters.

[0034] The multimodal sensor includes at least an onboard microwave dielectric sensor mounted on the front bumper, a polarized binocular camera mounted on the top of the windshield, a road surface temperature measurement module mounted on the chassis, a lidar, and a hyperspectral camera. Taking one type of onboard microwave dielectric sensor as an example, it has a beam angle of 80° and a detection distance of 0-150m, and can calculate the equivalent dielectric constant of the road surface through the phase difference of the reflected signal. The polarized binocular camera can acquire images of the road surface ahead. The road surface temperature measurement module can use a non-contact infrared thermometer to obtain the road surface temperature. The lidar is used to emit orthogonally polarized light to calculate the polarization angle dispersion. The hyperspectral camera is used to acquire continuous spectral data of the road surface ahead.

[0035] The multi-dimensional road surface perception data collected by the aforementioned sensors is used to construct a feature parameter set for identifying mirages. This application provides a mirage feature library storing road surface mirage features. By comparing the real-time collected feature parameter set with the mirage feature library, it can be determined whether a mirage has appeared on the road surface. The construction of the feature parameter set requires further analysis of the multi-dimensional road surface perception data.

[0036] Because the physical nature of mirages leads to kinematic differences in their optical behavior compared to real objects, the movement patterns of areas resembling water surfaces can be detected based on road surface images. The dynamic characteristic reflecting these movement patterns is the characteristic displacement deviation in the road surface image. Therefore, it is necessary to obtain the characteristic displacement deviation from the acquired road surface images, which becomes one of the characteristic parameters for subsequent mirage identification. Real water bodies and dry asphalt or concrete pavements have drastically different spectral reflectance characteristics in specific infrared bands. Although mirages visually resemble water, their physical material is still asphalt or similar substances. These materials do not have strong absorption characteristics in the same infrared band, and their reflectance is much higher than that of water. Therefore, the absorption rate in different bands of the spectral data can also be used as a characteristic parameter for mirage identification. In addition, the equivalent dielectric constant, polarization angle dispersion, and refractive index gradient of the road surface, further calculated from data collected by vehicle-mounted microwave dielectric sensors and lidar, can also be used to identify road surface mirages.

[0037] The polarization angle dispersion is obtained through data analysis from a polarization binocular camera and a lidar. Specifically, a laser beam with a specific polarization state is emitted from the polarization lidar towards the suspected water surface area, and the polarization binocular camera is simultaneously triggered to collect the reflected optical signals. For each detection point (i.e., pixel), based on the light intensity information of the orthogonal polarization components received by the lidar, combined with the multi-angle polarization images captured by the polarization camera, a set of Stokes parameters [S0, S1, S2, S3] is jointly calculated. Here, S0 represents the total light intensity, S1 represents the light intensity difference between the horizontal and vertical polarization components, S2 represents the light intensity difference between the +45° and -45° polarization components, and S3 represents the light intensity difference between right-handed and left-handed circularly polarized light. Using the inverted Stokes parameters, the polarization angle of each detection point is calculated. The formula for calculating the polarization angle is: The polarization angle is used to describe the vibration direction of the electric field vector of a light wave. Statistical analysis was performed on the polarization angles calculated from all detection points within the suspected water surface area over multiple consecutive time frames, and the standard deviation of the polarization angles was calculated. The optical properties of a real water surface are relatively stable, and its polarization angle distribution is concentrated with a small standard deviation. However, in mirage areas, due to drastic and irregular fluctuations in air density, the polarization angles exhibit strong randomness and irregular fluctuations, resulting in highly dispersed values ​​in both time series and spatial distribution. Consequently, the calculated standard deviation will be significantly larger.

[0038] The refractive index gradient is calculated using road surface temperatures collected by a non-contact infrared thermometer. Specifically, the surface temperature of the road surface at different distances in front of the vehicle is measured using a non-contact infrared thermometer, and the low-altitude temperature distribution close to the road surface is estimated accordingly. Since the electromagnetic waves emitted by the lidar are affected by atmospheric water vapor absorption during propagation, the approximate humidity information along the path can be deduced by analyzing the attenuation characteristics of the echo signal. Substituting the obtained temperature and humidity data into the Edlen formula, i.e. Where A and B are constants, T is temperature, e is humidity, P is atmospheric pressure (provided by vehicle-mounted sensors), and n is the air refractive index. The refractive index gradient is obtained by differentiating the calculated refractive index value by altitude, dn / dz. The core condition for mirage phenomena is the presence of a strong temperature inversion layer, which causes the air refractive index near the road surface to decrease sharply with increasing altitude, resulting in a huge refractive index gradient. Under normal atmospheric conditions, this gradient is very small and tends to be zero.

[0039] Therefore, the final feature parameter set mainly includes the following features: characteristic displacement deviation, spectral data, equivalent dielectric constant of the road surface, polarization angle dispersion, and refractive index gradient. It should be noted that the above features are only possible examples; other parameters involved in identifying mirages and real wading areas in practical applications can also be added to the feature parameter set.

[0040] S102: Match each feature parameter in the feature parameter set with its corresponding preset mirage feature, so as to determine the number of feature parameters that match the preset mirage feature by means of optical feature anomalies and spatial displacement anomalies on the road surface.

