A vehicle-mounted fragrance control method, control system and pet fragrance box

Abstract

The method of the present invention provides a vehicle-mounted fragrance control method, a control system and a pet fragrance box, by arranging a temperature sensor, a humidity sensor, an infrared sensor and a photometric sensor on the vehicle body to collect environmental characteristic values, using a vehicle-mounted microphone to collect the voice of the pet and identifying the pet's voice characteristic data through a voice recognition system, using a vehicle-mounted camera to collect the image of the pet and identifying the type characteristic data of the pet through a target detection system, inputting the above-mentioned characteristic data into a fragrance decision model, and outputting the predicted result according to the neural network, thereby controlling the fragrance box at the bottom of the vehicle body to start the corresponding concentration and type of fragrance to be directional released to the pet. The present invention can release the pet concentration and control the air concentration range according to relevant control parameters, ensuring that the fragrance gas is always kept in the pet area, does not conflict with human use, and meets the needs of both humans and pets.

Description

Technical Field

[0001] The present invention relates to the field of vehicle technology, and in particular to a vehicle-mounted fragrance control method, a control system and a pet fragrance box. Background Art

[0002] At present, car fragrances are widely used in cars. They can kill bacteria in the air inside the car to keep the air clean and purify the air. In addition, car fragrances can also create a fresh and pleasant air atmosphere in the narrow space inside the car to keep the driver and passengers in the car clear-headed and calm. It can not only effectively reduce the occurrence of traffic accidents and benefit driving safety, but also add elegance to the car.

[0003] Existing in-car fragrance systems mainly provide related services for drivers and passengers, and the modes need to be set by users, or are simple centralized adaptive modes. In addition, the types of fragrances are relatively simple, and only the concentration is controlled. There is no fragrance air function for pets, and the installation positions are mostly at the air outlet or above the cabin. At the same time, the concentration is too high, and it is not a pet-type fragrance reagent. Pets have a sensitive sense of smell and may feel uncomfortable after smelling human fragrance. Summary of the invention

[0004] The present invention provides a vehicle-mounted fragrance control method, characterized in that the method specifically comprises:

[0005] S100: When the vehicle is started, characteristic data of the pet in the vehicle is obtained.

[0006] S200: Processing the characteristic data using a fragrance decision model to calculate fragrance release parameters.

[0007] S300: Adjusting the type and concentration of the fragrance in the vehicle according to the fragrance release parameter.

[0008] In the above technical scheme, pet fragrance boxes are arranged in the car according to needs. The pet fragrance boxes are equipped with common animal fragrance essential oils, concentration control chips, and area control chips. In addition, temperature sensors, humidity sensors, infrared sensors, and photometric sensors are arranged on the car body to collect environmental characteristic values. The pet's voice is collected by a car microphone and the pet's voice characteristic data is identified by a voice recognition system. The pet's image is collected by a car camera and the pet's type characteristic data is identified by a target detection system. The above-mentioned characteristic data are input into the fragrance decision model, and the predicted results are output according to the neural network, so as to control the fragrance box at the bottom of the car body to start the targeted release of fragrance of corresponding concentration and type to the pet.

[0009] Furthermore, the acquisition of characteristic data of the pet in the vehicle at least includes:

[0010] Environmental characteristic values ​​obtained based on the temperature, humidity, infrared and light intensity inside the car.

[0011] The pet emotion feature value is obtained based on the collected pet sound spectrum in the car.

[0012] The pet category feature value is obtained based on the collected pet image in the car.

[0013] In the above technical solution, the temperature, humidity, infrared and luminosity in the car are acquired through temperature sensors, humidity sensors, infrared sensors and luminosity sensors respectively, which can ensure the real-time acquisition of the environmental characteristic values ​​in the car, provide accuracy and efficiency in controlling the pet fragrance box, and input the spectral noise of the pet sound in the car into the trained neural network to obtain the pet emotion characteristic value; the image of the pet in the car is input into the trained neural network to obtain the pet category characteristic value.

