Clothes identification and comfort linkage control method and system based on YOLOv8

Through the YOLOv8 model, the clothing type and external environmental conditions are identified in real time, and the control command is generated to automatically adjust the air conditioning temperature, solving the problem of the lack of real-time response of the existing air conditioning control methods and achieving more accurate satisfaction with the comfort of drivers and passengers.

CN118977538APending Publication Date: 2024-11-19SHANGHAI PUFAFEN ELECTRONIC TECH CO LTD
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202411372055.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-29
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The existing air conditioning control methods lack real-time response when adjusting the comfort of drivers and passengers. They rely on preset compensation temperatures and ambient temperatures and cannot be dynamically adjusted to meet the real-time needs of members, which affects the safety and comfort of drivers.

Method used

The clothing recognition technology based on the YOLOv8 model is adopted to identify the type of clothing and external environmental conditions of the driver and passengers in real time, and generate control instructions to automatically adjust the air conditioning temperature to improve the comfort of the driver and passengers.

Benefits of technology

It achieves accurate satisfaction of the comfort of drivers and passengers, and dynamically adjusts the air conditioning settings by capturing clothing information and weather conditions in real time, improving the accuracy and efficiency of automatic adjustment of air conditioners.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118977538A_ABST
    Figure CN118977538A_ABST
Patent Text Reader

Abstract

The invention provides a clothes identification and comfort linkage control method and system based on YOLOv8, and relates to the technical field of air conditioner control, and the method comprises the steps: S1, based on a YOLOv8 model, carrying out the learning prediction of the clothes type of a person in a preset environment, and optimizing the model; and S2, deploying the optimized YOLOv8 model in a preset environment, and automatically adjusting the temperature of the air conditioner by recognizing the change of clothes of people in the environment in real time, so that the comfort of the people is improved. According to the method, the clothes of people can be classified and recognized, the recognition result is linked with the environment temperature, and the requirement for improving the comfort of a user through automatic adjustment of the air conditioner is met.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning control, and in particular to a clothing recognition and comfort linkage control method and system based on YOLOv8. Background Art

[0002] In the car cockpit system, it is very important to improve the comfort of the driver and passengers. The comfort of the driver and passengers is affected by the external environment temperature and the type and thickness of clothing.

[0003] At the same outside temperature, people wearing heavy clothes may feel overheated, while people wearing thin clothes may feel too cold. Therefore, real-time identification of the clothing type of drivers and passengers is important for automatic adjustment of the air conditioner. However, currently, the improvement of driver and passenger comfort through clothing changes mainly relies on the driver and passengers manually adjusting the air conditioning temperature, which can affect driver safety.

[0004] The invention patent with the existing publication number CN118182058A discloses a vehicle air conditioning control method, device, vehicle and storage medium, including: obtaining clothing type, indoor temperature and ambient temperature; clothing type is the type of clothing worn by the occupant, indoor temperature is the temperature in the cab, and ambient temperature is the temperature of the environment outside the cab; determining the compensation temperature according to the clothing type; determining the target temperature according to the indoor temperature, ambient temperature and compensation temperature; controlling the vehicle air conditioning to blow out the wind of the target temperature. By comprehensively referring to environmental factors (indoor temperature and ambient temperature) and the influence of the clothing worn by the occupant on the occupant's body temperature, the appropriate target temperature is determined, and then the vehicle air conditioning is controlled to blow out the wind of the target temperature, and the temperature in the cab is adjusted to the target temperature, realizing the automatic adjustment of the temperature in the cab, without the need for frequent manual operation by the occupant. The patent has the following defects: the invention adjusts the air conditioning in the car based on the clothing type and ambient temperature through the preset compensation temperature, which is a relatively static control method, does not consider the real-time needs of the members, and relies on the pre-defined compensation temperature and ambient temperature for calculation. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a clothing recognition and comfort linkage control method and system based on YOLOv8.

