Dynamic calibration method, device and electronic equipment for driver face posture angle

Abstract

The application provides a dynamic calibration method and device for a driver's face posture angle and electronic equipment, which is applied to the field of computer vision and includes the following steps: collecting a face image of a driver, extracting a face feature and a current face posture angle; obtaining a target observation offset angle from an observation offset angle management module based on the face feature, and the calculation process of the target observation offset angle includes the following steps: continuously collecting multiple face posture angle samples of the driver when detecting that the vehicle is in a straight driving state; identifying a main angle interval of the multiple face posture angle samples based on statistical quantity in angle value distribution; determining the target observation offset angle according to the main angle interval; and subtracting the target observation offset angle from the current face posture angle to obtain a calibrated face posture angle. The method can obtain a corresponding observation offset angle according to a face feature, correct the current face posture angle, and thus improve the accuracy of judging whether a driver is in a dangerous driving state.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of computer vision, and particularly relates to a dynamic calibration method, device and electronic equipment for a driver face posture angle. BACKGROUND

[0002] A driver monitoring system (DMS) is a system for monitoring the state of a driver. The driver monitoring system analyzes the head, face and eye movements of the driver through a camera and the like, so as to determine in real time whether the driver has dangerous behaviors such as fatigue and distraction, and thus prevent the driver from driving dangerously.

[0003] In order to avoid the driver monitoring system from blocking the line of sight of the driver and causing interference to the driver, the camera of the driver monitoring system is generally not installed in front of the driver, but is installed on a center console or an instrument panel in the vehicle. Therefore, the face posture of the driver presents a certain angle, and even when the driver is in a normal driving posture, the face of the driver collected by the camera of the driver monitoring system also has a posture offset. In the related art, the installation angle of the camera is manually corrected to solve the offset angle of the face of the driver. However, since different drivers have different heights and driving habits, different drivers have different face offset angles. Therefore, the same correction method cannot adapt to different drivers. SUMMARY

[0004] The present application provides a dynamic calibration method, device and electronic equipment for a driver face posture angle. The method can collect the face features of the driver to obtain an observation offset angle, and correct the current face posture angle of the driver according to the observation offset angle, so as to solve the problem that the face angle of the driver is inaccurate and the accuracy of determining that the driver is in a dangerous driving state is low.

[0005] In a first aspect, a dynamic calibration method for a driver face posture angle is provided. The method comprises:

[0006] collecting a face image of a driver, and extracting face features and a current face posture angle of the driver from the face image;

[0007] obtaining a target observation offset angle corresponding to the face features from an observation offset angle management module, the observation offset angle management module being configured to store face features and target observation offset angles corresponding to the face features of different drivers; the target observation offset angle is calculated by: continuously collecting a plurality of face posture angle samples of the driver when detecting that the vehicle is in a straight driving state; identifying a main angle interval of the plurality of face posture angle samples based on a statistical number in terms of angle value distribution; and determining the target observation offset angle according to the main angle interval;

[0008] Subtracting the target observation offset angle from the current face posture angle, a calibrated face posture angle is obtained;

[0009] The identification of the main angle interval of the plurality of face posture angle samples in the angle value distribution based on the statistical quantity comprises:

[0010] The preset angle total interval is divided into a plurality of continuous angle subintervals;

[0011] The first quantity of the plurality of face posture angle samples falling into each angle subinterval is counted, and the sample quantity proportion of each angle subinterval is calculated;

[0012] The angle subinterval with the sample quantity proportion greater than the preset proportion, or the continuous plurality of angle subintervals with the sum of the sample quantity proportions greater than the preset proportion, is determined as the main angle interval.

[0013] Through the above method, the face image of the driver is collected, and the face feature and the current face posture angle of the driver are extracted from the face image. In actual application, due to individual differences of the driver, the current face posture angle of the driver collected by the driver monitoring system is different for drivers with different heights and different sitting habits. If the current face posture angle is directly used as a reference, the driver who is normally driving will be misjudged as being in dangerous driving. In this case, based on the face feature, the corresponding target observation offset angle is obtained from the observation offset angle management module, the current face posture angle is subtracted from the target observation offset angle, and a calibrated face posture angle is obtained. That is, under a large sample quantity, the accuracy of calculating the sample quantity proportion of each angle subinterval is higher, so in the case where the second quantity is greater than the first threshold, the angle subinterval with the sample quantity proportion greater than the preset proportion, or the continuous plurality of angle subintervals with the sum of the sample quantity proportions greater than the preset proportion, is determined as the main angle interval, so that the confidence of the main angle interval is higher, the target observation offset angle is calculated through the main angle interval, the current face posture angle is corrected through the target observation offset angle, and a calibrated face posture angle is obtained, thereby improving the accuracy of the driver monitoring system in real-time detecting dangerous behaviors such as fatigue and distraction of the driver.

[0014] In combination with the first aspect, in some possible implementation manners, the identification of the main angle interval of the plurality of face posture angle samples in the angle value distribution based on the statistical quantity comprises:

[0015] The preset angle total interval is divided into a plurality of continuous angle subintervals;

[0016] The first quantity of the plurality of face posture angle samples falling into each angle subinterval is counted, and the sample quantity proportion of each angle subinterval is calculated;

[0017] The angle sub-interval with the sample quantity ratio greater than the preset ratio, or the continuous multiple angle sub-intervals with the sum of sample quantity ratios greater than the preset ratio, is determined as the main angle interval.

[0018] In combination with the first aspect, in some possible implementation manners, the target observation offset angle is determined according to the main angle interval, including:

[0019] A first average value of all face posture angle samples in the main angle interval is calculated, and the first average value is determined as the target observation offset angle.

[0020] Through the above method, the average level of multiple face posture angles in the angle value distribution is reflected by the first average value, the influence of extreme values is reduced, and the target observation offset angle is more reliable.

[0021] In combination with the first aspect, in some possible implementation manners, before the sample quantity ratio of each angle sub-interval is calculated, the method further includes:

[0022] A second quantity of the multiple face posture angle samples falling into the angle total interval is counted.

[0023] In a case where the second quantity is greater than a first threshold, the step of calculating the sample quantity ratio of each angle sub-interval is performed.

