Image recognition-based sentry warning method and system

Abstract

The invention provides a sentinel warning method and system based on image recognition, and belongs to the technical field of image processing, and the method specifically comprises the steps: obtaining the remaining electric quantity and a safety electric quantity threshold value of a vehicle-mounted storage battery of a vehicle, and carrying out the determination of a use safety value of the vehicle-mounted storage battery through the combination of use cycle data, the problem probability of the parking time period is determined by using the parking vehicle in the sentry warning mode in the parking time period within the latest preset time at the current place and the historical alarm data; determining the image monitoring frequency of the sentry warning mode in the parking period based on the use safety value, the problem probability and the historical traffic busy degree of the vehicle-mounted storage battery, and monitoring the vehicle environment based on the image monitoring frequency to obtain an environment image; and the image monitoring frequency of the sentry warning mode is dynamically adjusted based on the recognition result of the image recognition of the environment image in the parking period, so that the safety of the vehicle is further improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and particularly relates to a sentinel warning method and system based on image recognition. Background Art

[0002] Existing intelligent vehicles are often equipped with monitoring cameras and a sentinel mode. Through the settings of the monitoring cameras and the sentinel mode, when the vehicle has abnormal vibrations or other situations, videos will be recorded and pushed to the user. However, at the same time, due to the application of the sentinel mode, a large amount of electrical energy will be lost, and in severe cases, it may even cause the on-vehicle battery of the vehicle to run out of power and unable to start normally. Therefore, how to adaptively turn on the sentinel mode and give warnings has become a technical problem to be solved urgently.

[0003] In order to solve the above technical problems, in the invention patent CN202211493729.8 "Sentinel Mode Recommendation Method, Device, Equipment and Medium", by presetting sentinel mode recommendation conditions, the sentinel mode can be recommended when the conditions are met, and there is no need to recommend and turn on the sentinel mode when the conditions are not met, thus realizing the adaptive recommendation and turning on of the sentinel mode. However, there are the following technical problems: The existing technical solutions ignore turning on the sentinel mode according to the real-time state of the on-vehicle battery. Specifically, when the on-vehicle battery has different remaining capacities, the risk of power shortage of the vehicle caused by the use of the sentinel mode also varies greatly. Therefore, if the sentinel mode cannot be turned on based on the remaining capacity of the on-vehicle battery, it is impossible to ensure the safety of the vehicle while ensuring the safety of the vehicle's battery.

[0004] The existing technical solutions ignore selecting a differentiated sentinel warning mode according to the recognition results of image recognition. Specifically, for different monitoring locations, due to the differences in the number of vehicles and pedestrians at different times, the probability of scratching also varies to a certain extent. Therefore, if the above factors cannot be combined to select a differentiated sentinel warning mode, it is impossible to ensure the safety of the vehicle while ensuring the safety of the vehicle's battery.

[0005] In view of the above technical problems, the present invention provides a sentinel warning method and system based on image recognition. Summary of the Invention

[0006] To achieve the purpose of the present invention, the present invention adopts the following technical solutions: According to one aspect of the present invention, a sentinel warning method based on image recognition is provided.

[0007] A sentinel warning method based on image recognition, characterized in that it specifically includes: S1 Obtain the remaining power of the vehicle's on-vehicle battery and the safety power threshold, and combine the usage cycle data to determine the usage safety value of the on-vehicle battery. When the usage safety value meets the requirements, proceed to the next step; S2 Determine the parking period corresponding to the current moment, and determine the historical traffic busyness of the parking period through the historical vehicle flow and historical pedestrian flow at the current location on different dates. Based on the usage safety value, determine whether the historical traffic busyness meets the requirements. If so, proceed to the next step; if not, conduct vehicle environment monitoring through the real-time sentry warning mode. S3 Use the parked vehicles in the parking period in the sentry warning mode and the historical alarm data at the current location in the most recent preset time to determine the problem probability of the parking period, and determine the image monitoring frequency of the sentry warning mode for the parking period based on the usage safety value of the on-vehicle battery, the problem probability, and the historical traffic busyness. S4 Conduct vehicle environment monitoring based on the image monitoring frequency to obtain environmental images, and dynamically adjust the image monitoring frequency of the sentry warning mode based on the recognition results of the image recognition of the environmental images in the parking period.