[0006] According to a YOLOv8-based clothing recognition and comfort linkage control method and system provided by the present invention, the scheme is as follows:

[0007] In a first aspect, a clothing recognition and comfort linkage control method based on YOLOv8 is provided, the method comprising:

[0008] Step S1: Based on the YOLOv8 model, learn and predict the clothing categories of people in the preset environment and optimize the model;

[0009] Step S2: Deploy the optimized YOLOv8 model in a preset environment, and automatically adjust the air-conditioning temperature to improve the comfort of personnel by real-time identification of changes in clothing of personnel in the environment.

[0010] Preferably, the step S1 comprises:

[0011] Step S1.1: Collecting data required for training, including clothing image data of people in a preset environment, external environment temperature, body temperature, time and weather conditions;

[0012] Step S1.2: Use EiSeg segmentation and annotation software to annotate the clothing of the person in the image, where the clothing classification is divided into short sleeves, thin long sleeves and thick long sleeves; where summer or spring and autumn long sleeves and coats are defined as thin long sleeves, and winter long sleeves and coats are defined as thick long sleeves;

[0013] Step S1.3: Convert the obtained annotation file format into label data in YOLO format that can be recognized by the YOLOv8 model;

[0014] Step S1.4: Classify the data into image data and label data, and randomly divide them into training set, validation set, and test set according to the proportion;

[0015] Step S1.5: Use the YOLOv8 model to train the training set, use the validation set to verify the training results of each round, save the best weights, load the model using the best weight file, and use the test set to evaluate the performance;

[0016] Step S1.6: Optimize the model by gradually adjusting the YOLOv8 model parameters and data enhancement techniques.

[0017] Preferably, the step S1.1 includes: collecting personnel video information through a camera in a preset environment, and then disassembling the video into image information; crawling the external environment temperature, body temperature, time and weather conditions at each moment through the network.

[0018] Preferably, when the preset environment position is not fixed, the computer obtains the latitude and longitude information of the real-time position of the preset environment through GPS, and uses the latitude and longitude information and calls the weather API interface through the Internet to obtain the external environment temperature, body temperature, time and weather conditions of the current location.

[0019] Preferably, the step S1.4 includes: the image data and the corresponding label data are a set of data, these data are randomly shuffled by code, and the data are divided into 80% training set, 10% verification set and 10% test set.

[0020] Preferably, the step S1.5 comprises: training the YOLOv8 model using the training set, and then using the validation set to evaluate the training results of each round through evaluation indicators including loss value, recall rate, and precision, so as to determine the performance of the current model;

[0021] If the performance of the current round of model on the validation set is better than the previous best model, the weight of the current model is saved as the best weight. This step is repeated recursively until the end of training, and the final best weight can be saved. Finally, the final best weight model is used to predict and evaluate the performance of the test set.

[0022] Preferably, step S2 comprises:

[0023] Step S2.1: deploy the optimized model;

[0024] Step S2.2: deploy cameras in the preset environment to capture the video stream of people's clothing information in real time; crawl the external environment temperature, body temperature, time and weather conditions at every moment on the network end;

[0025] Step S2.3: Transmitting the captured video stream and external environment temperature, body temperature, time, and weather condition data to a data processing unit;

[0026] Step S2.4: The data processing unit loads the optimized YOLOv8 model, inputs the acquired video stream into the YOLOv8 model for prediction, and outputs the category of the person's clothing as the result;

[0027] Step S2.5: Generate control instructions according to the recognition results;

[0028] Step S2.6: After receiving the control instruction, the control module in the system adjusts the air-conditioning temperature.

[0029] Preferably, generating a control instruction according to the recognition result in step S2.5 includes:

[0030] By collecting the current external environment temperature, body temperature, time and weather conditions;

[0031] If the climate is hot, if it is recognized that the person is wearing thick long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing thin long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing short sleeves, then a temperature control command is generated to raise the air conditioner temperature;

[0032] If the climate is cold and it is recognized that the person is wearing thick long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing thin long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing short sleeves, a temperature control command is generated to lower the air conditioning temperature.