[0041] Traditional mirage recognition methods rely on static historical meteorological data or single features for identification. However, single features are susceptible to environmental noise interference. For example, in visual detection methods, strong light, shadows, and rain reflections can lead to misjudgments of texture anomalies. When using millimeter-wave radar, the reflected waves from real puddles may be confused with the unobstructed features of a mirage. To adapt to mirage recognition in complex scenarios, this application collects multimodal road perception data and constructs a multidimensional feature parameter set, where each feature parameter can be used for mirage recognition. To improve the reliability of the recognition results, each feature parameter in the feature parameter set needs to be matched with its corresponding preset mirage features. The mirage feature library stores typical mirage features corresponding to different feature parameters. These mirage features describe the theoretical values ​​of the optical and spatial geometric parameters that a mirage phenomenon should exhibit under specific environmental conditions.

[0042] After matching each feature parameter sequentially, it can be determined whether the feature parameter matches the preset mirage features. If they do not match, it indicates that the current indicators have identified optical feature anomalies or spatial displacement anomalies on the road surface ahead. Optical feature anomalies reflect the difference in optical appearance between the road surface area to be identified and real water bodies, while spatial displacement anomalies refer to whether the overall movement between the vehicle and the area to be identified follows spatial displacement laws due to continuous vehicle movement. Therefore, based on the optical feature anomalies and spatial displacement anomalies present on the road surface, it is possible to determine which feature parameters do not match the preset mirage features, and thus determine the number of matching feature parameters.

[0043] S103: Determine the confidence level used to reflect whether a mirage exists on the road surface based on the number of feature parameters.

[0044] If multiple feature parameters match their corresponding preset mirage features, the reliability of the mirage recognition result is higher. Therefore, based on the number of feature parameters that match the preset mirage features, the confidence level reflecting the presence of a mirage on the road surface can be determined. Confidence level is an indicator used to quantify the credibility of a mirage on the road surface; the higher the confidence level, the more accurate the recognition result of a mirage on the road surface. The correspondence between confidence level and the number of feature parameters is preset. This application embodiment lists 5 feature parameters for judging road surface mirages. When the number of matched feature parameters is greater than or equal to 4, the confidence level is greater than 95%; when the number of feature parameters is 3, the confidence level is in the range of [75%, 95%]; when the number of matched feature parameters is less than 3, the confidence level is less than 70%. The specific confidence score is determined based on the importance of the feature parameters at the current moment. The importance of different feature parameters varies depending on the type of area the vehicle is driving in. If, within the current area, a feature parameter that matches the preset mirage feature has high importance, then the confidence score will ultimately fall at the higher value within the aforementioned confidence score range. It should be noted that the feature parameter set is updated in real time. When feature parameters are added or deleted, if the total number of feature parameters in the feature parameter set changes, then the criteria for judging the confidence score also need to be updated accordingly.

[0045] S104: Control the vehicle to perform mirage prompting operations based on the confidence level corresponding to the confidence level.

[0046] If a mirage is detected on the road ahead, the vehicle needs to provide a mirage warning, such as a voice prompt or vibration alert, to help the driver understand that the wading area is a mirage rather than real water, thus preventing the driver from instinctively braking due to visual misperception of water. However, traditional mirage recognition methods only support two states: triggered or not triggered. This means the system's final feedback to the user is either a warning or no warning, which can prevent the driver from accurately assessing the road conditions ahead, potentially leading to overreaction or complete disregard for the warning. Furthermore, vehicle control systems have limited computing resources; failing to tier the warning strategies would result in a significant waste of computational resources.

[0047] Therefore, in order to improve user experience, optimize the utilization of computing resources, and reduce system power consumption, this embodiment of the application adopts a graded prompting strategy based on confidence level. The control system can control the vehicle to perform the mirage prompting operation corresponding to the confidence level calculated above. If the confidence level is a high confidence level, that is, a confidence level greater than 95%, it means that the current mirage is highly credible and a prompt should be triggered immediately to avoid unnecessary braking by the driver or the autonomous driving system due to misjudgment. The mirage prompting operation performed at this time is a high-intensity prompting operation, including but not limited to the head-up display (HUD) projecting semi-transparent rainbow stripes, displaying the text "Optical phenomenon ahead", and voice broadcasting "Optical phenomenon ahead". The brightness of the HUD can be adaptively adjusted according to the current vehicle speed. The faster the vehicle speed, the higher the brightness. In this way, in the scenario of high-speed driving, the stronger brightness prompt can ensure that the driver can understand the road conditions ahead in a timely manner. If the confidence level is medium (between 75% and 95%), it indicates a high probability of a mirage. However, the driver's experience is still needed to confirm the existence of a mirage. In this case, a simple warning is sufficient. The mirage warning would involve a slight vibration of the steering wheel at a frequency of 5Hz and an amplitude of 0.3mm. Conversely, if the confidence level is low (less than 70%), the likelihood of a mirage is low, and the control system will not trigger a warning to avoid frequent false alarms that could distract the driver and reduce system reliability.