[0008] The beneficial effects of the present invention are as follows: 1. By combining the remaining power of the vehicle's on-vehicle battery, the safety power threshold, and the usage cycle data to determine the usage safety value of the on-vehicle battery, it not only takes into account the differences in the safety power threshold for different vehicles due to differences in standby power consumption, but also takes into account the differences in the aging degree of the on-vehicle battery caused by the usage cycle data. Thus, it realizes the accurate assessment of the usage safety state based on the remaining power of the on-vehicle battery, and also lays a foundation for further realizing the differential setting of the sentry warning mode.

[0009] 2. By determining the image monitoring frequency of the sentry warning mode for the parking period based on the usage safety value of the on-vehicle battery, the problem probability, and the historical traffic busyness, it not only takes into account the usage safety state of the on-vehicle battery, but also takes into account the historical problem situations, pedestrian flow, and vehicle flow in different parking periods. It not only ensures the usage safety of the battery, but also ensures the differences in the requirements for the sentry warning mode under different problem probabilities and traffic flows.

[0010] 3. By dynamically adjusting the image monitoring frequency of the sentry warning mode based on the recognition results of the image recognition of the environmental images in the parking period, it realizes the dynamic adjustment of the image monitoring frequency of the sentry warning mode based on the differences in the pedestrian flow in the recognition results of the image recognition. On the basis of ensuring the monitoring reliability, it also reduces unnecessary power consumption.

[0011] A further technical solution lies in that the safe power threshold is determined according to the power consumption of the vehicle during a specified period of parking and the power consumption required for vehicle startup.

[0012] A further technical solution lies in that the usage cycle data includes the number of usage cycles of the vehicle-mounted battery and the depth of charge and discharge of the battery at different usage cycle numbers.

[0013] A further technical solution lies in that when the usage safety value does not meet the requirements, the sentry warning mode is suspended, and it is determined that vehicle environment monitoring cannot be performed.

[0014] A further technical solution lies in that the method for determining the usage safety value of the vehicle-mounted battery is as follows: Taking the deviation between the remaining power of the vehicle-mounted battery of the vehicle and the safe power threshold as the safe remaining power, and determining the power attenuation rate of the vehicle-mounted battery at different charge and discharge cycle numbers through the usage cycle data of the vehicle-mounted battery; Based on the average value of the power attenuation rate of the vehicle-mounted battery and the number of charge and discharge cycles with a power attenuation rate greater than the preset attenuation rate, determining the attenuation evaluation value of the vehicle-mounted battery, and determining whether there is an aging problem with the vehicle-mounted battery through the attenuation evaluation value. If so, proceed to the next step. If not, determine the usage safety value of the vehicle-mounted battery based on the safe remaining power; Based on the number of usage cycles of the vehicle-mounted battery, the depth of charge and discharge at different usage cycle numbers, and the attenuation evaluation value, determining the aging evaluation value of the vehicle-mounted battery, and correcting the safe remaining power based on the aging evaluation value to obtain the usage safety value of the vehicle-mounted battery.

[0015] A further technical solution lies in that the method for determining the traffic flow busyness degree of the parking period is as follows: Determining the traffic flow evaluation value of the current location during the parking period on different dates through the traffic flow and pedestrian flow of the current location during the parking period on different dates, and taking the maximum value of the traffic flow evaluation value as the reference evaluation value. Determining whether the accuracy of the reference evaluation value meets the requirements by the number of dates with a deviation less than the preset deviation from the reference evaluation value. If so, taking the reference evaluation value as the traffic flow busyness degree of the parking period. If not, determining the traffic flow busyness degree of the parking period through the average value of the reference evaluation value.

[0016] A further technical solution lies in that determining whether the traffic flow busyness degree meets the requirements based on the usage safety value specifically includes: Determine the busyness threshold of the vehicle during different parking periods using the usage safety value, and when the traffic busyness is greater than the busyness threshold, determine that the traffic busyness does not meet the requirements.

[0017] A further technical solution lies in that the method for determining the image monitoring frequency is as follows: Determine the image-based monitoring frequency of the sentry warning mode during the parking period through the usage safety value of the vehicle-mounted battery, and determine whether the problem probability is greater than a preset probability value. If so, determine the image monitoring frequency of the sentry warning mode through the image-based monitoring frequency. If not, proceed to the next step; Determine the frequency correction amount through the problem probability and the historical traffic busyness, and determine the image monitoring frequency of the sentry warning mode during the parking period based on the product of the frequency correction amount and the image-based monitoring frequency.