[0033] In a second aspect, a clothing recognition and comfort linkage control system based on YOLOv8 is provided, the system comprising:

[0034] Module M1: Based on the YOLOv8 model, learn and predict the clothing categories of people in the preset environment and optimize the model;

[0035] Module M2: Deploy the optimized YOLOv8 model in a preset environment, automatically adjust the air-conditioning temperature by real-time identification of changes in people's clothing in the environment, and improve people's comfort.

[0036] Preferably, the module M1 comprises:

[0037] Module M1.1: Collect the data required for training, including clothing image data of people in the preset environment, external environment temperature, body temperature, time and weather conditions;

[0038] Module M1.2: Use EiSeg segmentation and annotation software to annotate the clothing of people in the image, where clothing is classified into short sleeves, thin long sleeves, and thick long sleeves; summer or spring and autumn long sleeves and coats are defined as thin long sleeves, and winter long sleeves and coats are defined as thick long sleeves;

[0039] Module M1.3: Convert the obtained annotation file format into label data in YOLO format that can be recognized by the YOLOv8 model;

[0040] Module M1.4: Classify the data into image data and label data, and randomly divide the data into training set, validation set, and test set in proportion;

[0041] Module M1.5: Use the YOLOv8 model to train the training set, use the validation set to verify the training results of each round, save the best weights, load the model using the best weights file, and use the test set to evaluate performance;

[0042] Module M1.6: Optimize the model by gradually adjusting YOLOv8 model parameters and data enhancement technology;

[0043] The module M1.1 includes: collecting personnel video information through the camera in the preset environment, and then disassembling the video into image information; crawling the external environment temperature, body temperature, time and weather conditions at every moment through the network;

[0044] When the preset environment position is not fixed, the computer obtains the latitude and longitude information of the real-time position of the preset environment through GPS, and uses the latitude and longitude information and the weather API interface through the Internet to obtain the external environment temperature, body temperature, time and weather conditions of the current position;

[0045] The module M1.4 includes: image data and corresponding label data are a set of data, these data are randomly shuffled by code, and the data are divided into 80% training set, 10% verification set and 10% test set;

[0046] The module M1.5 includes: the YOLOv8 model is trained using the training set, and then the validation set is used to evaluate the results of each round of training through evaluation indicators including loss value, recall rate, and precision, so as to determine the performance of the current model;

[0047] If the performance of the current round of model on the validation set is better than the previous best model, the weight of the current model is saved as the best weight. Repeat this step recursively until the training is completed, and it can be saved as the final best weight. Finally, the final best weight model is used to predict and evaluate the performance of the test set.

[0048] The module M2 comprises:

[0049] Module M2.1: Deploy the optimized model;

[0050] Module M2.2: Deploy cameras in the preset environment to capture the video stream of people's clothing information in real time; crawl the external environment temperature, body temperature, time and weather conditions at every moment on the network end;

[0051] Module M2.3: Transmit the captured video stream and external environment temperature, body temperature, time, and weather condition data to the data processing unit;

[0052] Module M2.4: The data processing unit loads the optimized YOLOv8 model, inputs the acquired video stream into the YOLOv8 model for prediction, and outputs the category of the person's clothing as the result;

[0053] Module M2.5: Generate control instructions based on recognition results;

[0054] Module M2.6: After receiving the control command, the control module in the system adjusts the air conditioning temperature;

[0055] Particularly preferably, the module M2.1 converts the optimized model into a TensorRT engine, and uses the 8155 chip and development tool chain to integrate the converted TensorRT model into an embedded application, and finally deploys the integrated application to the vehicle end equipped with the 8155 chip;

[0056] The module M2.5 generates control instructions according to the recognition results, including:

[0057] By collecting the current external environment temperature, body temperature, time and weather conditions;

[0058] If the climate is hot, if it is recognized that the person is wearing thick long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing thin long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing short sleeves, then a temperature control command is generated to raise the air conditioner temperature;

[0059] If the climate is cold and it is recognized that the person is wearing thick long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing thin long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing short sleeves, a temperature control command is generated to lower the air conditioning temperature.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The present invention is based on the YOLOv8 model and chip, implements the capture and analysis of clothing information, and dynamically adjusts the air-conditioning settings according to real-time information and weather conditions, thereby solving the need for automatic adjustment of the air-conditioning to improve user comfort and more accurately meeting the comfort requirements of passengers.