[0024] Through the above method, when the sample quantity is small, the calculated main angle interval is very unstable, and therefore, based on the statistical principle, the accuracy of calculating the sample quantity ratio of each angle sub-interval is high when the sample quantity is large. Therefore, in the case where the second quantity is greater than the first threshold, the angle sub-interval with the sample quantity ratio greater than the preset ratio, or the continuous multiple angle sub-intervals with the sum of sample quantity ratios greater than the preset ratio, is determined as the main angle interval, so that the confidence of the main angle interval is higher.

[0025] In combination with the first aspect, in some possible implementation manners, the target observation offset angle corresponding to the face feature is acquired from the observation offset angle management module, including:

[0026] In a case where the face feature matches any candidate face feature in the observation offset angle management module, it is determined whether there is a target observation offset angle in the observation offset angle management module;

[0027] In a case where there is no target observation offset angle in the observation offset angle management module, it is determined whether there is a historical offset angle in the observation offset angle management module based on the face feature;

[0028] In a case where there is a historical offset angle in the observation offset angle management module, the historical offset angle is determined as the target observation offset angle.

[0029] In the case where the face feature matches any candidate face feature in the observation offset angle management module and there is no target observation offset angle in the observation offset angle management module, the target observation offset angle is determined according to the historical offset angle, so that the current face posture angle is corrected by the historical offset angle to improve the accuracy of the current face posture angle.

[0030] In combination with the first aspect, in some possible implementation manners, the method further includes:

[0031] In the case where there is no historical offset angle in the observation offset angle management module, a second average value of all interval angles in the main angle interval is calculated;

[0032] The second average value is determined as the target observation offset angle.

[0033] In the case where there is no historical offset angle in the observation offset angle management module, the second average value of all interval angles in the main angle interval is determined as the target observation offset angle, so that the current face posture angle is calibrated by all interval angles in the main angle interval, and the accuracy of the current face posture angle is higher.

[0034] In combination with the first aspect, in some possible implementation manners, before the plurality of face posture angle samples of the driver are continuously collected in the case where it is detected that the vehicle is in a straight driving state, the method further includes:

[0035] The speed of the vehicle and the steering wheel angle are obtained;

[0036] In the case where the speed of the vehicle is greater than a preset speed and the steering wheel angle is in a preset steering wheel angle interval, it is determined that the vehicle is in a straight driving state.

[0037] In the above method, whether the vehicle is in a straight driving state is determined based on the speed of the vehicle and the steering wheel angle, and the accuracy is high.

[0038] In combination with the first aspect, in some possible implementation manners, after the current face posture angle is subtracted by the target observation offset angle to obtain a calibrated face posture angle, the method further includes:

[0039] Whether the driver is in dangerous driving is determined based on the calibrated face posture angle;

[0040] In the case where the driver is in dangerous driving, a dangerous driving alarm information is triggered.

[0041] By the above method, whether the driver is in dangerous driving is determined based on the calibrated face posture angle, so that the accuracy of the judgment result is higher, and the problem of distraction alarm false alarm caused by inaccurate driving face posture angle is avoided.

[0042] In a second aspect, a device for dynamic calibration of a driver's face posture angle is provided, and the device comprises:

[0043] A collection and extraction module is configured to collect a face image of a driver and extract a face feature and a current face posture angle of the driver from the face image.

[0044] An acquisition module is configured to acquire a corresponding target observation offset angle from an observation offset angle management module based on the face feature, and the observation offset angle management module is configured to store face features of different drivers and corresponding target observation offset angles. The calculation process of the target observation offset angle comprises: continuously collecting a plurality of face posture angle samples of the driver when it is detected that the vehicle is in a straight driving state; identifying a main angle interval of the plurality of face posture angle samples based on a statistical number in the angle value distribution; and determining the target observation offset angle according to the main angle interval.

[0045] A calculation module is configured to subtract the target observation offset angle from the current face posture angle to obtain a calibrated face posture angle; divide a preset angle total interval into a plurality of continuous angle subintervals; count a first number of the plurality of face posture angle samples falling into each angle subinterval, and calculate a sample number proportion of each angle subinterval; determine the angle subinterval with a sample number proportion greater than a preset proportion, or a plurality of continuous angle subintervals with a sum of sample number proportions greater than the preset proportion, as the main angle interval.

[0046] In a third aspect, an electronic device is provided, comprising a memory and a processor. The memory is configured to store executable program code, and the processor is configured to call and run the executable program code from the memory, so that the electronic device executes the method performed by the above-mentioned dynamic calibration method of the driver's face posture angle.

[0047] In a fourth aspect, a computer program product is provided, and the computer program product comprises computer program code. When the computer program code runs on a computer, the computer executes the method performed by the above-mentioned dynamic calibration method of the driver's face posture angle.

[0048] In a fifth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores computer program code. When the computer program code runs on a computer, the computer executes the method performed by the above-mentioned dynamic calibration method of the driver's face posture angle. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 is a schematic diagram of an implementation environment of a dynamic calibration method for a driver face posture angle provided by an embodiment of the present application.

[0050] Figure 2 is a schematic flowchart of a dynamic calibration method for a driver face posture angle provided by an embodiment of the present application.

[0051] Figure 3 is a display schematic diagram of a calibrated face posture angle provided by an embodiment of the present application.

[0052] Figure 4 is a schematic flowchart of a method for calculating a target observation offset angle provided by an embodiment of the present application.

[0053] Figure 5 is a schematic flowchart of another dynamic calibration method for a driver face posture angle provided by an embodiment of the present application.

[0054] Figure 6 is a structural schematic diagram of a dynamic calibration device for a driver face posture angle provided by an embodiment of the present application.

[0055] Figure 7 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0056] The technical solutions in the present application will be described in detail below with reference to the drawings. In the description of the embodiments of the present application, unless otherwise specified, " / " represents the meaning of or, for example, A / B can represent A or B: "and / or" in the text only describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases of A alone, A and B together, and B alone. In addition, in the description of the embodiments of the present application, "multiple" means two or more than two.

[0057] Hereinafter, the terms "first" and "second" are only for descriptive purposes, and cannot be understood as implying or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more features.

[0058] The following embodiments of the present application are introduced in the process of Figure 1 , the implementation environment of the embodiments of the present application is introduced.

[0059] For example, as shown in Figure 1 , the implementation environment includes a driver monitoring system 110 and a microcontroller 120.