[0048] In one embodiment, before identifying a mirage, it is necessary to determine whether the vehicle meets the mirage recognition trigger conditions. Two triggering mechanisms are provided for mirage recognition: user-initiated triggering and vehicle-initiated passive triggering. Based on these different mechanisms, different methods exist for determining whether the vehicle meets the mirage recognition trigger conditions. It should be noted that, regardless of the triggering mechanism, before determining whether the mirage recognition trigger conditions are met, the distance between the current vehicle and the area to be identified (i.e., the suspected water surface area) must be detected. If the distance is far, the driver does not need to react to this area. Mirage recognition is only required when the vehicle is close to the area to be identified, that is, when the distance to the area is no greater than a preset distance threshold. Only under these circumstances will the control system trigger the adaptation process between the vehicle and the mirage recognition trigger conditions.

[0049] In the user-initiated triggering mechanism, users have the ability to actively initiate the mirage recognition process. If a user manually triggers the mirage recognition function, the system will respond to the user's trigger signal, directly determining that the vehicle meets the mirage recognition trigger conditions. At this point, the system has the authority for mirage recognition and executes the subsequent data collection and mirage recognition process. User-selectable triggering methods include voice, gesture, and app. For example, users can directly trigger the mirage recognition function via voice command, press or click a dedicated button or touchscreen option on the vehicle's center console or steering wheel, trigger mirage recognition by recognizing user gestures through the vehicle's camera or sensors, or remotely send trigger signals via a mobile app or vehicle networking service.

[0050] In the vehicle passive triggering mechanism, the mirage recognition process relies on the vehicle's detection results of the surrounding road and environment. In this case, it is only necessary to automatically determine whether the vehicle meets the mirage recognition triggering conditions based on the collected environmental data and the behavior data of the vehicle in front, without user intervention. Environmental data refers to the environmental data collected in real time by onboard sensors, such as temperature, humidity, and light sensors. Based on the environmental data, it can be determined whether the current environment will produce a mirage, i.e., whether it is high temperature, non-rainy weather, etc. For example, when the surface temperature is greater than 50℃ and the vertical temperature gradient ΔT / Δh > 3℃ / m, it is determined that a mirage will occur. Environmental data can only characterize the potential environmental conditions for the mirage phenomenon and cannot directly prove whether the current vehicle will actually be affected by the mirage. The behavior data of the vehicle in front represents the reaction behavior of the vehicle in front to the road conditions. If the behavior data of the vehicle in front can indirectly reflect that a mirage may appear on the road ahead, the subsequent mirage recognition process can be directly triggered to ensure driving safety and driving efficiency. By combining environmental and vehicle behavior data, the confidence and reliability of mirage recognition trigger conditions are effectively improved.

[0051] In one embodiment, since mirages only occur under specific conditions, determining whether mirage recognition of the road surface is necessary depends first on whether the current environment meets the necessary conditions for mirage formation. If the current environment has no possibility of mirage formation, then impossible scenarios can be filtered out, ensuring that the control system only performs subsequent analysis in scenarios where mirages are possible, thus reducing system power consumption. If the environmental data indicates low temperature, no light or weak light, precipitation, or strong winds, mirage formation is not possible under these conditions, and mirage recognition of the road surface is unnecessary. If the current environment meets the above environmental data requirements, it indicates that mirage formation is possible on the road surface. Whether to initiate the mirage recognition process requires further judgment based on the behavior of surrounding vehicles. Here, the behavior of surrounding vehicles mainly refers to the vehicles in front of the vehicle. If a vehicle changes lanes, then the behavior data of the vehicle in front in the lane where the vehicle is located needs to be re-acquired.

[0052] If a mirage exists on the road ahead, the driver may mistake it for real water. Drivers accustomed to braking in water will often subconsciously slow down. Therefore, the deceleration of the vehicle in front can indirectly reflect the presence of a road mirage. However, deceleration is actually influenced by various factors. Besides braking in water, sudden deceleration of the vehicle in front or poor driving habits can also cause the vehicle in front to slow down. Therefore, judging the existence of a road mirage solely based on the deceleration of a single vehicle is highly random and uncertain. Therefore, to improve the accuracy of identifying the behavior of vehicles in front, the data should not be limited to the closest vehicle. Instead, it needs to utilize vehicle-to-vehicle (V2V) communication technology to obtain the behavior data of multiple vehicles in front of the vehicle, all of which must be in the same lane as the current vehicle. In the V2V communication framework, the behavior data of the preceding vehicle is no longer just the data perceived by the following vehicle through its own sensors. It also includes state information reflecting the true motion state, such as speed, acceleration, and heading angle, which is directly measured and actively broadcast by the preceding vehicle through its own onboard sensors. After receiving the message from the preceding vehicle, the following vehicle decodes it and sends it to the autonomous driving domain controller via Controller Area Network (CAN) or Ethernet for decision-making, to determine whether the preceding vehicle has decelerated.