[0062] Other beneficial effects of the present invention will be explained in the specific implementation manner through the introduction of specific technical features and technical solutions. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0064] Figure 1 It is a schematic diagram of the overall process of the present invention. DETAILED DESCRIPTION

[0065] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0066] The embodiment of the present invention provides a clothing recognition and comfort linkage control method based on YOLOv8, which uses the YOLOv8 algorithm to identify the clothing changes of the driver and passengers in the car, and automatically adjusts the air conditioner according to the current temperature changes to improve the comfort of the driver and passengers. Figure 1 As shown, the method specifically includes:

[0067] Step S1: Based on the YOLOv8 model, learn and predict the categories of clothing of people in the preset environment (such as the in-car environment, in this case the drivers and passengers in the car) and optimize the model.

[0068] Specifically, step S1 includes:

[0069] Step S1.1: Collect the data required for training, including the clothing image data of the driver and passengers in the car, the ambient temperature outside the car, the body temperature, time and weather conditions.

[0070] Specifically, the in-car camera collects video information of the driver and passengers, and then breaks the video into image information; crawls the network to obtain the ambient temperature, body temperature, time and weather conditions outside the vehicle at every moment. When the vehicle is moving, the computer obtains the longitude and latitude information of the vehicle's real-time location through the GPS module, and uses the longitude and latitude information and calls the weather API interface through the Internet to obtain the ambient temperature, body temperature, time and weather conditions outside the vehicle at the current location.

[0071] Step S1.2: Use EiSeg segmentation and annotation software to annotate the clothing of the driver and passengers in the image. EISeg is an automatic annotation software developed based on interactive segmentation tools. The clothing is classified into short sleeves, thin long sleeves and thick long sleeves; thin long sleeves and thick long sleeves are manually defined, and long sleeves and coats in summer or spring and autumn are thin long sleeves, and thick long sleeves in winter.

[0072] Step S1.3: Convert the obtained annotation file format into label data in the YOLO format that can be recognized by the YOLOv8 model.

[0073] Step S1.4: Classify the data into image data and label data, and randomly divide them into training set, validation set, and test set in proportion.

[0074] Specifically, the image data and the corresponding label data are a set of data. After the data are randomly shuffled using the code, the data are divided into 80% training set, 10% validation set and 10% test set.

[0075] Step S1.5: Use the YOLOv8 model to train the training set, use the validation set to verify the training results of each round, save the best weights, load the model using the best weight file, and use the test set to evaluate the performance.

[0076] Specifically, the YOLOv8 model is trained using the training set, and then the validation set is used to evaluate the results of each round of training through evaluation indicators including loss value, recall rate, and precision (after each round of training, the model uses the current weights to infer all samples in the validation set, generates prediction results, and then compares these prediction results with the true labels of the validation set to calculate a series of evaluation indicators) to determine the performance of the current model.

[0077] If the performance of the current round of model on the validation set is better than the previous best model, the weight of the current model is saved as the best weight. This step is repeated recursively until the end of training, and the final best weight can be saved. Finally, the final best weight model is used to predict and evaluate the performance of the test set.

[0078] Step S1.6: Optimize the model by gradually adjusting the YOLOv8 model parameters and data enhancement techniques.

[0079] Step S2: Deploy the trained YOLOv8 model to the vehicle computer, and automatically adjust the air-conditioning temperature by real-time recognition of changes in the clothing of the driver and passengers in the vehicle, so as to improve the comfort of the driver and passengers.