[0014] Furthermore, the step S200 specifically includes:

[0015] S201: performing regularization and normalization processing on the feature data.

[0016] S202: Classify the processed feature data to obtain fragrance parameters and probabilities corresponding to each feature data; the fragrance parameters include at least fragrance concentration and fragrance type; the probabilities include at least the probability corresponding to the fragrance concentration and the probability corresponding to the fragrance type.

[0017] S203: Determine whether there are fragrance parameters that do not match the current pet type based on the pet category feature value and the preset category catalog; if so, remove the fragrance parameters that do not match the conditions, calculate the average value of the probabilities corresponding to the remaining fragrance parameters, and use the category with the largest average value as the fragrance release parameter; otherwise, calculate the average value of the probabilities corresponding to all the fragrance parameters, and use the category with the largest average value as the fragrance release parameter.

[0018] In the above technical solution, the fragrance decision model is a neural network, which is a type of supervised learning. The characteristic of supervised learning is that each sample in the training sample set has a corresponding label, that is, each sample has its own corresponding category. The result obtained by the activation function of the neural network is actually the probability of which category the feature data belongs to. For example, the training sample set has two categories: fragrance concentration and fragrance type. After the feature data is input into the neural network, the result category and category corresponding probability are the probability of a certain fragrance type and the probability of a certain concentration, respectively.

[0019] Furthermore, the preset category catalog includes pet types and fragrance parameters associated with the pet types.

[0020] In the above technical solution, associations are established in advance between different pet types and corresponding fragrance concentrations and fragrance types and stored, and individual pet characteristics are pre-set, such as body shape, breed, age, preferences and health status. Corresponding associations between corresponding fragrance concentrations and fragrance types of pet individual characteristics are established accordingly, which can more accurately control the pet fragrance box to release suitable fragrance.

[0021] Furthermore, the fragrance decision model is any one of a convolutional neural network, a deep neural network and a BP neural network.

[0022] Furthermore, the step S300 specifically includes:

[0023] S301: defining the fragrance release parameter as a working parameter of a fragrance regulator.

[0024] S302: adjusting the type and concentration of the in-vehicle fragrance according to the working parameters, and emitting a fragrance matching the fragrance release parameters.

[0025] In the above technical solution, the fragrance box releases the pet concentration and controls the air concentration range according to relevant control parameters to ensure that the fragrance gas always remains in the pet area and does not conflict with human use.

[0026] As another preferred embodiment, the present invention further provides a pet fragrance box control system, the system at least comprising:

[0027] The acquisition module is used to acquire the characteristic data of the pet in the vehicle when the vehicle is started, and send the characteristic data to the calculation module.

[0028] The calculation module is used to construct a fragrance decision model to process the characteristic data, calculate the fragrance release parameters, and send the adjustment parameters to the control module.

[0029] The control module is used to adjust the type and concentration of the fragrance in the vehicle according to the fragrance release parameters.

[0030] Furthermore, the acquisition module at least includes:

[0031] The sensor module is used to obtain environmental characteristic values ​​according to the temperature, humidity, infrared and light intensity in the vehicle; the sensor module is provided with at least a temperature sensor, a humidity sensor, an infrared sensor and a light intensity sensor.

[0032] The sound recognition module is used to obtain the pet emotion feature value based on the frequency spectrum of the pet sound in the car.

[0033] The object detection module is used to obtain pet category feature values ​​based on the image of the pet in the car.

[0034] Furthermore, the computing module is embedded with a fragrance decision model, and the fragrance decision model at least includes:

[0035] The input layer is used to normalize and regularize the feature data, and send the processed feature data to the logical discrimination layer.

[0036] The logic discrimination layer is used to classify the processed feature data to obtain the corresponding fragrance parameters and corresponding probabilities of each feature data; the fragrance parameters include at least fragrance concentration and fragrance type; the probability includes at least the probability corresponding to the fragrance concentration and the probability corresponding to the fragrance type.