[0053] When data on the behavior of vehicles ahead indicates that a predetermined number of vehicles have slowed down, it suggests that multiple vehicles have decelerated. This indicates a common source of interference, causing multiple drivers to make similar evasive responses. This interference source is likely a road mirage, which makes it difficult for drivers to accurately assess road conditions, leading them to brake instinctively. Assuming a predetermined number of vehicles slowing down collectively increases the confidence in the existence of a mirage, significantly reducing the probability of the system falsely triggering mirage recognition due to the random behavior of individual vehicles. The predetermined number is pre-set and can be flexibly adjusted based on actual road conditions. A larger predetermined number is used when there are many vehicles to improve the accuracy of road condition recognition in congested traffic. It should be noted that identifying the presence of a predetermined number of vehicles slowing down can be done by identifying consecutive vehicles ahead or by identifying scattered vehicles within the recognition area. If the target is dispersed vehicles, it is no longer enough to determine whether there is a group deceleration behavior by simply counting the number of vehicles. Instead, it is necessary to analyze the behavior data of the vehicles in front within the identification area. If the proportion of the vehicles in front that decelerates to the total number of vehicles in the identification area meets a preset ratio, it is also considered that a preset number of vehicles in front have decelerated.

[0054] The characteristics of water splashing on sensors are highly similar to those of a mirage. Multiple vehicles ahead will collectively slow down due to the water splashed by other vehicles, which obstructs their vision. Therefore, after identifying a collective slowdown by vehicles ahead, it is necessary to further analyze the actual driving conditions to determine whether the slowdown is caused by the presence of a wading area. The data on the vehicles ahead also includes data detected by the vehicles' own sensors. By analyzing this data, it is possible to determine whether the nearest target vehicle is splashing water. If splashing water is detected, it indicates that the collective slowdown is most likely due to the presence of a real water area, thus ruling out the mirage hypothesis. Only when the target vehicle is not splashing water does it suggest that the current slowdown may be due to misidentification of the water surface, in which case it is necessary to trigger the subsequent mirage identification process.

[0055] In one embodiment, mirage identification is based on a set of feature parameters, which are determined by multi-dimensional road surface sensing data acquired in real time by a multi-modal sensor. The multi-dimensional road surface sensing data includes at least road surface images and spectral data. The road surface images are used to analyze the characteristic displacement deviations caused by dynamic displacement in the water surface area to be identified, while the spectral data is used to analyze band absorption rates.

[0056] Feature displacement deviation refers to the unstable motion vectors exhibited by the same area to be identified in consecutive multi-frame road surface images due to the dynamic distortion of the optical path caused by a mirage. The motion pattern and amplitude of this vectors significantly deviate from the expected rigid displacement caused by the vehicle itself. Since the motion characteristics of mirages and real water surfaces are fundamentally different, feature displacement deviation in road surface images can distinguish the dynamic characteristics of static mirages from those of real water surfaces. Furthermore, the spectrum reflects the differences in optical reflection characteristics between mirages and real water surfaces. Therefore, feature displacement deviation and spectral data can be added to the feature parameter set as feature parameters for mirage identification, and used for subsequent mirage recognition. In addition to feature displacement deviation and spectral data, the feature parameter set also includes the road surface equivalent dielectric constant, polarization angle dispersion, and refractive index gradient. Since the calculation process for the road surface equivalent dielectric constant, polarization angle dispersion, and refractive index gradient has already been described, the determination process of feature displacement deviation will be explained later.

[0057] In one embodiment, both the real water surface and the mirage undergo dynamic displacement due to the real-time movement of the vehicle, but their physical nature is completely different. Therefore, the real water surface and the road surface mirage can be distinguished by characteristic motion patterns. This application embodiment achieves mirage identification through two types of characteristic displacement deviations: feature points and optical flow vectors.

[0058] Feature points, also known as key points, refer to pixels in a road surface image that possess prominent and easily identifiable attributes. Several feature points are selected within the area of ​​water surface to be identified, and the movement of each feature point is tracked in consecutive frames of road surface images. By analyzing the movement patterns of the feature points, it is possible to determine whether the area represents real water or a mirage. Because water movement is affected by factors such as wind, gravity, and the water's own inertia, changes in ripples caused by wind can lead to irregular movement of individual feature points. The direction and magnitude of the movement are random. A mirage, on the other hand, is caused by the road surface heating the air, forming a group of air lenses with uneven density. It is a refraction phenomenon that occurs when light propagates in a non-uniform medium, rather than actual physical movement. Even with wind, it exhibits more regular feature point displacements. Therefore, by detecting whether irregular movement occurs in feature points within the area to be identified, it is possible to distinguish between mirages and real water surfaces at a microscopic level.

[0059] Irregular motion reflects the dispersion of the motion directions of all feature points within the area to be identified. Therefore, the displacement deviation of feature points can be determined by the standard deviation between the displacement vectors of each feature point. A large displacement deviation indicates inconsistent motion of the feature points, suggesting that the feature points in the area are undergoing irregular motion. A small displacement deviation indicates that the motion directions of the feature points are generally consistent, suggesting that the area is experiencing stable and regular motion. The displacement deviation reflects the degree of disorder in the motion of feature points at the microscopic level, and can be used to distinguish between road surfaces and mirages.