[0080] Specifically, step S2 includes:

[0081] Step S2.1: Deploy the optimized model on the vehicle side. This includes converting the optimized model into a TensorRT engine, integrating the converted TensorRT model into embedded applications using the 8155 chip and development tool chain, and finally deploying the integrated application to the vehicle side equipped with the 8155 chip. TensorRT is particularly suitable for applications that require real-time or high-performance reasoning, such as autonomous driving, video analysis, etc. Train the model in other frameworks first, then optimize and deploy it using TensorRT to obtain the best performance and efficiency.

[0082] Step S2.2: Deploy a camera on the vehicle computer to capture the video stream of the driver and passengers' clothing information in real time; crawl the outdoor ambient temperature, body temperature, time and weather conditions at every moment on the network end.

[0083] Step S2.3: The captured video stream and external environment temperature, body temperature, time, and weather condition data are transmitted to a data processing unit.

[0084] Step S2.4: The data processing unit loads the optimized YOLOv8 model, inputs the acquired video stream into the YOLOv8 model for prediction, and outputs the category of the driver's clothing as the result.

[0085] Step S2.5: Generate control instructions according to the recognition results.

[0086] Specifically, it includes collecting the current ambient temperature, perceived temperature, time and weather conditions;

[0087] In the same season, thick long sleeves and thin long sleeves can be identified by material, texture, etc. Thick long sleeves are usually made of heavier materials (such as wool, wool, down, etc.), while thin long sleeves are made of lighter materials (such as cotton, linen, thin textiles, etc.).

[0088] If the climate is hot (ambient temperature is higher than 30°C), if the driver and passenger are wearing thick long sleeves, a control command to reduce the air-conditioning temperature by 1°C (the set temperature is reduced by 1°C at a time) is generated; if the driver and passenger are wearing thin long sleeves, a control command to reduce the air-conditioning temperature by 1°C is generated; if the driver and passenger are wearing short sleeves, a control command to increase the air-conditioning temperature by 1°C is generated;

[0089] If the climate is in low temperature (ambient temperature is lower than 15℃), if it is recognized that the driver and passengers are wearing thick long sleeves, a control command to increase the air-conditioning temperature by 1℃ is generated. If it is recognized that the driver and passengers are wearing thin long sleeves, a control command to increase the air-conditioning temperature by 1℃ is generated. If it is recognized that the driver and passengers are wearing short sleeves, a control command to lower the air-conditioning temperature by 1℃ is generated.

[0090] Step S2.6: After receiving the control command, the comfort control module in the system makes specific adjustments to the air-conditioning temperature.

[0091] The present invention also provides a clothing recognition and comfort linkage control system based on YOLOv8, which can be implemented by executing the process steps of the clothing recognition and comfort linkage control method based on YOLOv8, that is, those skilled in the art can understand the clothing recognition and comfort linkage control method based on YOLOv8 as a preferred implementation of the clothing recognition and comfort linkage control system based on YOLOv8. The system specifically includes the following contents:

[0092] Module M1: Based on the YOLOv8 model, learn and predict the clothing categories of the drivers and passengers in the car and optimize the model.

[0093] Specifically, the module M1 includes:

[0094] Module M1.1: Collect the data required for training, including the clothing image data of the driver and passengers in the car, the ambient temperature outside the car, the body temperature, time and weather conditions.

[0095] Specifically, the in-car camera collects video information of the driver and passengers, and then breaks the video into image information; crawls the network to obtain the ambient temperature, body temperature, time and weather conditions outside the vehicle at every moment. When the vehicle is moving, the computer obtains the longitude and latitude information of the vehicle's real-time location through the GPS module, and uses the longitude and latitude information and calls the weather API interface through the Internet to obtain the ambient temperature, body temperature, time and weather conditions outside the vehicle at the current location.

[0096] Module M1.2: Use EiSeg segmentation and annotation software to annotate the clothing of the driver and passengers in the image. EISeg is an automatic annotation software developed based on interactive segmentation tools. The clothing is classified into short sleeves, thin long sleeves and thick long sleeves; thin long sleeves and thick long sleeves are manually defined, and long sleeves and coats in summer or spring and autumn are thin long sleeves, and thick long sleeves in winter.