[0037] The judgment and calculation layer is used to judge whether there are fragrance parameters that do not meet the current pet type based on the pet category feature value and the preset category catalog. If so, the fragrance parameters that do not meet the conditions are eliminated, and the average value of the probabilities corresponding to the remaining fragrance parameters is calculated, and the category with the largest average value is used as the fragrance release parameter. Otherwise, the average value of the probabilities corresponding to all the fragrance parameters is calculated, and the category with the largest average value is used as the fragrance release parameter.

[0038] As another preferred embodiment, the present invention also provides a pet fragrance box, which is set at any place in the car, and the pet fragrance box is communicatively connected to a computer-readable storage medium, and a computer program is stored on the computer-readable storage medium; the computer program is controlled and executed by one or more processors to implement the vehicle-mounted fragrance control method as described above.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] Temperature sensors, humidity sensors, infrared sensors and photometric sensors are arranged on the vehicle body to collect environmental characteristic values, the vehicle-mounted microphone is used to collect the pet's voice, and the pet's voice characteristic data is identified by the sound recognition system, the vehicle-mounted camera is used to collect the pet's image, and the pet's type characteristic data is identified by the target detection system, and the above-mentioned characteristic data are input into the fragrance decision model, and the prediction result of the neural network output is controlled to control the fragrance box at the bottom of the vehicle body to start the targeted release of fragrance of corresponding concentration and type to the pet. The present invention can release the pet concentration and control the air concentration range according to relevant control parameters, so as to ensure that the fragrance gas is always kept in the pet area, does not conflict with human use, and meets the needs of humans and pets at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 The present invention is a flow chart of a vehicle-mounted fragrance control method.

[0042] Figure 2 Schematic diagram of fragrance release.

[0043] Figure 3 Schematic diagram of the control system of the present invention.

[0044] Figure 4 FIG. 4 is a schematic diagram of a preferred embodiment of the pet fragrance box of the present invention. DETAILED DESCRIPTION

[0045] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. The drawings show preferred embodiments of the present invention. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein.

[0046] Please refer to Figure 1 The present invention provides a vehicle-mounted fragrance control method, characterized in that the method specifically comprises:

[0047] S100: When the vehicle is started, characteristic data of the pet in the vehicle is obtained.

[0048] S200: Processing the characteristic data using a fragrance decision model to calculate fragrance release parameters.

[0049] S300: Adjusting the type and concentration of the fragrance in the vehicle according to the fragrance release parameter.

[0050] During the specific implementation process, pet fragrance boxes are arranged in the car according to needs. The pet fragrance boxes are equipped with common animal fragrance essential oils, concentration control chips, and area control chips. In addition, temperature sensors, humidity sensors, infrared sensors, and photometric sensors are arranged on the car body to collect environmental characteristic values. The car-mounted microphone is used to collect the pet's voice, and the pet's voice characteristic data is identified through the voice recognition system. The car-mounted camera is used to collect the pet's image and the pet's type characteristic data is identified through the target detection system. The above-mentioned characteristic data are input into the fragrance decision model, and the predicted results are output according to the neural network, so as to control the fragrance box at the bottom of the car body to start the targeted release of fragrance of corresponding concentration and type to the pet.

[0051] In this embodiment, the step of obtaining characteristic data of the pet in the vehicle at least includes:

[0052] Environmental characteristic values ​​obtained based on the temperature, humidity, infrared and light intensity inside the car.

[0053] The pet emotion feature value is obtained based on the collected pet sound spectrum in the car.

[0054] The pet category feature value is obtained based on the collected pet image in the car.