[0060] The driver monitoring system 110 is a driving assistance system of a vehicle. The driver monitoring system 110 collects a face image of a driver through a visual sensor, extracts information from the face image to obtain a face feature, and monitors a state of the driver in real time through the face feature to ensure driving safety. That is, the driver monitoring system 110 judges whether the driver is driving while tired or distracted by capturing a driving picture of the driver in real time.

[0061] In the related art, the visual sensor of the driver monitoring system 110 cannot be installed in front of the driver, which will block the driver's view and cause interference to the driver. In actual application, the visual sensor is installed on a center console or an instrument panel of the vehicle, and presents a certain angle with a face posture of the driver. When the driver maintains normal driving, the face image of the driver collected by the visual sensor is offset from the face of the driver. Before the driver monitoring system 110 judges whether the driver is driving while tired or distracted through the face image, the face image needs to be calibrated to eliminate the influence of the face posture observation offset.

[0062] The microcontroller 120 is configured to acquire the face image of the driver extracted by the driver monitoring system 110, and process the face image of the driver to eliminate the angle offset of the face image, so as to facilitate judging whether the driver is in dangerous driving according to the corrected face image.

[0063] To solve at least one of the above technical problems, the embodiment of the present application provides a dynamic calibration method for a face posture angle of a driver. The method can collect a face image of a driver, extract a face feature and a current face posture angle of the driver from the face image, correct the current face posture angle according to an observation offset angle obtained according to the face feature, and obtain a calibrated face posture angle, so as to solve the problem that the face angle of the driver is inaccurate and the accuracy of judging whether the driver is in dangerous driving is low.

[0064] Figure 2 FIG. 1 is a schematic flowchart of a dynamic calibration method for a face posture angle of a driver provided by an embodiment of the present application.

[0065] As shown in FIG. 2, for example, the dynamic calibration method for the face posture angle of the driver provided by the embodiment of the present application is described taking the microcontroller as an execution subject. The method 200 includes the following steps. Figure 2

[0066] Step 201: Collect a face image of a driver, and extract a face feature and a current face posture angle of the driver from the face image.

[0067] ​The face image refers to original pixel data containing a face. The face image can be a photo, a video frame, or the like visual data. The face feature refers to a feature vector extracted from the face image by a deep learning model, and the face feature can represent the uniqueness of the face. The current face posture angle refers to the three-dimensional space deviation of the face relative to the camera, and the current face posture angle includes a pitch angle, a yaw angle, and a roll angle. The pitch angle is used to represent the nodding action of the driver's head up and down, the yaw angle is used to represent the turning action of the driver's head left and right, and the roll angle is used to represent the tilting action of the driver's head to one side.

[0068] It should be understood that in the process of driving the vehicle, the face image of the driver needs to be collected to determine whether the driver is in fatigue driving.

[0069] In some embodiments, the face image of the driver is collected based on a visual sensor.

[0070] In some embodiments, the face image of the driver is input into a face algorithm module to obtain a face frame and a current face posture angle; the position of the face in the face image is determined based on the face frame; the key feature points (i.e., contour points) of the face image are extracted in the face frame; and the contour points are converted into a feature vector to obtain the face feature of the driver.

[0071] In step 202, the corresponding target observation offset angle is obtained from the observation offset angle management module based on the face feature. The calculation process of the target observation offset angle includes: continuously collecting a plurality of face posture angle samples of the driver when it is detected that the vehicle is in a straight driving state; identifying a main angle interval based on a statistical number on the angle value distribution of the plurality of face posture angle samples; and determining the target observation offset angle according to the main angle interval.

[0072] In actual application, there are individual differences among drivers. For drivers of different heights and different sitting habits, the face angle of the driver collected by the driver monitoring system is different. For example, for a driver with a higher height, the face of the driver has a small pitch angle compared to the driver monitoring system; for a driver with a left or right sitting habit, the face of the driver has a small yaw angle compared to the driver monitoring system. However, the driver monitoring system will use a unified reference line to determine whether the driver is in dangerous driving. If the front of the driver monitoring system is directly used as the reference, the driver in normal driving will be misjudged as in dangerous driving. In this case, the target observation offset angle needs to be obtained.

[0073] The observation offset angle management module is configured to manage the observation offset angle of the driver. In some embodiments, the observation offset angle management module records the facial feature of the driver and the observation offset angle of the driver. The facial posture angle sample is a facial angle sample of the driver collected by the driver monitoring system. The facial posture angle sample includes a pitch angle sample, a yaw angle sample, and a roll angle sample. The main angle interval is an angle range in which the number of facial posture angle samples is most concentrated and the frequency of occurrence is highest. The meaning of the angle value distribution of the plurality of facial posture angle samples is to analyze the overall distribution of the plurality of facial posture angle samples. The target observation offset angle is the deviation angle of the driver relative to the theoretical front of the driver.

[0074] For the completeness and clarity of the scheme, the process of calculating the target observation offset angle will be described in detail in the following embodiments.

[0075] It should be understood that when the vehicle is in a non-straight state, the facial posture angle sample of the driver has a large offset angle relative to the theoretical front of the driver, and therefore the error of calculating the target observation offset angle according to the facial posture angle sample is large. In this case, when it is detected that the vehicle is in a straight state, the plurality of facial posture angle samples of the driver are continuously collected.

[0076] In a possible implementation, it is determined whether the facial feature matches any candidate facial feature in the observation offset angle management module.

[0077] It should be understood that in actual applications, the observation offset angle management module may not store the target observation offset angle of the driver, and therefore it is necessary to determine whether the facial feature of the driver matches any candidate facial feature in the observation offset angle management module to obtain the target observation offset angle of the driver.

[0078] In some embodiments, in the case where the facial feature matches any candidate facial feature in the observation offset angle management module, it is determined whether there is a target observation offset angle in the observation offset angle management module.

[0079] The facial feature matching any candidate facial feature in the observation offset angle management module indicates that the identity information (Identity document, ID) of the driver is stored in the observation offset angle management module, and therefore there may be a target observation offset angle corresponding to the driver in the observation offset angle management module.

[0080] In some embodiments, in the case where there is no target observation offset angle in the observation offset angle management module, it is determined whether there is a historical offset angle in the observation offset angle management module based on the facial feature.

[0081] The historical offset angle is an offset angle calculated by historical facial posture angle samples of the driver. The historical offset angle is used to correct the current facial posture angle of the driver.