[0097] Module M1.3: Convert the obtained annotation file format into label data in YOLO format that can be recognized by the YOLOv8 model.

[0098] Module M1.4: Classify the data into image data and label data, and randomly divide the data into training set, validation set, and test set in proportion.

[0099] Specifically, the image data and the corresponding label data are a set of data. After the data are randomly shuffled using the code, the data are divided into 80% training set, 10% validation set and 10% test set.

[0100] Module M1.5: Use the YOLOv8 model to train the training set, use the validation set to verify the results of each round of training, save the best weights, load the model using the best weight file, and use the test set to evaluate performance.

[0101] Specifically, the YOLOv8 model is trained using the training set, and then the validation set is used to evaluate the results of each round of training through evaluation indicators including loss value, recall rate, and precision to determine the performance of the current model.

[0102] If the performance of the current round of model on the validation set is better than the previous best model, the weight of the current model is saved as the best weight. This step is repeated recursively until the end of training, and the final best weight can be saved. Finally, the final best weight model is used to predict and evaluate the performance of the test set.

[0103] Module M1.6: Optimize the model by gradually adjusting YOLOv8 model parameters and data enhancement technology.

[0104] Module M2: Deploy the trained YOLOv8 model to the vehicle computer, automatically adjust the air-conditioning temperature by real-time recognition of changes in the clothing of the driver and passengers in the vehicle, and improve the comfort of the driver and passengers.

[0105] Specifically, the module M2 includes:

[0106] Module M2.1: Deploy the optimized model on the vehicle side. This includes converting the optimized model into a TensorRT engine, integrating the converted TensorRT model into embedded applications using the 8155 chip and development tool chain, and finally deploying the integrated application to the vehicle side equipped with the 8155 chip. TensorRT is particularly suitable for applications that require real-time or high-performance reasoning, such as autonomous driving, video analysis, etc. Train the model in other frameworks first, then optimize and deploy it using TensorRT to obtain the best performance and efficiency.

[0107] Module M2.2: Deploy a camera on the vehicle computer to capture the video stream of the driver and passengers' clothing information in real time; crawl the outdoor ambient temperature, body temperature, time and weather conditions at every moment on the network end.

[0108] Module M2.3: Transmits the captured video stream and external environment temperature, body temperature, time, and weather condition data to the data processing unit.

[0109] Module M2.4: The data processing unit loads the optimized YOLOv8 model, inputs the acquired video stream into the YOLOv8 model for prediction, and outputs the category of the driver's clothing as the result.

[0110] Module M2.5: Generate control instructions based on recognition results. Specifically, it includes collecting the current ambient temperature, perceived temperature, time and weather conditions. If the climate is hot, it is recognized that the driver and passengers are wearing thick long sleeves, then a control command to reduce the air-conditioning temperature by 1°C is generated; if the driver and passengers are wearing thin long sleeves, then a control command to reduce the air-conditioning temperature by 1°C is generated; if the driver and passengers are wearing short sleeves, then a control command to increase the air-conditioning temperature by 1°C is generated; if the climate is cold, it is recognized that the driver and passengers are wearing thick long sleeves, then a control command to increase the air-conditioning temperature by 1°C is generated; if the driver and passengers are wearing thin long sleeves, then a control command to increase the air-conditioning temperature by 1°C is generated; if the driver and passengers are wearing short sleeves, then a control command to reduce the air-conditioning temperature by 1°C is generated.

[0111] Module M2.6: After receiving the control command, the comfort control module in the system makes specific adjustments to the air-conditioning temperature.

[0112] In the field of image vision, traditional clothing category classification methods mainly include methods based on feature extraction or shallow neural networks. In comparison, yolov8 has the advantages of high precision, high efficiency, and high robustness. It can respond to application scenarios in real time and provide more reliable support for smart air conditioners.