[0060] In addition to identifying mirages through the motion patterns of feature points, it is also possible to combine the overall motion between the vehicle and the area to be identified with whether it follows kinematic laws, i.e., to determine whether the motion of the entire area to be identified conforms to the expectations determined by the vehicle's motion, thus identifying mirages on the road surface. In this embodiment, the overall motion pattern of the area is evaluated by the displacement of the optical flow vector. The optical flow vector represents the direction and distance of motion of a pixel in the road surface image from the previous frame to the current frame, and can be calculated using the Lucas-Kanade optical flow method. The consistency between the actual motion observed in the road surface image and the theoretical motion predicted based on physical laws is evaluated using the optical flow vector. In this embodiment, the consistency is quantified by the optical flow displacement ratio. Therefore, by calculating the optical flow displacement ratio corresponding to the optical flow vector in the area to be identified, it can be determined whether the area to be identified is a mirage or a real water surface.

[0061] Specifically, since the image features of a real water surface undergo regular displacement as the vehicle moves, while a mirage is a virtual image created by the refraction of light and may not move regularly with the vehicle, it is possible to preliminarily determine whether the area to be identified is a mirage by determining whether the optical flow vector corresponding to the area to be identified in multiple road surface images exhibits random movement. If the displacement direction and magnitude of the optical flow vector are inconsistent, it indicates that the area to be identified may be a road surface mirage.

[0062] However, real water surfaces under strong winds can also exhibit highly chaotic ripples, leading to a high degree of randomness in optical flow and being misinterpreted as mirages. While ripples on the water surface can cause random and irregular movement of the optical flow vector at individual feature points, the water body itself is a fixed entity on the road surface. As the vehicle moves forward, the relative position changes between the water body and the vehicle strictly adhere to kinematic principles. Therefore, despite the interference caused by surface ripples, the displacement reflected by the average, macroscopic optical flow vector over the entire area will highly match the expected displacement predicted by the vehicle's kinematic model. Conversely, mirages are virtual images, not physical entities. Their movement is not caused by the displacement of objects in three-dimensional space, but by the distortion and jittering effects produced by light passing through a layer of hot air with constantly changing density. The movement exhibited by this optical illusion is completely undetermined by vehicle motion; its optical flow vector is a product of optical distortion, and the final calculated optical flow displacement ratio will significantly deviate from the theoretically expected value.

[0063] Therefore, the vehicle's driving parameters are obtained, including vehicle speed, acceleration, steering angle, and camera mounting position, angle, and focal length. Using an inverse perspective transformation model, the displacement vector that a pixel fixed on the ground should produce in the image within one frame interval is calculated. This displacement vector is the predicted expected displacement between the vehicle and the water surface area to be identified. Then, the average value of each optical flow vector within the area to be identified is calculated to obtain the actual displacement generated by the optical flow vector within that area. Both the expected and actual displacements are taken as modulo values, and the ratio between the actual and expected displacements is calculated to obtain the optical flow displacement ratio corresponding to the optical flow vector.

[0064] In one embodiment, the above content describes the calculation process of each feature parameter. Subsequently, each feature parameter in the feature parameter set needs to be matched with its corresponding preset mirage feature in turn. If a match can be made, it means that the region to be identified can be determined to be a mirage through the feature parameter.

[0065] Each feature parameter corresponds to a different preset mirage feature. The extremely high dielectric constant of real water surface will significantly change the equivalent dielectric properties of the road surface, while the dielectric constant of the dry road surface below the mirage remains unchanged and is no different from the surrounding normal road surface. Therefore, by matching the equivalent dielectric constant of the road surface with the numerical range corresponding to the mirage feature, it can be determined whether the area to be identified is a mirage.

[0066] Real water surfaces exhibit near-mirror reflection, producing reflected light with a highly consistent polarization angle, resulting in low dispersion. Mirages, however, are created by the refraction of light due to hot air currents, causing the polarization direction of the reflected light to become completely random and disordered, leading to extremely high dispersion of the polarization angle. Therefore, the polarization angle dispersion needs to be matched against a pre-defined polarization angle dispersion threshold; if it exceeds this threshold, it is identified as a mirage.

[0067] The core condition for mirages is the presence of a strong temperature inversion layer. This causes the refractive index of the air near the road surface to decrease sharply with increasing altitude, resulting in a huge refractive index gradient. Therefore, the refractive index gradient needs to be matched with a refractive index gradient threshold. If the refractive index gradient is greater than the threshold, it is identified as a mirage.

[0068] Matching characteristic displacement deviations focuses on the differences in dynamic motion patterns. Real water surfaces, due to their actual existence, have optical flow displacement ratios close to the theoretical values. However, mirages, being optical illusions, exhibit actual optical flow vector displacements that significantly deviate from the theoretical values ​​due to the randomness of light refraction. By matching characteristic displacement deviations with their corresponding theoretical values, it can be determined whether the characteristic displacement deviation exceeds a preset multiple of the theoretical value, thus effectively identifying mirages. If the optical flow displacement ratio is close to the theoretical value and the actual displacement is close to the expected displacement, it indicates that the overall movement of the area to be identified is basically synchronized with the vehicle, ultimately identifying it as a real water surface. Conversely, if the optical flow displacement ratio significantly deviates from the preset multiple of the theoretical value, it indicates that the actual displacement randomly deviates from the expected displacement, and the overall movement of the area to be identified is not synchronized with the vehicle, in which case it is identified as a mirage.