[0082] It should be understood that in actual applications, it takes a period of time to calculate the target observation offset angle, and if the current target observation offset angle has not been calculated, it can be determined whether there is a historical offset angle in the observation offset angle management module, so as to correct the current facial posture angle of the driver by the historical offset angle.

[0083] In some embodiments, in the case that there is a historical offset angle in the observation offset angle management module, the historical offset angle is determined as the target observation offset angle.

[0084] In the case that there is a historical offset angle in the observation offset angle management module, it means that the historical offset angle has been calibrated to the historical facial angle of the driver, and thus determining the historical offset angle as the target observation offset angle can calibrate the current facial posture angle of the driver.

[0085] In this implementation, in the case that the facial feature matches any candidate facial feature in the observation offset angle management module and there is no target observation offset angle in the observation offset angle management module, the target observation offset angle is determined according to the historical offset angle, so as to calibrate the current facial posture angle by the historical offset angle, thereby improving the accuracy of the current facial posture angle.

[0086] In some embodiments, in the case that there is no historical offset angle in the observation offset angle management module, a second average value of all interval angles in the main angle interval is calculated; and the second average value is determined as the target observation offset angle.

[0087] In the case that there is no historical offset angle in the observation offset angle management module, it means that the target observation offset angle has not been calculated at the current time and the facial angle of the driver has not been calibrated, and in this case, the target observation offset angle can be determined by the second average value of all interval angles in the main angle interval.

[0088] For example, if the main angle interval is [-5°, 0°), the second average value is -3°, that is, the target observation offset angle is -3°.

[0089] In this implementation, in the case that there is no historical offset angle in the observation offset angle management module, the second average value of all interval angles in the main angle interval is determined as the target observation offset angle, so as to calibrate the current facial posture angle by all interval angles in the main angle interval, thereby making the current facial posture angle more accurate.

[0090] For the completeness and clarity of the scheme description, the process of determining the main angle interval will be described in detail in the following examples.

[0091] Step 203, subtracting the target observation offset angle from the current face posture angle to obtain the calibrated face posture angle.

[0092] Wherein, the face posture angle is used to analyze whether the driver is in a dangerous driving state. The calculation formula of the face posture angle is: calibrated face posture angle = current face posture angle-target observation offset angle.

[0093] It can be understood that the face posture angle includes the pitch angle, the yaw angle and the roll angle, and then the calibrated face posture angle is composed of the calibrated pitch angle, the calibrated yaw angle and the calibrated roll angle.

[0094] As shown in FIG. 8, a display diagram of a calibrated face posture angle is shown. The current face posture angle of the driver is offset to the left compared with the driver monitoring system, and the calibrated face posture angle is parallel to the driver monitoring system. Since the current face posture angle has a deviation compared with the theoretical front of the driver, and the target observation offset angle is the deviation angle, the current face posture angle is subtracted by the target observation offset angle, and the calibrated face posture angle can be obtained. Figure 3 Figure 3

[0095] It can be understood that in actual application, after the above-mentioned step 203 is executed, the following steps can also be executed.

[0096] In a possible implementation, whether the driver is in a dangerous driving is determined based on the calibrated face posture angle.

[0097] In some embodiments, in the case that the driver is in a dangerous driving, a dangerous driving alarm information is triggered.

[0098] Wherein, the dangerous driving alarm information can be used to prompt the driver that the current is in a fatigue driving or a distraction driving. In the embodiments of the present application, the dangerous driving alarm information can include an alarm prompt text and an alarm prompt audio. For example, the alarm prompt text and the alarm prompt audio can be "slow down and stop at the service station immediately".

[0099] In some embodiments, the alarm prompt text and the alarm prompt audio are generated based on the calibrated face posture angle, the alarm prompt text is displayed and the alarm prompt audio is played.

[0100] ​​In some embodiments, the alarm prompt text and the alarm prompt audio are stored in an alarm prompt storage medium. For example, the alarm prompt storage medium includes a USB disk, a mobile hard disk, a magnetic disk, or an optical disk.

[0101] In this implementation, whether the driver is in dangerous driving is determined based on the calibrated facial posture angle, so that the accuracy of the determination result is higher, and the problem of false alarm caused by inattentive driving due to inaccurate driving facial posture angle is avoided.

[0102] The embodiment of the present application provides a dynamic calibration method for a driver's facial posture angle. The method can collect a face image of the driver, and extract a facial feature and a current facial posture angle of the driver from the face image. In actual application, due to individual differences of the driver, the current facial posture angle of the driver collected by the driver monitoring system is different for drivers with different heights and different sitting habits. If the current facial posture angle is directly used as a reference, a driver who is normally driving will be misjudged as being in dangerous driving. In this case, the target observation offset angle is obtained from the observation offset angle management module based on the facial feature, the current facial posture angle is subtracted by the target observation offset angle, and a calibrated facial posture angle is obtained. That is, under a large sample quantity, the accuracy of calculating the sample quantity proportion of each angle subinterval is higher, so in the case that the second quantity is greater than the first threshold, the angle subinterval with a sample quantity proportion greater than a preset proportion, or a plurality of continuous angle subintervals with a sample quantity proportion sum greater than the preset proportion, is determined as a main angle interval, so that the confidence of the main angle interval is higher, the target observation offset angle is calculated through the main angle interval, the current facial posture angle is corrected through the target observation offset angle, and a calibrated facial posture angle is obtained, thereby improving the accuracy of the driver monitoring system in real-time detecting dangerous behaviors such as driver fatigue and inattention of the driver.

[0103] For the completeness and clarity of the scheme, the process of calculating the target observation offset angle will be described in detail taking Figure 4 as an example. As shown in Figure 4 , an exemplary schematic flowchart of a method for calculating a target observation offset angle of the present application is shown, taking a microcontroller as an example of an execution subject. The method 400 includes the following steps.

[0104] Step 401, determining whether the vehicle is in a straight driving state.

[0105] It should be understood that when the vehicle is in the non-straight state, the facial posture angle of the driver will obviously deviate from the driving direction relative to the forward direction of the vehicle. For example, if the vehicle is in a turning state, the face and head of the driver will significantly deviate to the turning direction to observe the trajectory of the curve and the situation of the intersection. In this case, if the driver is determined to be in a dangerous driving state based on the facial posture angle of the driver when the vehicle is in the non-straight state, the normal straight-ahead behavior of the driver will be mistakenly classified as an abnormal posture, thereby reducing the accuracy of determining whether the driver is in a dangerous driving state. To avoid the above-mentioned situation, it is determined whether the vehicle is in a straight state.