[0055] During the specific implementation process, the temperature, humidity, infrared and light intensity in the car are obtained through temperature sensors, humidity sensors, infrared sensors and light intensity sensors respectively, which can ensure the real-time acquisition of the environmental characteristic values ​​in the car, provide accuracy and efficiency in controlling the pet fragrance box, and input the spectral noise of the pet sound in the car into the trained neural network to obtain the pet emotion characteristic value; the image of the pet in the car is input into the trained neural network to obtain the pet category characteristic value.

[0056] The pet emotion feature values ​​obtained based on the frequency spectrum of the pet sound in the car specifically include:

[0057] The pet sound is obtained through the vehicle microphone, the amplitude of the pet sound waveform is normalized, the waveform envelope and the amplitude average of the normalized waveform are calculated, all the maximum values ​​of the waveform envelope are obtained, the peak value is determined according to the maximum value, the amplitude average value and the preset time, and the waveform is divided into multiple segments according to the peak value; the segment average value is calculated, if the segment average value is greater than the average value of the left and right adjacent segments, the segment is marked as a segment to be identified, otherwise it is marked as a noise segment; the noise segments on the left and right adjacent sides of the segment to be identified are used to reduce the noise of the segment to be identified; the frequency spectrum of each segment to be identified after noise reduction is input into the trained neural network to obtain the category and the probability corresponding to the category, and the pet emotion characteristic value is obtained according to the category of each segment to be identified, the probability and the time interval between the segments to be identified.

[0058] The pet category feature values ​​obtained based on the image of the pet in the car specifically include:

[0059] The pet image is acquired through a vehicle-mounted camera, the pet face is located based on a target detection network, and the basic parts of the pet face are located using a feature point detection algorithm to extract the pet's facial image features, and the facial image features are input into a pet recognition model based on a neural network to obtain the pet category and the corresponding probability of the pet category, and the pet type feature data is obtained according to the object category and the corresponding probability of the pet category.

[0060] In this embodiment, the step S200 specifically includes:

[0061] S201: performing regularization and normalization processing on the feature data.

[0062] S202: Classify the processed feature data to obtain fragrance parameters and probabilities corresponding to each feature data; the fragrance parameters include at least fragrance concentration and fragrance type; the probabilities include at least the probability corresponding to the fragrance concentration and the probability corresponding to the fragrance type.

[0063] S203: Determine whether there are fragrance parameters that do not match the current pet type based on the pet category feature value and the preset category catalog; if so, remove the fragrance parameters that do not match the conditions, calculate the average value of the probabilities corresponding to the remaining fragrance parameters, and use the category with the largest average value as the fragrance release parameter; otherwise, calculate the average value of the probabilities corresponding to all the fragrance parameters, and use the category with the largest average value as the fragrance release parameter.

[0064] Among them, the fragrance decision model is a neural network, which is a type of supervised learning. The characteristic of supervised learning is that each sample in the training sample set has a corresponding label, that is, each sample has its own corresponding category. The result obtained by the activation function of the neural network is actually the probability of which category the feature data belongs to. For example, the training sample set has two categories: fragrance concentration and fragrance type. After the feature data is input into the neural network, the result category and category corresponding probability are the probability of a certain fragrance type and the probability of a certain concentration, respectively.

[0065] For example, the training sample set has two categories: fragrance concentration and fragrance type. After inputting each feature data into the neural network, the following situation is obtained:

[0066] The probability of the fragrance concentration corresponding to the environmental characteristic value is 0.9, and the corresponding probability of calming fragrance is 0.5, the probability of soothing fragrance is 0.3, and the probability of relaxing fragrance is 0.3;

[0067] The probability of the pet's emotion characteristic value corresponding to the fragrance concentration is 0.8, and the corresponding probability of calming fragrance is 0.6, the probability of soothing fragrance is 0.2, and the probability of relaxing fragrance is 0.2;

[0068] The probability of the fragrance concentration corresponding to the pet category feature value is 0.7, and the corresponding probability of calming fragrance is 0.3, the probability of soothing fragrance is 0.6, and the probability of relaxing fragrance is 0.4;

[0069] It should be understood that a higher probability of fragrance concentration represents a higher fragrance concentration, that is, probability corresponds to concentration.