[0069] The matching of spectral data utilizes the difference in optical reflection characteristics between mirages and real water surfaces. In order to accurately pinpoint the few bands that best characterize the unique absorption characteristics of water bodies from the full spectrum data, so as to maximize computational efficiency and recognition reliability, it is first necessary to determine the key bands in the spectral data used to identify mirages. The key bands are [0.48, 0.55, 0.65, 0.72, 0.85, 0.97, 1.20, 1.45, 1.65, 1.94, 2.05, 2.22, 3.40, 3.68, 3.96, 4.24, 4.52, 4.80, 5.08, 5.36, 8.50, 9.70, 10.6, 11.3], of which [0.48, 0.55, 0.65, 0.72, 0.85] represent the visible and near-infrared bands, [0.97, 1.20, 1.45, 1.94] represents the strong water vapor absorption band, [1.65, 2.05, 2.22] represents the shortwave infrared band, and [3.40, 3.68, 3.96, ... [4.24, 4.52, 4.80, 5.08, 5.36] represents the mid-wave infrared band, and [8.50, 9.70, 10.6, 11.3] represents the thermal infrared band. Because real water surfaces exhibit very strong absorption valleys in the near-infrared and short-wave infrared bands, their spectral curves show a significant downward trend. In contrast, the absorption characteristics of asphalt, cement, and other road materials in mirage surfaces differ drastically from those of water in these bands, resulting in relatively flat spectral curves. Therefore, the reflection intensity or wavelength distribution of mirages and real water surfaces in certain key bands will differ significantly from those of real water surfaces. By matching key bands with key bands in the water surface spectral data, it is possible to determine whether there are differences between the two, thus further distinguishing between mirages and real water surfaces.

[0070] In one example, the preset mirage features and real water surface features corresponding to the feature parameters are shown in Table 1: Table 1

[0071] As shown in Table 1, it is determined whether the feature parameter set matches the preset mirage features in the table. If they match, it indicates that the feature parameter identifies the area to be identified as a road surface mirage. It should be noted that the preset mirage features corresponding to the spectral data in Table 1 are features of key bands. However, in the actual judgment process, this single judgment standard is no longer used alone. It is also necessary to combine whether the difference between the spectral data and the spectral data of the real water surface exceeds a preset threshold to distinguish between mirages and real water surfaces.

[0072] In one embodiment, not all bands are equally important in distinguishing between real water surfaces and mirages. Therefore, when calculating the difference between the spectral features of real water surfaces and mirages, key bands need to be assigned different basic weights to highlight the band information with the most diagnostic value. First, the key bands are matched with the key bands in the water surface spectral data, and the reflectance of each key band is obtained. The difference in reflectance between each key band is then calculated, which is the band difference degree of the current key band. The difference quantifies the degree of dissimilarity between the current area to be identified and the real water surface in the key band. The greater the band difference degree, the less like water it is in that key band. The band difference degrees are weighted and summed according to the basic weights corresponding to each key band to obtain the difference degree between the spectral data and the water surface spectral data. If the difference degree is greater than the difference degree threshold, the current area is determined to be mismatched with the real water surface spectrum, i.e., a mirage. Otherwise, it is the real water surface. For example, in the 1.65μm band, the reflectivity of dry asphalt pavement is very high, assumed to be 40%, while the standard reflectivity of water is 5%, resulting in a difference of 35%. In the 0.65μm band, asphalt may also be dark-colored, with an assumed reflectivity of 11%, very close to water's 10%, resulting in a difference of 1%. Assuming a base weight of 0.9 for the 1.65μm band and 0.1 for the 0.65μm band, the final calculated difference is 31.6%. By assigning different base weights to key bands, the importance of certain key bands is amplified in different scenarios, reducing the impact of accidental noise and interference in non-sensitive bands on the overall judgment and improving the accuracy of the difference.

[0073] It should be noted that the importance of key bands varies depending on the environment and region. Therefore, to improve the accuracy of spectral identification, it is necessary to dynamically adjust the basic weights of different key bands so that spectral matching can adapt to interference under different geographical locations and weather conditions, thereby maintaining high identification accuracy and preventing misjudgments. First, the region type and environmental data of the vehicle's location are determined. The region type is determined based on the vehicle's current geographical location, including desert highways, coastal cities, and high-altitude mountain roads. Environmental data affecting key bands mainly include temperature and humidity. Then, based on the region type and / or environmental data, the basic weights corresponding to the designated key bands affected by the interference are adjusted. The adjustment logic is to increase the basic weights of key bands whose data is most reliable and effective in the current environment, and decrease the basic weights of key bands with poor data quality and high interference. Tables 2 and 3 show examples of adjustment strategies corresponding to environmental data and region type, respectively, where H represents humidity and T represents temperature. Table 2

[0074] Table 3

[0075] The data in Tables 1 and 2 are for illustrative purposes only. The specific adjustment values ​​can be flexibly set according to actual testing needs, and this application does not impose any restrictions on them.