[0106] In a possible implementation, the vehicle speed and the steering wheel angle of the vehicle are acquired; and the vehicle is determined to be in a straight state when the vehicle speed is greater than a preset vehicle speed and the steering wheel angle is in a preset steering wheel angle interval.

[0107] The steering wheel angle is used to determine the steering angle of the front wheel of the vehicle. The preset vehicle speed is a vehicle speed automatically determined by the microcontroller, which is not limited in the embodiments of the present application. For example, the preset vehicle speed is 0 km / h. The preset steering wheel angle interval is an angle interval automatically determined by the microcontroller, which is not limited in the embodiments of the present application. For example, the preset steering wheel angle interval is [-5°, 5°].

[0108] In some embodiments, the vehicle speed is acquired based on a vehicle speed sensor, and the steering wheel angle is acquired based on a steering wheel sensor.

[0109] It should be understood that the vehicle speed greater than the preset vehicle speed indicates that the vehicle is in a driving state, and if the steering wheel angle is stable in the range of [-5°, 5°], it indicates that the vehicle keeps straight driving.

[0110] In a possible implementation, the vehicle is determined to be in a non-straight state when the vehicle speed of the vehicle is less than or equal to a preset vehicle speed or the steering wheel angle is not in a preset steering wheel angle interval.

[0111] In this possible implementation, whether the vehicle is in a straight state is determined based on the vehicle speed and the steering wheel angle of the vehicle, and the accuracy is relatively high.

[0112] Optionally, after step 401, the following steps 402-403 are executed according to actual conditions, or the following step 404 is executed.

[0113] Step 402: When the vehicle is in a straight state, a target observation deviation angle is calculated.

[0114] It should be understood that, since the driver monitoring system needs a unified baseline to determine whether the driver is in dangerous driving, different facial posture angles of different drivers need to be calibrated through the unified baseline. In this case, the target observation offset angle is calculated when the vehicle is in a straight driving state.

[0115] The target observation offset angle is used to calibrate the facial posture angle of the driver.

[0116] In a possible implementation, when it is detected that the vehicle is in a straight driving state, a plurality of facial posture angle samples of the driver are continuously collected; a main angle interval of the plurality of facial posture angle samples is identified based on a statistical quantity in terms of angle value distribution; and the target observation offset angle is determined according to the main angle interval.

[0117] The facial posture angle sample is a facial angle sample of the driver collected by the driver monitoring system. The facial posture angle sample includes a pitch angle sample, a yaw angle sample, and a roll angle sample. The main angle interval is an angle range in which the number of facial posture angle samples is most concentrated and the frequency of occurrence is highest. The plurality of facial posture angle samples in terms of angle value distribution means to analyze the overall distribution of the plurality of facial posture angle samples. The target observation offset angle is a deviation angle of the driver compared to the theoretical front of the driver.

[0118] In this implementation, when the vehicle is in a straight driving state, the plurality of facial posture angle samples of the driver are dynamically learned, the angle range with the highest frequency of occurrence is determined based on the plurality of facial posture angle samples, and the target observation offset angle is calculated according to the angle range, that is, the plurality of facial posture angle samples when the vehicle is in a non-straight driving state are filtered out, the ratio of noise data in the plurality of facial posture angle samples is reduced, the proportion of effective data is increased, and the target observation offset angle can be dynamically adjusted according to the main angle interval, which significantly improves the accuracy of the baseline for determining whether the driver is in dangerous driving.

[0119] In order to describe the above implementation in more detail, the following describes the above implementation in several parts.

[0120] The first part describes the content of continuously collecting a plurality of facial posture angle samples of the driver when it is detected that the vehicle is in a straight driving state.

[0121] It should be understood that when the vehicle is in a straight driving state, the driver's face should be oriented in the same direction as the driving direction, and at this time, the plurality of facial posture angle samples of the driver are collected, which can truly reflect the natural state of the driver when he is in a focused driving state.

[0122] In some embodiments, the video acquisition module acquires a face image of the driver, and the face image is preprocessed according to preset image parameters.

[0123] The preset image parameters include an image format and an image resolution.

[0124] In some embodiments, the preprocessed face image is input into a face algorithm module to obtain face features and a facial pose angle sample.

[0125] The face features are used to identify identity features of the driver.

[0126] In the second part, the content of identifying a main angle interval of the plurality of facial pose angle samples based on a statistical number of angle value distributions is described.

[0127] It should be understood that the plurality of facial pose angle samples can not be the same angle value. By identifying the main angle interval, some transient angle deviations are automatically ignored, thereby reducing the interference of angle deviations.

[0128] In one possible implementation, the preset angle total interval is divided into a plurality of continuous angle subintervals, and a second number of the plurality of facial pose angle samples falling into the angle total interval is counted. It is determined whether the second number is greater than a first threshold value.

[0129] The preset angle total interval is an angle interval automatically determined by the microcontroller, and embodiments of the present application do not limit this. For example, the range of the angle total interval can be [-90°, 90°). It can be understood that in actual applications, [-90°, 90°) has covered the maximum deflection angle of the face of the driver.

[0130] The first threshold value is a numerical value automatically determined by the microcontroller, and embodiments of the present application do not limit this. For example, the first threshold value can be 4500.

[0131] In some embodiments, the angle total interval is divided into a plurality of continuous angle subintervals according to a preset interval. The preset interval is an angle interval automatically determined by the microcontroller, and embodiments of the present application do not limit this. For example, the preset interval can be 5°. That is, the angle total interval is divided into 36 angle subintervals. For example, [-90°, 85°), [-85°, 80°)... [-5°, 0°), [0°, 5°)... [80°, 85°), [85°, 90°).

[0132] It can be understood that, since the face posture angle samples include the pitch angle samples, the yaw angle samples and the roll angle samples, the preset angle total interval includes a pitch angle total interval, a yaw angle total interval and a roll angle total interval, and the angle sub-interval includes a pitch angle sub-interval, a yaw angle sub-interval and a roll angle sub-interval.