[0018] On the other hand, the present invention provides a sentry warning system based on image recognition, adopting the above-mentioned sentry warning method based on image recognition, which is characterized in that it specifically includes: A battery evaluation module, a busyness evaluation module, a warning mode determination module, and a warning mode adjustment module; Among them, the battery evaluation module is responsible for obtaining the remaining power and the safety power threshold of the vehicle-mounted battery of the vehicle, and determining the usage safety value of the vehicle-mounted battery in combination with the usage cycle data; The busyness evaluation module is responsible for determining the parking period corresponding to the current moment, and determining the historical traffic busyness of the parking period through the historical vehicle flow and historical pedestrian flow at the current location on different dates; The warning mode determination module is responsible for using the parked vehicles in the sentry warning mode during the parking period at the current location within the recent preset time and the historical alarm data to determine the problem probability of the parking period, and determining the image monitoring frequency of the sentry warning mode during the parking period based on the usage safety value of the vehicle-mounted battery, the problem probability, and the historical traffic busyness; The warning mode adjustment module is responsible for monitoring the vehicle environment based on the image monitoring frequency to obtain an environmental image, and dynamically adjusting the image monitoring frequency of the sentry warning mode based on the recognition result of the image recognition of the environmental image during the parking period.

[0019] Other features and advantages will be described in the subsequent specification, and part of them will become obvious from the specification or be understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification and the drawings.

[0020] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following specifically gives preferred embodiments and, in conjunction with the accompanying drawings, makes a detailed description as follows. Description of the Drawings

[0021] By referring to the drawings and describing in detail its exemplary embodiments, the above and other features and advantages of the present invention will become more obvious; Figure 1 is a flowchart of a sentinel warning method based on image recognition; Figure 2 is a flowchart of a method for determining the safe usage value of a vehicle-mounted battery; Figure 3 is a flowchart of another possible method for determining the safe usage value of a vehicle-mounted battery; Figure 4 is a flowchart of a method for determining the problem probability during the parking period; Figure 5 is a flowchart of a method for determining the image monitoring frequency; Figure 6 is a framework diagram of a sentinel warning system based on image recognition; Detailed Embodiments

[0022] In order to enable those skilled in the art to better understand the technical solutions in this specification, the following will clearly and completely describe the technical solutions in the embodiments of this specification in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all the embodiments. Based on the embodiments of this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this specification.

[0023] Problems with the existing technical solutions: The existing sentinel warning mode often performs real-time monitoring and automatically starts recording when the vehicle has abnormal vibrations or other situations. It not only fails to consider the overuse of the sentinel warning mode caused by differences in the state of the vehicle's on-board battery, resulting in situations such as the on-board battery running out of power, but also fails to consider historical warning events such as vehicle scratches and thefts at different parking positions during different parking periods, thus unable to achieve dynamic adjustment of the sentinel warning mode.

[0024] To solve the above technical problems, the following technical means are mainly adopted: First, determine the safety battery margin of the vehicle's on-vehicle battery based on the remaining power of the on-vehicle battery and the safety power threshold, and determine the aging coefficient according to the ratio of the number of usage cycles to the number of life cycles of the on-vehicle battery. Determine the usage safety value by multiplying the ratio of the safety battery margin to the rated capacity of the on-vehicle battery by the aging coefficient. When the usage safety value does not meet the requirements, the sentinel warning mode is not enabled at this time. When the usage safety value meets the requirements, proceed to the next step; Then, determine the historical vehicle flow and historical pedestrian flow during the parking period at the current location on different dates according to the parking period corresponding to the current time, and determine the historical flow busyness of the parking period according to the ratio of the average value of the historical vehicle flow and historical pedestrian flow during the parking period on different dates to the preset vehicle flow and pedestrian flow. When the historical flow busyness is large, perform vehicle environment monitoring through the real-time sentinel warning mode. When the historical flow busyness is small, determine the problem probability of the parking period by using the parked vehicles in the sentinel warning mode and the historical alarm data during the parking period at the current location within the nearest preset time. Specifically, determine the problem probability of the parking period according to the ratio of the historical alarm data to the parked vehicles in the sentinel warning mode, and determine the image monitoring frequency of the sentinel warning mode for the parking period based on the minimum value of the preset image monitoring frequency corresponding to the usage safety value of the on-vehicle battery, the problem probability, and the historical flow busyness; Finally, perform vehicle environment monitoring based on the image monitoring frequency to obtain environmental images, and dynamically adjust the image monitoring frequency of the sentinel warning mode based on the recognition result of the image recognition of the environmental images during the parking period. Specifically, when there is an alarm event in the recognition result of the image recognition of the environmental images during the parking period, perform vehicle environment monitoring through the real-time sentinel warning mode. When there is no alarm event, dynamically adjust the image monitoring frequency according to the vehicle flow and pedestrian flow in the recognition result of the image recognition of the environmental images during the parking period.