[0076] In addition, if Figure 2 As shown, Figure 2 This is a schematic diagram of a road surface recognition device provided in an embodiment of this application. The device includes: The recognition module 201 is used to determine the feature parameter set for recognizing the mirage when the vehicle meets the mirage recognition triggering conditions; The matching module 202 is used to match each feature parameter contained in the feature parameter set with its corresponding preset mirage feature, so as to determine the number of feature parameters that match the preset mirage feature by means of optical feature anomalies and spatial displacement anomalies existing on the road surface. The confidence level determination module 203 is used to determine the confidence level reflecting whether a mirage exists on the road surface based on the number of feature parameters. The control module 204 is used to control the vehicle to perform mirage prompting operations according to the confidence level corresponding to the confidence level.

[0077] Optionally, the identification module is specifically used to collect multi-dimensional road surface sensing data; wherein, the multi-dimensional road surface sensing data includes at least road surface images and spectral data; Determine the characteristic displacement deviation of the water surface region to be identified in multiple road surface images due to dynamic displacement; The characteristic displacement deviation and the spectral data are used as feature parameters for identifying mirages, and Add to the feature parameter set.

[0078] Optionally, the feature displacement deviation includes the optical flow displacement ratio and the feature point displacement deviation amount. The recognition module is specifically used to determine whether the feature points and optical flow vectors corresponding to the area to be recognized in multiple road images produce random motion. If so, based on the vehicle's driving parameters, predict the expected displacement between the vehicle and the water surface area to be identified; Based on the actual displacement and the expected displacement generated by the optical flow vector, determine the optical flow displacement ratio corresponding to the optical flow vector; The displacement deviation of each feature point is determined based on the standard deviation between the displacement vectors of each feature point.

[0079] Optionally, the confidence determination module is specifically used to match the feature displacement deviation with its corresponding theoretical value of feature displacement in order to determine whether the feature displacement deviation is greater than a preset multiple of the theoretical value of feature displacement. The key bands in the spectral data used to identify mirages are determined, and the key bands are matched with the key bands in the water surface spectral data to determine whether the difference between the two is greater than a preset difference.

[0080] Optionally, the confidence determination module is specifically used to match the key band with the key band in the water surface spectral data to determine the band difference degree corresponding to the key band. The difference between the spectral data and the water surface spectral data is obtained by weighted summation based on the basic weights corresponding to each key band.

[0081] Optionally, a confidence determination module is used to determine the area type and environmental data of the vehicle. Based on the region type and / or the environmental data, the basic weights corresponding to the specified key bands in the key bands are adjusted.

[0082] Optionally, before determining the feature parameter set for mirage recognition based on the collected multi-dimensional road surface perception data, specifically in response to the trigger signal sent by the user, the vehicle is used to determine whether it meets the mirage recognition trigger condition. Alternatively, collect environmental data and data on the behavior of vehicles ahead; Based on the environmental data and the preceding vehicle's behavior data, determine whether the vehicle meets the mirage recognition trigger conditions.

[0083] Optionally, it is specifically used to determine whether the current environment can generate a mirage based on the environmental data; If so, based on the preceding vehicle behavior data, determine whether a preset number of preceding vehicles have decelerated. If so, based on the behavior data of the target vehicle closest to the vehicle, and if it is determined that the target vehicle does not splash water, the vehicle is determined to meet the mirage recognition trigger condition.

[0084] Regarding the apparatus in the above embodiments, the specific manner in which each unit performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0085] Figure 3 This is a schematic diagram of the structure of a vehicle provided in an embodiment of this application.

[0086] For example, such as Figure 3 As shown, the vehicle includes a memory 301 and a processor 302. The memory 301 stores executable program code 3011, and the processor 302 is used to call and execute the executable program code 3011 to perform the road surface recognition method.

[0087] This embodiment can divide the vehicle into functional modules according to the above method example. For example, each function can be assigned to a separate module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0088] When each functional module is divided according to its corresponding function, the vehicle may include: The recognition module is used to determine the set of feature parameters for recognizing mirages when the vehicle meets the mirage recognition trigger conditions; The matching module is used to match each feature parameter contained in the feature parameter set with its corresponding preset mirage feature, so as to determine the number of feature parameters that match the preset mirage feature by means of optical feature anomalies and spatial displacement anomalies on the road surface. The confidence level determination module is used to determine the confidence level reflecting whether a mirage exists on the road surface based on the number of feature parameters. The control module is used to control the vehicle to perform mirage prompting operations based on the confidence level corresponding to the confidence level.

[0089] It should be noted that all relevant content of each step involved in the above method embodiments can be referenced from the functional description of the corresponding functional module, and will not be repeated here.

[0090] The vehicle provided in this embodiment is used to perform the above-described road mirage recognition method, and therefore can achieve the same effect as the above implementation method.

[0091] When using integrated units, the vehicle may include a processing module and a storage module. The processing module is used to control and manage the vehicle's actions. The storage module supports the vehicle in executing program code and data.

[0092] The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits as disclosed in this application. The processor may also be a combination of computing functions, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

[0093] This embodiment also provides a computer-readable storage medium (including but not limited to disk storage, CD-ROM, optical storage, etc.) storing computer program code. When the computer program code is run on a computer, the computer executes the above-mentioned related method steps to realize a vehicle remote control method provided in the above embodiment.

[0094] This embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to implement a vehicle remote control method provided in the above embodiment.