[0133] It should be understood that, since the angle total interval is the maximum deflection angle of the face of the driver, a plurality of face posture angle samples fall within the range of the angle total interval, in which case, the plurality of face posture angle samples falling into the angle total interval need to be accumulated, that is, the second number of the plurality of face posture angle samples falling into the angle total interval is counted.

[0134] In some embodiments, in the case that the second number is greater than the first threshold, the first number of the plurality of face posture angle samples falling into each angle sub-interval is counted, and the sample number proportion of each angle sub-interval is calculated; the angle sub-interval with the sample number proportion greater than the preset proportion, or the continuous multiple angle sub-intervals with the sum of the sample number proportions greater than the preset proportion, are determined as the main angle interval.

[0135] It can be understood that, counting the plurality of face posture angle samples falling into the angle total interval requires time. For example, it may take 5 minutes to count the second number of the plurality of face posture angle samples falling into the angle total interval to be greater than 4500.

[0136] It should be understood that, when the sample number is small, the calculated main angle interval is very unstable, and therefore, based on the principle of statistics, the accuracy of calculating the sample number proportion of each angle sub-interval is higher when the sample number is large, and therefore, in the case that the second number is greater than the first threshold, the sample number proportion of each angle sub-interval is calculated.

[0137] The preset proportion is a value determined automatically by the microcontroller, which is not limited in the embodiments of the present application. For example, the preset proportion can be 80%.

[0138] The calculation formula of the sample number proportion of each angle sub-interval is: sample number proportion = first number / second number*100%.

[0139] The sample number proportion is used to represent the confidence of the main angle interval; the first number is the number of the plurality of face posture angle samples falling into each angle sub-interval, and the second number is the number of the plurality of face posture angle samples falling into the angle total interval.

[0140] The angle sub-interval with the sample number proportion greater than the preset proportion, or the continuous multiple angle sub-intervals with the sum of the sample number proportions greater than the preset proportion, are determined as the main angle interval, indicating that the angle sub-interval has a higher reliability as the main angle interval.

[0141] In this implementation, when the sample quantity is small, the calculated main angle interval is very unstable, and therefore, based on the statistical principle, the accuracy of calculating the sample quantity proportion of each angle sub-interval is higher when the sample quantity is large, and therefore, in the case where the second quantity is greater than the first threshold value, the angle sub-interval with a sample quantity proportion greater than the preset proportion or the continuous multiple angle sub-intervals with a sample quantity proportion sum greater than the preset proportion are determined as the main angle interval, so that the confidence of the main angle interval is higher.

[0142] In some embodiments, in the case where the second quantity is less than the first threshold value, the multiple face posture angle samples of the driver are continuously collected.

[0143] It should be understood that the second quantity being less than the first threshold value indicates that the face posture angle samples are less, and in order to make the calculated sample quantity proportion more reliable, the multiple face posture angle samples of the driver are continuously collected.

[0144] Thirdly, the content of determining the target observation offset angle according to the main angle interval is described.

[0145] It should be understood that since the average value can offset the error between the highest and lowest face posture angle samples, the average value can represent the general level of the multiple face posture angle samples, and therefore, determining the first average value of all the face posture angle samples as the target observation offset angle can filter out the accidental face offset of the driver.

[0146] In one possible implementation, the first average value of all the face posture angle samples in the main angle interval is calculated, and the first average value is determined as the target observation offset angle.

[0147] In some embodiments, all the face posture angle samples in the main angle interval are accumulated to obtain an accumulated sample sum, and the accumulated sample sum is divided by the number of all the face posture angle samples to obtain the first average value of all the face posture angle samples in the main angle interval.

[0148] In this implementation, the average level of the multiple face posture angles in the angle value distribution is reflected by the first average value, and the influence of extreme values is reduced, so that the target observation offset angle is more reliable.

[0149] In step 403, the target observation offset angle is stored in the observation offset angle management module corresponding to the face feature of the driver.

[0150] It should be understood that in actual applications, since the heights and driving habits of different drivers are different, the target observation offset angles of different drivers are different, and therefore, the face features of different drivers need to be stored corresponding to the target observation offset angles.

[0151] In one possible implementation, the facial feature management module determines whether the driver's facial features exist in the observation offset angle management module.

[0152] In some embodiments, if the driver's facial features are present in the observation offset angle management module, the driver's target observation offset angle is updated to the observation offset angle management module based on the driver's facial features.

[0153] In some embodiments, if the driver's facial features are not present in the observation offset angle management module, the target observation offset angle is stored in the observation offset angle management module in correspondence with the driver's facial features.

[0154] In this implementation, the facial features of different drivers and the target observation offset angle are stored in the observation offset angle management module, so that the target observation offset angle is more adapted to the driving habits of different drivers.

[0155] Step 404: When the vehicle is not traveling straight, end the calculation of the target observation offset angle.

[0156] It should be understood that when a vehicle is not traveling straight, the driver's head will shift to observe the road conditions in order to ensure driving safety. In this case, if the target observation offset angle is calculated based on the driver's facial posture angle when the vehicle is not traveling straight, it will cause the target observation offset angle to deviate from the driver's normal forward-looking angle. This may easily misclassify the driver's normal forward-looking behavior as an abnormal posture. Therefore, the calculation of the target observation offset angle should be stopped when the vehicle is not traveling straight.

[0157] This application provides a method for calculating the target observation offset angle. When the vehicle is traveling straight, the method calculates the target observation offset angle and stores the target observation offset angle in the observation offset angle management module in correspondence with the driver's facial features. This allows the target observation offset angle to be obtained based on the facial features, thereby calibrating the driver's current facial posture angle and improving the accuracy of testing whether the driver is fatigued.

[0158] Figure 5 This is a schematic flowchart illustrating another dynamic calibration method for driver facial posture angles provided in this application embodiment.

[0159] For example, such as Figure 5 As shown, taking a microcontroller as the execution subject as an example, this application describes a dynamic calibration method for driver's facial posture angle. The method 500 includes the following steps.

[0160] Step 501, the face algorithm module detects the face frame of the driver in the face image and the current face posture angle.

[0161] It can be understood that the specific implementation of the face algorithm module detecting the face frame of the driver in the face image and the current face posture angle can refer to the related description of the foregoing step 201, which will not be repeated here.

[0162] Step 502, determine whether the vehicle is in a straight-ahead state.