[0025] The following will be elaborated in detail from two perspectives: system - type embodiments and method - type embodiments.

[0026] To solve the above problems, according to one aspect of the present invention, as Figure 1 shown, a sentinel warning method based on image recognition is provided, which is characterized by specifically including: S1 Obtain the remaining power of the vehicle's on - vehicle battery and the safety power threshold, and determine the usage safety value of the on - vehicle battery in combination with the usage cycle data. When the usage safety value meets the requirements, proceed to the next step; Furthermore, the safety power threshold is determined according to the power consumption of the vehicle during the specified duration of parking and the power consumption required for vehicle startup.

[0027] Specifically, the usage cycle data includes the usage cycle times of the vehicle-mounted battery and the charge-discharge depth of the battery at different usage cycle times.

[0028] In one possible embodiment, as Figure 2 shown, the method for determining the usage safety value of the vehicle-mounted battery in the above step S1 is as follows: S11: Take the deviation between the remaining power of the vehicle-mounted battery of the vehicle and the safety power threshold as the safe remaining power, and determine whether the safe remaining power meets the requirements. If so, proceed to the next step; if not, determine the usage safety value of the vehicle-mounted battery based on the safe remaining power. S12: Obtain the usage cycle data of the vehicle-mounted battery of the vehicle, and determine the usage cycle times of the vehicle-mounted battery of the vehicle based on the usage cycle data. Determine whether there is an aging problem with the vehicle-mounted battery through the usage cycle times. If so, proceed to step S14; if not, proceed to the next step. S13: Determine the power attenuation rate of the vehicle-mounted battery at different charge-discharge cycle times through the usage cycle data of the vehicle-mounted battery, and determine the attenuation evaluation value of the vehicle-mounted battery based on the average value of the power attenuation rate of the vehicle-mounted battery and the charge-discharge cycle times with the power attenuation rate greater than the preset attenuation rate. Determine whether there is an aging problem with the vehicle-mounted battery through the attenuation evaluation value. If so, proceed to step S14; if not, determine the usage safety value of the vehicle-mounted battery based on the safe remaining power. In one possible embodiment, the attenuation evaluation value of the vehicle-mounted battery is determined according to the product of the ratio of the average value of the power attenuation rate of the vehicle-mounted battery to the charge-discharge cycle times with the power attenuation rate greater than the preset attenuation rate in all charge-discharge cycle times.

[0029] S14: Determine the aging evaluation value of the vehicle-mounted battery based on the usage cycle times of the vehicle-mounted battery, the charge-discharge depth at different usage cycle times, and the attenuation evaluation value, and correct the safe remaining power based on the aging evaluation value to obtain the usage safety value of the vehicle-mounted battery.

[0030] Specifically, first determine the usage cycle times with a higher charge-discharge depth according to the charge-discharge depth at different usage cycle times, then determine the charge-discharge cycle aging coefficient according to the usage cycle times with a higher charge-discharge depth and the ratio of the usage cycle times to the service life times of the vehicle-mounted battery, and finally determine the aging evaluation value of the vehicle-mounted battery according to the sum of the weights of the charge-discharge aging coefficient and the attenuation evaluation value.

[0031] It can be understood that when the usage safety value does not meet the requirements, the sentinel warning mode is suspended, and it is determined that vehicle environment monitoring cannot be performed.

[0032] In another possible embodiment, as Figure 3 shown, the method for determining the usage safety value of the vehicle-mounted battery in the above step S1 is: Taking the deviation between the remaining power of the vehicle-mounted battery of the vehicle and the safety power threshold as the safety remaining power, and determining the power attenuation rate of the vehicle-mounted battery at different charge and discharge cycle times through the usage cycle data of the vehicle-mounted battery; Based on the average value of the power attenuation rate of the vehicle-mounted battery and the number of charge and discharge cycles with a power attenuation rate greater than the preset attenuation rate, determining the attenuation evaluation value of the vehicle-mounted battery, and determining whether there is an aging problem with the vehicle-mounted battery through the attenuation evaluation value. If so, proceed to the next step. If not, determine the usage safety value of the vehicle-mounted battery based on the safety remaining power; Based on the usage cycle times of the vehicle-mounted battery, the charge and discharge depth at different usage cycle times, and the attenuation evaluation value, determining the aging evaluation value of the vehicle-mounted battery, and correcting the safety remaining power based on the aging evaluation value to obtain the usage safety value of the vehicle-mounted battery.