[0070] Furthermore, the average values ​​of the new probability values ​​of the same classification of these three feature data are calculated as follows: the probability of fragrance concentration is 0.8, the probability of calming fragrance is 0.47, the probability of soothing fragrance is 0.37, and the probability of relaxing fragrance is 0.30. The classifications with the largest average values ​​are fragrance concentration of 0.8 and calming fragrance.

[0071] In this embodiment, the preset category directory includes pet types and fragrance parameters associated with the pet types.

[0072] During the specific implementation process, different pet types are pre-associated with corresponding fragrance concentrations and fragrance types and stored, and individual pet characteristics are pre-set, such as body shape, breed, age, preferences and health status. Corresponding associations between fragrance concentrations and fragrance types corresponding to individual pet characteristics are established accordingly, which can more accurately control the pet fragrance box to release suitable fragrance.

[0073] For example, if it is an everyday pet, it will be marked as a regular customer as an input end of the multimodal system; in addition, other environmental sensors on the vehicle body will also pass the collected environmental features into the fragrance decision model; the fragrance decision model outputs control parameters to the fragrance control according to the habit category, and then controls the pet fragrance box on the vehicle body to release the corresponding concentration and type of fragrance.

[0074] When calculating the average value of the new probability values ​​of the same classification of the three feature data, the fragrance parameters that do not match the current pet type will be excluded based on the preset pet individual characteristics.

[0075] In this embodiment, the fragrance decision model is any one of a convolutional neural network, a deep neural network and a BP neural network.

[0076] In this embodiment, the step S300 specifically includes:

[0077] S301: defining the fragrance release parameter as a working parameter of a fragrance regulator.

[0078] S302: adjusting the type and concentration of the in-vehicle fragrance according to the working parameters, and emitting a fragrance matching the fragrance release parameters.

[0079] In the above technical solution, the fragrance box releases the pet concentration and controls the air concentration range according to relevant control parameters to ensure that the fragrance gas always remains in the pet area and does not conflict with human use.

[0080] For example, please refer to Figure 2 The smiley face represents a pet, and the concentration and range are controlled in the pet activity area at a height of 0-50cm at the bottom of the cabin.

[0081] In the above technical solution, the fragrance box releases the pet concentration and controls the air concentration range according to relevant control parameters to ensure that the fragrance gas always remains in the pet area and does not conflict with human use.

[0082] Please refer to Figure 3 As another preferred embodiment, the present invention further provides a pet fragrance box control system, the system at least comprising:

[0083] The acquisition module is used to acquire the characteristic data of the pet in the vehicle when the vehicle is started, and send the characteristic data to the calculation module.

[0084] The calculation module is used to construct a fragrance decision model to process the characteristic data, calculate the fragrance release parameters, and send the adjustment parameters to the control module.

[0085] The control module is used to adjust the type and concentration of the fragrance in the vehicle according to the fragrance release parameters.

[0086] In this embodiment, the acquisition module at least includes:

[0087] The sensor module is used to obtain environmental characteristic values ​​according to the temperature, humidity, infrared and light intensity in the vehicle; the sensor module is provided with at least a temperature sensor, a humidity sensor, an infrared sensor and a light intensity sensor.

[0088] The sound recognition module is used to obtain the pet emotion feature value based on the frequency spectrum of the pet sound in the car.

[0089] The object detection module is used to obtain pet category feature values ​​based on the image of the pet in the car.

[0090] In this embodiment, the computing module is embedded with a fragrance decision model, and the fragrance decision model at least includes:

[0091] The input layer is used to normalize and regularize the feature data, and send the processed feature data to the logical discrimination layer.

