Home remote monitoring and early warning system and method based on personalized threshold value

Abstract

The invention relates to the technical field of home remote monitoring and control, in particular to a home remote monitoring and early warning system and method based on a personalized threshold value. Obtaining first home data and second home data of the target; dividing the second home data according to the target behavior to obtain third home data; obtaining an area surface material according to the first home data; according to the area surface material, the first home data and the third home data, obtaining a home sub-area; obtaining a first personalized threshold value of the home sub-region, and obtaining first home early warning data in combination with the third home data; determining an early warning sub-region according to the first home early warning data; and obtaining sub-region data of the early warning sub-region cleaned by the second intelligent device, obtaining second home early warning data according to the third home data and the sub-region data, and adjusting and controlling the second intelligent device according to the second home early warning data. According to the invention, through the first home early warning data and the second home early warning data, early warning and cleaning of pet hair are accurately realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of home remote monitoring and control, in particular to a home remote monitoring and early warning system and method based on personalized threshold. BACKGROUND

[0002] Pets bring joy and companionship to families, but they also bring some unavoidable troubles. For example, pets often shed hair during daily activities, especially after the molting season or intense exercise, the hair loss is more serious. Hair adheres to furniture, floors, bed linen and other items, not only increasing the difficulty of cleaning, but also affecting air quality, leading to an increase in allergens in the home, and thus potentially threatening the health of family members and pets.

[0003] Although there are vacuum cleaners, roller brushes and other cleaning tools on the market that can help clean up, these traditional cleaning methods often have limited effectiveness in dealing with pet hair. On the one hand, it requires a lot of time and effort, and it is difficult to achieve comprehensive coverage; on the other hand, existing cleaning tools often cannot accurately identify the specific location of the hair loss, and often can only rely on the user's subjective judgment to clean up, resulting in poor cleaning results, and even some important areas may be overlooked.

[0004] With the advancement of technology, the rapid development of artificial intelligence and Internet of Things technology, intelligent home solutions have begun to enter the market. By utilizing advanced sensing technology, machine learning algorithms and image recognition technology, real-time monitoring of pet activity can be achieved, and the area of pet hair loss can be accurately identified. However, existing technical solutions mostly focus on automated cleaning or static environment monitoring, lacking in-depth understanding and real-time feedback of dynamic pet activity and hair loss process. How to achieve intelligent monitoring, accurate early warning and efficient cleaning based on pet behavior in the home environment has become a difficult problem that the industry needs to solve.

[0005] Therefore, a home remote monitoring and early warning system and method based on personalized threshold are proposed. SUMMARY

[0006] The application aims to provide a home remote monitoring and early warning system and method based on personalized threshold, which obtains first home data and second home data of a target, divides the second home data according to target behavior to obtain third home data, obtains regional surface material according to the first home data, obtains a home sub-region according to the regional surface material, the first home data and the third home data, obtains a first personalized threshold of the home sub-region, obtains first home early warning data in combination with the third home data, determines a warning sub-region according to the first home early warning data, obtains sub-region data of the warning sub-region cleaned by a second intelligent device, obtains second home early warning data according to the third home data and the sub-region data, and adjusts and controls the second intelligent device according to the second home early warning data.

[0007] To achieve the above object, the application provides the following technical scheme.

[0008] A home remote monitoring and early warning system based on personalized threshold comprises:

[0009] A first intelligent device control module controls the first intelligent device to monitor a target, and obtains first home data and second home data of the target, wherein the first home data is pet trajectory data, and the second home data is pet behavior data; the second home data is divided and recognized according to target behavior to obtain third home data, wherein the third home data comprises target hair loss data and target influence range data;

[0010] A first intelligent device early warning module obtains regional surface material according to the first home data; clusters and divides the regional surface material, the first home data and the third home data to obtain a home sub-region; obtains a first personalized threshold according to the regional surface material of the home sub-region, and obtains first home early warning data according to the first personalized threshold and the third home data;

[0011] A second intelligent device control module determines a warning sub-region according to the first home early warning data of the home sub-region, controls the second intelligent device to clean the warning sub-region, and adjusts parameters of the second intelligent device according to the first home early warning data.

[0012] A second intelligent device early warning module obtains sub-region data of the warning sub-region cleaned by the second intelligent device, obtains second home early warning data according to the third home data and the sub-region data, and adjusts and controls the second intelligent device according to the second home early warning data.

[0013] The process of obtaining the third home data in the first intelligent device control module is as follows:

[0014] acquire a target behavior and a target parameter, the target parameter including a target species, a target age, a target gender and a target health condition, the target health condition including a target physiological indicator, hair health data and skin health data, the physiological indicator including heart rate data, body temperature data and blood indicators, the hair health data including pet body hair data and pet dropped hair data, the skin health data including a skin abnormal area and a skin severity coefficient, the skin severity coefficient being determined according to the influence of the skin abnormal area on hair loss;

[0015] split the second home data according to the target behavior to obtain second home split data;

[0016] construct a target behavior recognition model to recognize the second home split data and the target parameter, and acquire target hair loss data and target influence range data.