[0095] The beneficial effects of the above embodiments can be referred to the beneficial effects of the corresponding methods provided above, and will not be repeated here.

[0096] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0097] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0098] In the description of this disclosure, it should be understood that if the terms "upper", "lower", "front", "rear", "left" and "right" are used to indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the position or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this disclosure.

[0099] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. It should also be noted that 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 process, method, article, or apparatus. Unless otherwise specified, 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 the element.

[0100] The above are merely embodiments of this disclosure and are not intended to limit the scope of this disclosure. Various modifications and variations can be made to this disclosure by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of the claims of this disclosure.

Claims

1. A road surface recognition method, characterized in that, Applied to a vehicle control system, the method includes: If the vehicle meets the mirage recognition triggering conditions, determine the set of feature parameters used to identify the mirage; Each feature parameter in the feature parameter set is matched with its corresponding preset mirage feature to determine the number of feature parameters that match the preset mirage feature by using optical feature anomalies and spatial displacement anomalies on the road surface. Based on the number of the feature parameters, determine the confidence level used to reflect whether a mirage exists on the road surface; Based on the confidence level corresponding to the confidence level, the vehicle is controlled to perform a mirage prompt operation.

2. The road surface recognition method according to claim 1, characterized in that, Determine the set of feature parameters used to identify mirages, specifically including: Collect multi-dimensional road surface sensing data; wherein, the multi-dimensional road surface sensing data includes at least road surface images and spectral data; Determine the characteristic displacement deviation of the water surface region to be identified in multiple road surface images due to dynamic displacement; The characteristic displacement deviation and the spectral data are used as feature parameters for identifying mirages and added to the feature parameter set.

3. The road surface recognition method according to claim 2, characterized in that, The characteristic displacement deviation is the optical flow displacement ratio or the displacement deviation of the feature point. Determining the characteristic displacement deviation of the water surface area to be identified in multiple road surface images due to dynamic displacement specifically includes: The displacement deviation of each feature point is determined based on the standard deviation between the displacement vectors of each feature point. Alternatively, determine whether the optical flow vectors corresponding to the regions to be identified in multiple road surface images exhibit random motion; If so, based on the vehicle's driving parameters, predict the expected displacement between the vehicle and the water surface area to be identified; The optical flow displacement ratio corresponding to the optical flow vector is determined based on the actual displacement generated by the optical flow vector and the expected displacement.

4. The road surface recognition method according to claim 2, characterized in that, Each feature parameter in the feature parameter set is matched with its corresponding preset mirage feature, specifically including: The characteristic displacement deviation is matched with its corresponding theoretical characteristic displacement value to determine whether the characteristic displacement deviation is greater than a preset multiple of the theoretical characteristic displacement value. The key bands in the spectral data used to identify mirages are determined, and the key bands are matched with the key bands in the water surface spectral data to determine whether the difference between the two is greater than a preset difference.

5. The road surface recognition method according to claim 4, characterized in that, Matching the key bands with key bands in the water surface spectral data specifically includes: The key bands are matched with key bands in the water surface spectral data to determine the band difference degree corresponding to the key bands; The difference between the spectral data and the water surface spectral data is obtained by weighted summation based on the basic weights corresponding to each key band.

6. The road surface recognition method according to claim 5, characterized in that, Before weighted summation of the band differences according to the band weights corresponding to each key band, the method further includes: Determine the area type and environmental data where the vehicle is located; Based on the region type and / or the environmental data, the basic weights corresponding to the specified key bands in the key bands are adjusted.

7. The road surface recognition method according to claim 1, characterized in that, Before determining the feature parameter set used to identify mirages based on the collected multi-dimensional road surface perception data, the method further includes: In response to a trigger signal sent by the user, determine that the vehicle meets the mirage recognition trigger conditions; Alternatively, collect environmental data and data on the behavior of vehicles ahead; Based on the environmental data and the preceding vehicle's behavior data, determine whether the vehicle meets the mirage recognition trigger conditions.

8. The road surface recognition method according to claim 7, characterized in that, Based on the environmental data and the preceding vehicle's behavior data, determine whether the vehicle meets the mirage recognition trigger conditions, specifically including: Based on the environmental data, determine whether the current environment is capable of producing a mirage; If so, based on the preceding vehicle behavior data, determine whether a preset number of preceding vehicles have decelerated. If so, based on the behavior data of the target vehicle closest to the vehicle, and if it is determined that the target vehicle does not splash water, the vehicle is determined to meet the mirage recognition trigger condition.

9. A road surface recognition device, characterized in that, The device, used in a vehicle control system, includes: The recognition module is used to determine the set of feature parameters for recognizing mirages when the vehicle meets the mirage recognition trigger conditions; The matching module is used to match each feature parameter contained in the feature parameter set with its corresponding preset mirage feature, so as to determine the number of feature parameters that match the preset mirage feature by means of optical feature anomalies and spatial displacement anomalies existing on the road surface. The confidence level determination module is used to determine the confidence level reflecting whether a mirage exists on the road surface based on the number of the feature parameters. The control module is used to control the vehicle to perform a mirage prompt operation based on the confidence level corresponding to the confidence level.

10. A vehicle, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to: perform the road surface recognition method according to any one of claims 1-8.

Images