[0163] It can be understood that the specific implementation of determining whether the vehicle is in a straight-ahead state can refer to the related description of the foregoing step 401, which will not be repeated here.

[0164] Optionally, after step 502, the following steps 503-508 are executed according to actual conditions, or the following step 509 is executed.

[0165] Step 503, when the vehicle is in a straight-ahead state, determine the angle interval in which the plurality of face posture angle samples are located.

[0166] It can be understood that the specific implementation of determining the angle interval in which the plurality of face posture angle samples are located can refer to the related description of the foregoing step 402, which will not be repeated here.

[0167] Step 504, count the second number of the plurality of face posture angle samples falling into the total angle interval and the first number of the plurality of face posture angle samples falling into each angle sub-interval.

[0168] It can be understood that the specific implementation of counting the second number of the plurality of face posture angle samples falling into the total angle interval and the first number of the plurality of face posture angle samples falling into each angle sub-interval can refer to the related description of the foregoing step 402, which will not be repeated here.

[0169] Step 505, determine whether the second number is greater than a first threshold.

[0170] It can be understood that the specific implementation of determining whether the second number is greater than a first threshold can refer to the related description of the foregoing step 402, which will not be repeated here.

[0171] Optionally, after step 505, the following steps 506-508 are executed according to actual conditions, or the following step 509 is executed.

[0172] Step 506, in the case where the second number is greater than the first threshold, calculate the sample quantity proportion of each angle sub-interval.

[0173] It can be understood that the specific implementation of determining whether the sample quantity proportion of the angle sub-interval is greater than the preset proportion or determining whether the sum of the sample quantity proportions of the continuous multiple angle sub-intervals is greater than the preset proportion can be referred to the related description of the foregoing step 402, and will not be repeated here.

[0174] In step 507, it is determined whether the sample quantity proportion of the angle sub-interval is greater than the preset proportion or whether the sum of the sample quantity proportions of the continuous multiple angle sub-intervals is greater than the preset proportion.

[0175] It can be understood that the specific implementation of determining whether the sample quantity proportion of the angle sub-interval is greater than the preset proportion or determining whether the sum of the sample quantity proportions of the continuous multiple angle sub-intervals is greater than the preset proportion can be referred to the related description of the foregoing step 402, and will not be repeated here.

[0176] Optionally, after step 507, the following step 508 or the following step 509 is executed according to actual conditions.

[0177] In step 508, in the case that the sample quantity proportion of the angle sub-interval is greater than the preset proportion or the sum of the sample quantity proportions of the continuous multiple angle sub-intervals is greater than the preset proportion, the observation offset angle management module is updated.

[0178] It can be understood that the specific implementation of updating the observation offset angle management module in the case that the sample quantity proportion of the angle sub-interval is greater than the preset proportion or the sum of the sample quantity proportions of the continuous multiple angle sub-intervals is greater than the preset proportion can be referred to the related description of the foregoing step 403, and will not be repeated here.

[0179] In step 509, the calculation of the target observation offset angle is ended.

[0180] In some embodiments, the calculation of the target observation offset angle is ended when the vehicle is in a non-straight-ahead state.

[0181] It can be understood that the specific implementation of ending the calculation of the target observation offset angle when the vehicle is in a non-straight-ahead state can be referred to the related description of the foregoing step 404.

[0182] In some embodiments, the calculation of the target observation offset angle is ended when the second quantity is less than or equal to the first threshold.

[0183] It should be understood that the second quantity being less than or equal to the first threshold indicates that the sample of the plurality of face posture angle samples falling into the total angle interval is too small. In this case, if the main angle interval is calculated based on the sample, the calculated main angle interval will be very unstable. Therefore, the calculation of the target observation offset angle is ended when the second quantity is less than or equal to the first threshold.

[0184] In some embodiments, the target observation offset angle is determined when the proportion of the sample quantity of the angle sub-interval is less than or equal to the preset proportion and the sum of the proportions of the sample quantities of the continuous multiple angle sub-intervals is less than or equal to the preset proportion.

[0185] It should be understood that the proportion of the sample quantity of the angle sub-interval being less than or equal to the preset proportion and the sum of the proportions of the sample quantities of the continuous multiple angle sub-intervals being less than or equal to the preset proportion indicates that the observation offset angle learning fails, and in this case, updating the observation offset angle management module will cause inaccurate target observation offset angles to appear in the observation offset angle management model, and thus the determination of the target observation offset angle is ended.

[0186] Figure 6 FIG. 1 is a structural schematic diagram of a dynamic calibration device for a driver's face posture angle according to an embodiment of the present application.

[0187] In an example, the device 600 includes:

[0188] The acquisition and extraction module 601 is configured to acquire a face image of a driver and extract a face feature and a current face posture angle of the driver from the face image.

[0189] The acquisition module 602 is configured to acquire a corresponding target observation offset angle from an observation offset angle management module based on the face feature, and the calculation process of the target observation offset angle includes: continuously acquiring a plurality of face posture angle samples of the driver when it is detected that the vehicle is in a straight driving state; identifying a main angle interval of the plurality of face posture angle samples based on a statistical quantity in the angle value distribution; and determining the target observation offset angle according to the main angle interval.

[0190] The calculation module 603 is configured to subtract the target observation offset angle from the current face posture angle to obtain a calibrated face posture angle.

[0191] In a possible implementation, the device 600 includes:

[0192] The acquisition module 602 is further configured to divide a preset angle total interval into a plurality of continuous angle sub-intervals.

[0193] The acquisition module 602 is further configured to count a first quantity of the plurality of face posture angle samples falling into each angle sub-interval and calculate a proportion of the sample quantity of each angle sub-interval.

[0194] The acquisition module 602 is further configured to determine an angle sub-interval with a proportion of the sample quantity greater than a preset proportion or a continuous multiple angle sub-intervals with a sum of the proportions of the sample quantities greater than the preset proportion as the main angle interval.

[0195] In a possible implementation, the device 600 includes:

[0196] The acquisition module 602 is also used to calculate the first average value of all facial pose angle samples within the main angle interval, and to determine the first average value as the target observation offset angle.

[0197] In one possible implementation, the device 600 includes:

[0198] The acquisition module 602 is also used to count the second number of the multiple facial pose angle samples falling into the total angle interval;

[0199] The acquisition module 602 is also used to perform the step of calculating the percentage of the number of samples in each angle sub-interval when the second quantity is greater than the first threshold.