[0113] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0114] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A clothing recognition and comfort linkage control method based on YOLOv8, characterized in that: include: Step S1: Based on the YOLOv8 model, learn and predict the clothing categories of people in the preset environment and optimize the model; Step S2: Deploy the optimized YOLOv8 model in a preset environment, and automatically adjust the air-conditioning temperature to improve the comfort of personnel by real-time identification of changes in clothing of personnel in the environment.

2. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 1 is characterized in that: The step S1 comprises: Step S1.1: Collecting data required for training, including clothing image data of people in a preset environment, external environment temperature, body temperature, time and weather conditions; Step S1.2: Use EiSeg segmentation and annotation software to annotate the clothing of the person in the image, where the clothing classification is divided into short sleeves, thin long sleeves and thick long sleeves; where summer or spring and autumn long sleeves and coats are defined as thin long sleeves, and winter long sleeves and coats are defined as thick long sleeves; Step S1.3: Convert the obtained annotation file format into label data in YOLO format that can be recognized by the YOLOv8 model; Step S1.4: Classify the data into image data and label data, and randomly divide them into training set, validation set, and test set according to the proportion; Step S1.5: Use the YOLOv8 model to train the training set, use the validation set to verify the training results of each round, save the best weights, load the model using the best weight file, and use the test set to evaluate the performance; Step S1.6: Optimize the model by gradually adjusting the YOLOv8 model parameters and data enhancement techniques.

3. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 2 is characterized in that: The step S1.1 includes: collecting personnel video information through a camera in a preset environment, and then breaking the video into image information; crawling the external environment temperature, body temperature, time and weather conditions at each moment through the network.

4. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 3 is characterized in that: When the preset environment position is not fixed, the computer obtains the latitude and longitude information of the real-time position of the preset environment through GPS, and uses the latitude and longitude information and calls the weather API interface through the Internet to obtain the external environment temperature, body temperature, time and weather conditions of the current position.

5. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 2 is characterized in that: The step S1.4 includes: the image data and the corresponding label data are a set of data, these data are randomly shuffled by code, and the data are divided into 80% training set, 10% verification set and 10% test set.

6. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 2 is characterized in that: The step S1.5 includes: the YOLOv8 model is trained using the training set, and then the validation set is used to evaluate the results of each round of training through evaluation indicators including loss value, recall rate, and precision, so as to determine the performance of the current model; If the performance of the current round of model on the validation set is better than the previous best model, the weight of the current model is saved as the best weight. This step is repeated recursively until the end of training, and the final best weight can be saved. Finally, the final best weight model is used to predict and evaluate the performance of the test set.

7. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 1, characterized in that: The step S2 comprises: Step S2.1: deploy the optimized model; Step S2.2: deploy cameras in the preset environment to capture the video stream of people's clothing information in real time; crawl the external environment temperature, body temperature, time and weather conditions at every moment on the network end; Step S2.3: Transmitting the captured video stream and external environment temperature, body temperature, time, and weather condition data to a data processing unit; Step S2.4: The data processing unit loads the optimized YOLOv8 model, inputs the acquired video stream into the YOLOv8 model for prediction, and outputs the category of the person's clothing as the result; Step S2.5: Generate control instructions according to the recognition results; Step S2.6: After receiving the control instruction, the control module in the system adjusts the air-conditioning temperature.

8. The clothing recognition and comfort linkage control method based on YOLOv8 according to claim 1, characterized in that: The step S2.5 of generating a control instruction according to the recognition result includes: By collecting the current external environment temperature, body temperature, time and weather conditions; If the climate is hot, if it is recognized that the person is wearing thick long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing thin long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing short sleeves, then a temperature control command is generated to raise the air conditioner temperature; If the climate is cold and it is recognized that the person is wearing thick long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing thin long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing short sleeves, a temperature control command is generated to lower the air conditioning temperature.

9. A clothing recognition and comfort linkage control system based on YOLOv8, characterized in that: include: Module M1: Based on the YOLOv8 model, learn and predict the clothing categories of people in the preset environment and optimize the model; Module M2: Deploy the optimized YOLOv8 model in a preset environment, automatically adjust the air-conditioning temperature by real-time identification of changes in people's clothing in the environment, and improve people's comfort.