[0092] The logic discrimination layer is used to classify the processed feature data to obtain the corresponding fragrance parameters and corresponding probabilities of each feature data; the fragrance parameters include at least fragrance concentration and fragrance type; the probability includes at least the probability corresponding to the fragrance concentration and the probability corresponding to the fragrance type.

[0093] The judgment and calculation layer is used to judge whether there are fragrance parameters that do not meet the current pet type based on the pet category feature value and the preset category catalog. If so, the fragrance parameters that do not meet the conditions are eliminated, and the average value of the probabilities corresponding to the remaining fragrance parameters is calculated, and the category with the largest average value is used as the fragrance release parameter. Otherwise, the average value of the probabilities corresponding to all the fragrance parameters is calculated, and the category with the largest average value is used as the fragrance release parameter.

[0094] Among them, the fragrance decision-making model is trained through massive data of a large amount of pet olfactory behavior. The data of the operation process will also be updated to the cloud in a timely manner to help the model evolve.

[0095] It should be understood that the implementation process of the control module is as follows:

[0096] According to the adjusted fragrance parameters, the parameters of the fragrance regulator in the pet fragrance box are determined; according to the parameters of the fragrance regulator in the pet fragrance box, the fragrance regulator is controlled to adjust the pet fragrance box so that a fragrance matching the adjusted fragrance parameters is emitted through the pet fragrance box. Specifically, there is a fragrance control chip in the fragrance box, which will release fragrance gas by heating the corresponding fragrance essential oil according to the input parameters such as the fragrance type and concentration direction, control the control range of the air outlet direction, and control the concentration by temperature and heating time.

[0097] As another preferred embodiment, the present invention also provides a pet fragrance box, which is set at any place in the car, and the pet fragrance box is communicatively connected to a computer-readable storage medium, and a computer program is stored on the computer-readable storage medium; the computer program is controlled and executed by one or more processors to implement the vehicle-mounted fragrance control method as described above.

[0098] For example, please refer to Figure 4 Four pet fragrance boxes are arranged in the cockpit to build a vehicle-mounted fragrance control system, namely left pet fragrance box 1, left pet fragrance box 2, right pet fragrance box 1 and right pet fragrance box 2, with built-in common animal fragrance essential oils, concentration control chips, and area control chips; temperature sensors, humidity sensors, infrared sensors, and photometric sensors are arranged on the vehicle body to collect environmental characteristic values, and the voice of the pet is collected by the vehicle-mounted microphone, and the pet voice characteristic data is identified by the voice recognition system; the image of the pet is collected by the vehicle-mounted camera, and the type characteristic data of the pet is identified by the target detection system, and the above-mentioned characteristic data are input into the fragrance decision model, and the predicted result of the neural network output is controlled to control the fragrance box at the bottom of the vehicle body to start the directional release of the fragrance of the corresponding concentration and type to the pet. The present invention can release the pet concentration and control the air concentration range according to the relevant control parameters, so as to ensure that the fragrance gas is always kept in the pet area, does not conflict with human use, and meets the needs of humans and pets at the same time.

[0099] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present invention thereto. Various changes and modifications may be made therein by one of ordinary skill in the art without departing from the scope and spirit of the present invention. All such changes and modifications are intended to be included within the scope of the present invention as required by the appended claims.

[0100] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0101] The various system and method embodiments of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all of the functions of some modules according to embodiments of the present invention. The present invention can also be implemented as a system program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0102] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, for example, the division of functions is only a logical function division, and there may be other division methods in actual implementation, such as multiple tools or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0103] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0104] Although the present invention is described in conjunction with the above specific embodiments, it is obvious that those skilled in the art can make many substitutions, modifications and changes according to the above content. Therefore, all such substitutions, improvements and changes are included in the spirit and scope of the appended claims.