[0200] In one possible implementation, the device 600 includes:

[0201] The acquisition module 602 is specifically used to determine whether there is a target observation offset angle in the observation offset angle management module when the face feature matches any candidate face feature in the observation offset angle management module.

[0202] The acquisition module 602 is specifically used to determine whether there is a historical offset angle in the observation offset angle management module based on the face feature when there is no target observation offset angle in the observation offset angle management module.

[0203] The acquisition module 602 is specifically used to determine the historical offset angle as the target observation offset angle when there is a historical offset angle in the observation offset angle management module.

[0204] In one possible implementation, the device 600 includes:

[0205] The acquisition module 602 is specifically used to calculate the second average value of all interval angles within the main angle interval when there are no historical offset angles in the observation offset angle management module.

[0206] The acquisition module 602 is specifically used to determine the second average value as the target observation offset angle.

[0207] In one possible implementation, the device 600 includes:

[0208] The acquisition and extraction module 601 is also used to acquire the vehicle speed and steering wheel angle;

[0209] The data acquisition and extraction module 601 is also used to determine that the vehicle is in a straight-going state when the vehicle speed is greater than the preset vehicle speed and the steering wheel angle is within the preset steering wheel angle range.

[0210] In one possible implementation, the device 600 includes:

[0211] The calculation module 603 is also used to determine whether the driver is driving dangerously based on the calibrated facial posture angle;

[0212] The calculation module 603 is also used to trigger a dangerous driving alarm message when the driver is driving dangerously.

[0213] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0214] For example, such as Figure 7 As shown, the electronic device 700 includes a memory 701 and a processor 702. The memory 701 stores executable program code 703, and the processor 702 is used to call and execute the executable program code 703 to perform a dynamic calibration method for the driver's facial posture angle.

[0215] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of dynamic calibration of a driver's facial pose angle, characterized in that, The method comprises: Collecting a face image of a driver, and extracting a face feature and a current face posture angle of the driver from the face image; Obtaining a corresponding target observation offset angle from an observation offset angle management module based on the face feature, the observation offset angle management module being used to store face features of different drivers and corresponding target observation offset angles thereof; the calculation process of the target observation offset angle comprising: continuously collecting a plurality of face posture angle samples of the driver when it is detected that the vehicle is in a straight driving state; identifying a main angle interval based on a statistical quantity in the angle value distribution of the plurality of face posture angle samples; determining the target observation offset angle according to the main angle interval; Subtracting the target observation offset angle from the current face posture angle to obtain a calibrated face posture angle; The identification of the main angle interval based on the statistical quantity in the angle value distribution of the plurality of face posture angle samples comprises: Dividing a preset angle total interval into a plurality of continuous angle subintervals; Counting a first quantity of the plurality of face posture angle samples falling into each angle subinterval, and calculating a sample quantity proportion of each angle subinterval; Determining an angle subinterval with a sample quantity proportion greater than a preset proportion, or a plurality of continuous angle subintervals with a sum of sample quantity proportions greater than the preset proportion, as the main angle interval.

2. The method of claim 1, wherein, The determination of the target observation offset angle according to the main angle interval comprises: Calculating a first average value of all face posture angle samples in the main angle interval, and determining the first average value as the target observation offset angle.

3. The method of claim 1, wherein, Before the calculation of the sample quantity proportion of each angle subinterval, the method further comprises: Counting a second quantity of the plurality of face posture angle samples falling into the angle total interval; In a case where the second quantity is greater than a first threshold value, performing the step of calculating the sample quantity proportion of each angle subinterval.

4. The method of claim 1, wherein, The obtaining of the corresponding target observation offset angle from the observation offset angle management module based on the face feature comprises: In a case where the face feature matches any candidate face feature in the observation offset angle management module, determining whether there is a target observation offset angle in the observation offset angle management module; In a case where there is no target observation offset angle in the observation offset angle management module, determining whether there is a historical offset angle in the observation offset angle management module based on the face feature; In a case where there is a historical offset angle in the observation offset angle management module, determining the historical offset angle as the target observation offset angle.

5. The method of claim 4, wherein, The method further comprises: In a case where there is no historical offset angle in the observation offset angle management module, calculating a second average value of all interval angles in the main angle interval; Determining the second average value as the target observation offset angle.

6. The method of claim 1, wherein, Before the continuous collection of the plurality of face posture angle samples of the driver when it is detected that the vehicle is in a straight driving state, the method further comprises: Obtaining a vehicle speed and a steering wheel angle of the vehicle; In a case where the vehicle speed is greater than a preset vehicle speed and the steering wheel angle is in a preset steering wheel angle range, it is determined that the vehicle is in a straight driving state.

7. The method of claim 1, wherein, After the current face posture angle is subtracted by the target observation offset angle to obtain a calibrated face posture angle, the method further includes: Determining whether the driver is in dangerous driving based on the calibrated face posture angle. In a case where the driver is in dangerous driving, triggering a dangerous driving alarm information.

8. A device for dynamic calibration of driver face pose angle, characterized in that, The device includes: A collection and extraction module configured to collect a face image of a driver and extract a face feature and a current face posture angle of the driver from the face image; An acquisition module configured to acquire a corresponding target observation offset angle from an observation offset angle management module based on the face feature, the observation offset angle management module being configured to store face features of different drivers and corresponding target observation offset angles of the face features, and a calculation process of the target observation offset angle including: continuously collecting a plurality of face posture angle samples of the driver when it is detected that the vehicle is in a straight driving state; identifying a main angle range of the plurality of face posture angle samples based on a statistical number in terms of angle value distribution; and determining the target observation offset angle according to the main angle range; A calculation module configured to subtract the current face posture angle by the target observation offset angle to obtain a calibrated face posture angle; divide a preset total angle range into a plurality of continuous angle sub-ranges; count a first number of the plurality of face posture angle samples falling into each angle sub-range and calculate a sample number proportion of each angle sub-range; and determine an angle sub-range with a sample number proportion greater than a preset proportion or a plurality of continuous angle sub-ranges with a sum of sample number proportions greater than the preset proportion as the main angle range.

9. An electronic device, comprising: The electronic device includes: A memory configured to store executable program codes; A processor configured to call and run the executable program codes from the memory, so that the electronic device performs the method according to any one of claims 1 to 7.

Images