10. The clothing recognition and comfort linkage control system based on YOLOv8 according to claim 9, characterized in that: The module M1 comprises: Module M1.1: Collect the data required for training, including clothing image data of people in the preset environment, external environment temperature, body temperature, time and weather conditions; Module M1.2: Use EiSeg segmentation and annotation software to annotate the clothing of people in the image, where clothing is classified into short sleeves, thin long sleeves, and thick long sleeves; summer or spring and autumn long sleeves and coats are defined as thin long sleeves, and winter long sleeves and coats are defined as thick long sleeves; Module M1.3: Convert the obtained annotation file format into label data in YOLO format that can be recognized by the YOLOv8 model; Module M1.4: Classify the data into image data and label data, and randomly divide the data into training set, validation set, and test set in proportion; Module M1.5: Use the YOLOv8 model to train the training set, use the validation set to verify the training results of each round, save the best weights, load the model using the best weights file, and use the test set to evaluate performance; Module M1.6: Optimize the model by gradually adjusting YOLOv8 model parameters and data enhancement technology; The module M1.1 includes: collecting personnel video information through the camera in the preset environment, and then disassembling the video into image information; crawling the external environment temperature, body temperature, time and weather conditions at every moment through the network; When the preset environment position is not fixed, the computer obtains the latitude and longitude information of the real-time position of the preset environment through GPS, and uses the latitude and longitude information and the weather API interface through the Internet to obtain the external environment temperature, body temperature, time and weather conditions of the current position; The module M1.4 includes: image data and corresponding label data are a set of data, these data are randomly shuffled by code, and the data are divided into 80% training set, 10% verification set and 10% test set; The module M1.5 includes: the YOLOv8 model is trained using the training set, and then the validation set is used to evaluate the results of each round of training through evaluation indicators including loss value, recall rate, and precision, so as to determine the performance of the current model; If the performance of the current round of model on the validation set is better than the previous best model, the weight of the current model is saved as the best weight. Repeat this step recursively until the training is completed, and it can be saved as the final best weight. Finally, the final best weight model is used to predict and evaluate the performance of the test set. The module M2 comprises: Module M2.1: Deploy the optimized model; Module M2.2: Deploy cameras in the preset environment to capture the video stream of people's clothing information in real time; crawl the external environment temperature, body temperature, time and weather conditions at every moment on the network end; Module M2.3: Transmit the captured video stream and external environment temperature, body temperature, time, and weather condition data to the data processing unit; Module M2.4: The data processing unit loads the optimized YOLOv8 model, inputs the acquired video stream into the YOLOv8 model for prediction, and outputs the category of the person's clothing as the result; Module M2.5: Generate control instructions based on recognition results; Module M2.6: After receiving the control command, the control module in the system adjusts the air conditioning temperature; The module M2.5 generates control instructions according to the recognition results, including: By collecting the current external environment temperature, body temperature, time and weather conditions; If the climate is hot, if it is recognized that the person is wearing thick long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing thin long sleeves, then a temperature control command is generated to lower the air conditioner temperature; if it is recognized that the person is wearing short sleeves, then a temperature control command is generated to raise the air conditioner temperature; If the climate is cold and it is recognized that the person is wearing thick long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing thin long sleeves, a temperature control command is generated to increase the air conditioning temperature; if it is recognized that the driver and passengers are wearing short sleeves, a temperature control command is generated to lower the air conditioning temperature.

Citation Information

Cited By

  • Intelligent automobile seat temperature control system based on YOLOv5 algorithm

    CN119142230A

  • An intelligent temperature control system for car seats based on YOLOv5 algorithm

    CN119142230B

  • Vehicle-mounted air conditioner control method, vehicle and readable storage medium

    CN119974903A

  • On-vehicle air conditioner control method, vehicle and readable storage medium

    CN119974903B