Claims

1. A vehicle-mounted fragrance control method, characterized in that: The method specifically comprises: S100: When the vehicle is started, characteristic data of the pet in the vehicle is obtained; S200: Processing the characteristic data using a fragrance decision model to calculate fragrance release parameters; S300: adjusting the type and concentration of the fragrance in the vehicle according to the fragrance release parameter; Wherein, the obtaining of characteristic data of the pet in the vehicle at least includes: Environmental characteristic values ​​obtained based on the temperature, humidity, infrared rays and light intensity inside the vehicle; A pet emotion feature value obtained based on the collected pet sound spectrum in the car; A pet category feature value obtained based on the collected pet image in the car; The step S200 specifically includes: S201: performing regularization and normalization processing on the feature data; S202: Classifying the processed feature data to obtain fragrance parameters and probabilities corresponding to each feature data; the fragrance parameters at least include fragrance concentration and fragrance type; the probabilities at least include the probability corresponding to the fragrance concentration and the probability corresponding to the fragrance type; S203: judging whether there are fragrance parameters that do not match the current pet type according to the pet category feature value and the preset category catalog; if so, eliminating the fragrance parameters that do not match the condition, calculating the average value of the probabilities corresponding to the remaining fragrance parameters, and taking the category with the largest average value as the fragrance release parameter; otherwise, calculating the average value of the probabilities corresponding to all the fragrance parameters, and taking the category with the largest average value as the fragrance release parameter; The preset category directory includes pet types and Mutually associated fragrance parameters; The step S300 specifically includes: S301: Using the fragrance release parameter as a working parameter of a fragrance regulator; S302: adjusting the type and concentration of the in-vehicle fragrance according to the working parameters, and emitting a fragrance matching the fragrance release parameters.

2. The vehicle-mounted fragrance control method according to claim 1, characterized in that: The fragrance decision model is any one of a convolutional neural network, a deep neural network and a BP neural network.

3. A control system, using the vehicle-mounted fragrance control method according to any one of claims 1-2, characterized in that: The control system at least comprises: An acquisition module, used for acquiring characteristic data of a pet in the vehicle when the vehicle is started, and sending the characteristic data to a calculation module; A calculation module is used to construct a fragrance decision model to process the characteristic data, calculate the fragrance release parameters, and send the adjustment parameters to the control module; The control module is used to adjust the type and concentration of the fragrance in the vehicle according to the fragrance release parameters.

4. The control system according to claim 3, characterized in that: The acquisition module at least includes: A sensor module, used to obtain environmental characteristic values ​​according to the temperature, humidity, infrared and light intensity in the vehicle; the sensor module is provided with at least a temperature sensor, a humidity sensor, an infrared sensor and a light intensity sensor; A sound recognition module is used to obtain the pet emotion feature value based on the frequency spectrum of the pet sound in the car; The object detection module is used to obtain pet category feature values ​​based on the image of the pet in the car.

5. The control system according to claim 4, characterized in that: The fragrance decision model at least includes: The input layer is used to perform normalization and regularization on the feature data and send the processed feature data to the logical discrimination layer; A logic discrimination layer, for classifying the processed feature data to obtain the corresponding fragrance parameters and corresponding probabilities of each feature data; the fragrance parameters at least include fragrance concentration and fragrance type; the probability at least includes the probability corresponding to the fragrance concentration and the probability corresponding to the fragrance type; The judgment calculation layer is used to judge whether there are fragrance parameters that do not match the current pet type based on the pet category feature value and the preset category catalog. If so, the fragrance parameters that do not meet the conditions are eliminated, and the average value of the probabilities corresponding to the remaining fragrance parameters is calculated, and the category with the largest average value is used as the fragrance release parameter; otherwise, the average value of the probabilities corresponding to all the fragrance parameters is calculated, and the category with the largest average value is used as the fragrance release parameter.

6. A pet fragrance box, characterized in that: The pet fragrance box is set anywhere in the car, and the pet fragrance box is communicatively connected to a computer-readable storage medium, on which a computer program is stored; the computer program is controlled and executed by one or more processors to implement the vehicle-mounted fragrance control method as described in any one of claims 1-2.

Images