[0033] S2 Determine the parking period corresponding to the current moment, and determine the historical traffic flow density of the parking period through the historical traffic flow and historical pedestrian flow of the current location at different dates. Based on the usage safety value, determine whether the historical traffic flow density meets the requirements. If so, proceed to the next step. If not, perform vehicle environment monitoring through the real-time sentinel warning mode; Specifically, the method for determining the traffic flow density of the parking period is: Determine the traffic flow evaluation value of the current location at the parking period on different dates through the traffic flow and pedestrian flow at the parking period of the current location on different dates, and take the maximum value of the traffic flow evaluation value as the reference evaluation value. Determine whether the accuracy of the reference evaluation value meets the requirements by the number of dates with a deviation less than the preset deviation from the reference evaluation value. If so, take the reference evaluation value as the traffic flow density of the parking period. If not, determine the traffic flow density of the parking period through the average value of the reference evaluation value.

[0034] It should be further noted that determining whether the traffic flow density meets the requirements based on the usage safety value specifically includes: Determine the busyness threshold of the vehicle during different parking periods through the use safety value, and when the traffic busyness is greater than the busyness threshold, determine that the traffic busyness does not meet the requirements.

[0035] S3 Determine the problem probability of the parking period by using the parked vehicles in the sentry warning mode during the parking period at the current location in the recent preset time and the historical alarm data. Determine the image monitoring frequency of the sentry warning mode during the parking period based on the use safety value of the vehicle-mounted battery, the problem probability, and the historical traffic busyness; In one possible embodiment, as Figure 4 shown, the method for determining the problem probability of the parking period in step S2 above is: S21 Use the parked vehicles in the sentry mode during the parking period at the current location in the recent preset time as the parked monitoring vehicles, and determine the historical alarm times of the parked monitoring vehicles based on the historical alarm data of the parked monitoring vehicles. Determine whether the historical alarm times are within the preset number requirement range through the number of the parked monitoring vehicles. If so, determine the problem probability of the parking period through the historical alarm times and the number of the parked monitoring vehicles. If not, proceed to the next step; S22 Calculate the proportion of the number of dates with historical alarm data of the parked monitoring vehicles within the preset time according to the historical alarm data of the parked monitoring vehicles, and determine whether the proportion of the number of dates with historical alarm data is less than the preset ratio. If so, determine the problem probability of the parking period through the proportion of the number of dates with historical alarm data of the parked monitoring vehicles within the preset time. If not, proceed to the next step; S23 Use the dates with historical alarm data of the parked monitoring vehicles within the preset time as the alarm dates, and determine the problem evaluation amount for different alarm dates according to the number of parked monitoring vehicles and the historical alarm times on different alarm dates. Determine the serious problem dates and general problem dates in the alarm dates based on the problem evaluation amount; S24 Obtain the number and proportion of the serious problem dates in the alarm dates, and determine the problem probability of the parking period in combination with the number of the alarm dates and the problem evaluation amount.

[0036] In another possible embodiment, the method for determining the problem probability of the parking period in step S2 above is: Use the parked vehicles in the sentry mode during the parking period at the current location in the recent preset time as the parked monitoring vehicles, and determine the historical alarm times of the parked monitoring vehicles based on the historical alarm data of the parked monitoring vehicles; Take the date when the parked surveillance vehicle has historical alarm data within a preset time as the alarm date, and determine the problem evaluation quantity of different alarm dates according to the number of parked surveillance vehicles and the historical alarm times on different alarm dates. Judge whether there is an alarm date whose problem evaluation quantity does not meet the requirements. If so, determine the problem probability of the parking period through the maximum value of the problem evaluation quantity. If not, proceed to the next step; Based on the problem evaluation quantity, determine the serious problem dates and general problem dates in the alarm date. Determine the weight of the alarm date according to the time duration between the alarm date and the current date. Determine the sum of the weights of the problem evaluation quantity of the serious problem dates in the alarm date through the weight of the serious problem dates in the alarm date and the problem evaluation quantity. Judge whether the sum of the weights of the problem evaluation quantity of the serious problem dates in the warning date meets the requirements. If so, determine the problem probability of the parking period through the product of the proportion of the serious problem dates in the warning date and the sum of the weights of the problem evaluation quantity of the serious problem dates. If not, proceed to the next step; Obtain the number and proportion of the serious problem dates in the alarm date, and determine the problem probability of the parking period in combination with the number of the alarm dates and the problem evaluation quantity.

[0037] S4: Based on the image monitoring frequency, conduct vehicle environment monitoring to obtain environmental images, and dynamically adjust the image monitoring frequency of the sentry warning mode based on the recognition results of the image recognition of the environmental images during the parking period.

[0038] In one possible embodiment, as Figure 5 shown, the method for determining the image monitoring frequency in the above step S4 is as follows: S41: Determine the basic image monitoring frequency of the sentry warning mode during the parking period through the use safety value of the vehicle-mounted battery, and judge whether the problem probability is greater than the preset probability value. If so, determine the image monitoring frequency of the sentry warning mode through the basic image monitoring frequency. If not, proceed to the next step; S42: Determine the frequency correction quantity through the problem probability and the historical traffic busyness degree, and determine the image monitoring frequency of the sentry warning mode during the parking period based on the product of the frequency correction quantity and the basic image monitoring frequency.

[0039] It can be understood that dynamically adjusting the image monitoring frequency of the sentry warning mode based on the recognition results of the image recognition of the environmental images during the parking period specifically includes: Determine whether there is a preset alarm event during the parking period based on the recognition result of image recognition of the environmental image during the parking period. If so, switch the sentry warning mode to the real-time sentry warning mode. If not, proceed to the next step; Determine the environmental image including pedestrians or vehicles during the parking period based on the recognition result of the image recognition, and use it as the screened environmental image. Determine whether mode switching is required based on the proportion of the screened environmental image in the environmental image. If so, switch the sentry warning mode to the real-time sentry warning mode. If not, proceed to the next step; Determine the number of people and vehicle flow in the screened environmental image through the screened environmental image, and determine whether the number of people and vehicle flow in the screened environmental image meet the requirements based on the image monitoring frequency and monitoring duration during the parking period. If so, proceed to the next step. If not, switch the sentry warning mode to the real-time sentry warning mode; Divide the screened environmental image into short-term screened environmental images and long-term screened environmental images according to the time corresponding to the screened environmental image, and determine the short-term flow busyness during the parking period based on the number of short-term screened environmental images, the number of people and vehicle flow in the short-term screened environmental images during the parking period. Determine whether mode switching is required based on the short-term flow busyness. If so, switch the sentry warning mode to the real-time sentry warning mode. If not, proceed to the next step; Obtain the image monitoring frequency and monitoring duration during the parking period, and determine the real-time busyness during the parking period by combining the number of people and vehicle flow in the screened environmental image. Determine the frequency correction amount of the image monitoring frequency of the sentry warning mode based on the real-time busyness and the short-term flow busyness, and dynamically adjust the image monitoring frequency of the sentry warning mode based on the frequency correction amount.

[0040] System class embodiments On the other hand, as Figure 6 shown, the present invention provides a sentry warning system based on image recognition, adopting the above-mentioned sentry warning method based on image recognition, which is characterized in that it specifically includes: A battery evaluation module, a busyness evaluation module, a warning mode determination module, and a warning mode adjustment module; Wherein the battery evaluation module is responsible for obtaining the remaining power of the vehicle's on-vehicle battery and the safety power threshold, and determining the use safety value of the on-vehicle battery by combining the usage cycle data; The busyness evaluation module is responsible for determining the parking period corresponding to the current moment, and determining the historical flow busyness of the parking period through the historical vehicle flow and historical pedestrian flow of the current location during the parking periods on different dates; The warning mode determination module is responsible for determining the problem probability during the parking period by using the parked vehicles in the sentinel warning mode during the parking period at the current location within the recent preset time and historical alarm data, and determining the image monitoring frequency of the sentinel warning mode during the parking period based on the usage safety value of the vehicle-mounted battery, the problem probability, and the historical traffic busyness; The warning mode adjustment module is responsible for monitoring the vehicle environment based on the image monitoring frequency to obtain an environmental image, and dynamically adjusting the image monitoring frequency of the sentinel warning mode based on the recognition result of the image recognition of the environmental image during the parking period.

[0041] Based on the above embodiments, the present application achieves the following beneficial effects: 1. By combining the remaining power of the vehicle-mounted battery, the safety power threshold, and the usage cycle data to determine the usage safety value of the vehicle-mounted battery, it not only takes into account the differences in the safety power threshold for different vehicles due to differences in standby power consumption, but also takes into account the differences in the aging degree of the vehicle-mounted battery caused by the usage cycle data, thus realizing an accurate assessment of the usage safety state according to the remaining power of the vehicle-mounted battery, and also laying a foundation for further realizing the differential setting of the sentinel warning mode.

[0042] 2. By determining the image monitoring frequency of the sentinel warning mode during the parking period based on the usage safety value of the vehicle-mounted battery, the problem probability, and the historical traffic busyness, it not only takes into account the usage safety state of the vehicle-mounted battery, but also takes into account the historical problem situations, pedestrian flow, and vehicle flow during different parking periods, ensuring both the usage safety of the battery and the differences in the requirements for the sentinel warning mode under different problem probabilities and traffic flows.

[0043] 3. By dynamically adjusting the image monitoring frequency of the sentinel warning mode based on the recognition result of the image recognition of the environmental image during the parking period, it realizes the dynamic adjustment of the image monitoring frequency of the sentinel warning mode based on the difference in the pedestrian flow in the recognition result of the image recognition, reducing unnecessary power consumption on the basis of ensuring the monitoring reliability.

[0044] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. In particular, for the embodiments of the device, equipment, and non-volatile computer storage medium, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0045] The above description has been made of specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0046] The foregoing is only one or more embodiments of this specification and is not intended to limit this specification. For those skilled in the art, various changes and modifications can be made to one or more embodiments of this specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of this specification shall be included within the scope of the claims of this specification.

Claims

1. A sentinel warning method based on image recognition, characterized in that, Specifically, it includes: Obtain the remaining power of the vehicle's on-vehicle battery and the safety power threshold, and combine the use of cycle data to determine the usage safety value of the on-vehicle battery. When the usage safety value meets the requirements, proceed to the next step; Determine the parking period corresponding to the current moment, and determine the historical traffic flow and historical pedestrian flow of the parking period at the current location on different dates to determine the historical traffic flow busyness of the parking period. Based on the usage safety value, determine whether the historical traffic flow busyness meets the requirements. If so, proceed to the next step. If not, perform vehicle environment monitoring through the real-time sentry warning mode; Use the parked vehicles in the parking period in the sentry warning mode and historical alarm data at the current location in the recent preset time to determine the problem probability of the parking period. Based on the usage safety value of the on-vehicle battery, the problem probability, and the historical traffic flow busyness, determine the image monitoring frequency of the sentry warning mode for the parking period; Perform vehicle environment monitoring based on the image monitoring frequency to obtain an environment image, and dynamically adjust the image monitoring frequency of the sentry warning mode based on the recognition result of the image recognition of the environment image in the parking period.

2. The sentinel warning method based on image recognition according to claim 1, characterized in that, The safety power threshold is determined according to the power consumption of the vehicle during the specified duration of parking and the power consumption required for vehicle startup.

3. The sentinel warning method based on image recognition according to claim 1, wherein The usage cycle data includes the usage cycle times of the on-vehicle battery and the battery charge-discharge depth at different usage cycle times.

4. The sentinel warning method based on image recognition according to claim 1, characterized in that, When the usage safety value does not meet the requirements, suspend the activation of the sentry warning mode and determine that vehicle environment monitoring cannot be performed.

5. The sentinel warning method based on image recognition according to claim 1, characterized in that, The method for determining the usage safety value of the on-vehicle battery is: Take the deviation between the remaining power of the vehicle's on-vehicle battery and the safety power threshold as the safety remaining power, and use the usage cycle data of the on-vehicle battery to determine the power attenuation rate of the on-vehicle battery at different charge-discharge cycle times; Based on the average value of the power attenuation rate of the on-vehicle battery and the charge-discharge cycle times with a power attenuation rate greater than the preset attenuation rate, determine the attenuation evaluation value of the on-vehicle battery. Determine whether the on-vehicle battery has an aging problem through the attenuation evaluation value. If so, proceed to the next step. If not, determine the usage safety value of the on-vehicle battery based on the safety remaining power; Based on the usage cycle times of the on-vehicle battery, the charge-discharge depth at different usage cycle times, and the attenuation evaluation value, determine the aging evaluation value of the on-vehicle battery, and correct the safety remaining power based on the aging evaluation value to obtain the usage safety value of the on-vehicle battery.

6. The sentinel warning method based on image recognition according to claim 1, wherein The method for determining the traffic flow busyness of the parking period is: Determine the traffic evaluation value of the parking period at the current location on different dates based on the traffic flow and pedestrian flow during the parking periods at the current location on different dates, and take the maximum value of the traffic evaluation values as the reference evaluation value. Determine whether the accuracy of the reference evaluation value meets the requirements by the number of dates with a deviation from the reference evaluation value less than a preset deviation. If so, take the reference evaluation value as the traffic busyness of the parking period. If not, determine the traffic busyness of the parking period through the average value of the reference evaluation values.

7. The sentinel warning method based on image recognition according to claim 1, characterized in that, Based on the usage safety value, determine whether the traffic busyness meets the requirements, specifically including: Determine the busyness threshold of the vehicle during different parking periods through the usage safety value, and when the traffic busyness is greater than the busyness threshold, determine that the traffic busyness does not meet the requirements.

8. The sentinel warning method based on image recognition according to claim 1, characterized in that, The method for determining the problem probability of the parking period is: Regard the parked vehicles in the sentry mode during the parking period at the current location within the recent preset time as parked surveillance vehicles, and determine the historical alarm times of the parked surveillance vehicles based on the historical alarm data of the parked surveillance vehicles. Determine whether the historical alarm times are within the range of the preset number of times required through the number of parked surveillance vehicles. If so, determine the problem probability of the parking period through the historical alarm times and the number of parked surveillance vehicles. If not, proceed to the next step; Calculate the proportion of the number of dates with historical alarm data of the parked surveillance vehicles within the preset time based on the historical alarm data of the parked surveillance vehicles, and determine whether the proportion of the number of dates with historical alarm data is less than the preset proportion. If so, determine the problem probability of the parking period through the proportion of the number of dates with historical alarm data of the parked surveillance vehicles within the preset time. If not, proceed to the next step; Regard the dates with historical alarm data of the parked surveillance vehicles within the preset time as alarm dates, and determine the problem evaluation amounts for different alarm dates based on the number of parked surveillance vehicles and the historical alarm times on different alarm dates. Determine the serious problem dates and general problem dates in the alarm dates based on the problem evaluation amounts. Obtain the number and proportion of the serious problem dates in the alarm dates, and determine the problem probability of the parking period in combination with the number of alarm dates and the problem evaluation amounts.

9. The sentinel warning method based on image recognition according to claim 1, wherein, The method for determining the image monitoring frequency is: Determine the basic image monitoring frequency of the sentry warning mode during the parking period through the usage safety value of the vehicle-mounted battery, and determine whether the problem probability is greater than the preset probability value. If so, determine the image monitoring frequency of the sentry warning mode through the basic image monitoring frequency. If not, proceed to the next step; Determine the frequency correction amount through the problem probability and the historical traffic busyness, and determine the image monitoring frequency of the sentry warning mode during the parking period based on the product of the frequency correction amount and the basic image monitoring frequency.

10. A sentinel warning system based on image recognition, which adopts a sentinel warning method based on image recognition according to any one of claims 1-9, characterized in that, Specifically including: Battery evaluation module, busyness evaluation module, warning mode determination module, warning mode adjustment module; The battery evaluation module is responsible for obtaining the remaining power of the vehicle's on-vehicle battery and the safety power threshold, and determining the usage safety value of the on-vehicle battery by combining and using cycle data; The busyness evaluation module is responsible for determining the parking period corresponding to the current moment, and determining the historical traffic busyness of the parking period through the historical traffic flow and historical pedestrian flow of the current location during the parking periods on different dates; The warning mode determination module is responsible for using the parked vehicles in the sentry warning mode during the parking period at the current location within the recent preset time and the historical alarm data to determine the problem probability of the parking period, and determining the image monitoring frequency of the sentry warning mode for the parking period based on the usage safety value of the on-vehicle battery, the problem probability, and the historical traffic busyness; The warning mode adjustment module is responsible for performing vehicle environment monitoring based on the image monitoring frequency to obtain an environmental image, and dynamically adjusting the image monitoring frequency of the sentry warning mode based on the recognition result of the image recognition of the environmental image during the parking period.

Images