[0017] The target behavior recognition model includes a hair loss data recognition layer and an influence range recognition layer, and the construction process is as follows:

[0018] identify the target historical behavior data to acquire target historical parameters, target historical behaviors, target historical hair loss data and target historical influence ranges;

[0019] combine the target historical parameters, the target historical behaviors, the target historical hair loss data and the target historical influence ranges to obtain first historical training data and second historical training data, the first historical training data including the target historical parameters, the target historical behaviors and the target historical hair loss data, and the second historical training data including the target historical parameters, the target historical behaviors and the target historical influence ranges;

[0020] train the hair loss data recognition layer according to the first historical training data, and train the influence range recognition layer according to the second historical training data.

[0021] The division process of the home sub-area is as follows:

[0022] obtain fusion home data according to the first home data and the third home data, the fusion home data including pet trajectory data, target hair loss data corresponding to the pet trajectory data and target influence ranges;

[0023] divide the fusion home data according to the area surface material to obtain single-material home data, the single-material home data including the fusion home data under the same area surface material;

[0024] cluster and divide the single-material home data according to a clustering algorithm to obtain home sub-areas.

[0025] The first personalized threshold reflects the difficulty of hair cleaning of the surface material of the region; the setting process is: establishing an experimental material sample according to the surface material of the region, uniformly placing pet hair on the experimental material sample, and selecting a second intelligent device to clean the experimental material sample, the cleaning process being consistent;

[0026] The first personalized threshold is obtained according to the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; the hair cleaning efficiency is determined according to the weight of the pet hair that is cleaned, and the damage coefficient of the experimental material is obtained according to the difference between the cleaned region image after cleaning and the standard image.

[0027] The initial weight before the second intelligent device cleans the early warning sub-region is obtained, and the final weight after the second intelligent device cleans the early warning sub-region is obtained;

[0028] The sub-region cleaning weight is obtained according to the initial weight and the final weight;

[0029] The impurity image of the impurity obtained by the second intelligent device cleaning the early warning sub-region is obtained, the distribution proportion of pet hair is obtained by identifying the impurity image, the hair cleaning data and non-hair cleaning data are obtained according to the sub-region cleaning weight and the distribution proportion of pet hair, and the sub-region data is obtained according to the hair cleaning data and the non-hair cleaning data;

[0030] The second home early warning data is obtained according to the hair cleaning data, the non-hair cleaning data and the target shedding data in the third home data; and the working parameters of the second intelligent device are adjusted according to the second home early warning data.

[0031] A home remote monitoring and early warning method based on a personalized threshold, comprising:

[0032] The first intelligent device is controlled to monitor the target, and the first home data and the second home data of the target are obtained; the first home data is pet trajectory data, and the second home data is pet behavior data; the second home data is divided and identified according to the target behavior, and the third home data is obtained, the third home data including target shedding data and target influence range data;

[0033] The surface material of the region is obtained according to the first home data; the home sub-region is obtained by clustering and dividing according to the surface material of the region, the first home data and the third home data; the first personalized threshold is set according to the surface material of the home sub-region, and the first home early warning data is obtained according to the first personalized threshold and the third home data;

[0034] The early warning sub-region is determined according to the first home early warning data of the home sub-region, the second intelligent device is controlled to clean the early warning sub-region, and the parameters of the second intelligent device are adjusted according to the first home early warning data.

[0035] Obtain the sub-area data of the second intelligent device for the early warning sub-area cleaning, obtain the second home early warning data according to the third home data and the sub-area data, and adjust and control the second intelligent device according to the second home early warning data.

[0036] The third home data is obtained by:

[0037] Obtain the target behavior and the target parameter, the target parameter includes the target category, the target age, the target gender and the target health condition, the target health condition includes the target physiological index, the hair health data and the skin health data;

[0038] According to the target behavior, the second home data is split to obtain the second home split data;

[0039] The target behavior recognition model is constructed to identify the second home split data and the target parameter, and obtain the target hair loss data and the target influence range data.

[0040] The target behavior recognition model includes a hair loss data identification layer and an influence range identification layer, and the construction process is:

[0041] The target historical behavior data is identified to obtain the target historical parameter, the target historical behavior, the target historical hair loss data and the target historical influence range;

[0042] The target historical parameter, the target historical behavior, the target historical hair loss data and the target historical influence range are combined to obtain the first historical training data and the second historical training data, the first historical training data includes the target historical parameter, the target historical behavior and the target historical hair loss data, and the second historical training data includes the target historical parameter, the target historical behavior and the target historical influence range;

[0043] The hair loss data identification layer is trained according to the first historical training data, and the influence range identification layer is trained according to the second historical training data.

[0044] Compared with the prior art, the beneficial effects of the present application are:

[0045] 1、The second home data is split according to the target behavior to obtain the second home split data, the target behavior recognition model is constructed to identify the second home split data and the target parameter, the hair loss data of the pet is obtained by identifying the second home split data and the target parameter according to the hair loss data identification layer, the influence range data of the pet is obtained by identifying the second home split data and the target parameter according to the influence range identification layer, and the hair loss condition is accurately identified according to the behavior data of the pet.

[0046] 2、The application obtains fusion home data according to the first home data and the third home data; the fusion home data includes pet trajectory data, target hair loss data corresponding to the pet trajectory data and a target influence range; the fusion home data is divided according to a regional surface material to obtain single-material home data; the fusion home data under the same regional surface material is accurately obtained; and the single-material home data is clustered and divided by a clustering algorithm to accurately divide the area affected by pet hair.

[0047] 3、The application establishes an experimental material sample according to a regional surface material, uniformly places pet hair on the experimental material sample, selects a second intelligent device to clean the experimental material sample, obtains a first personalized threshold according to a hair cleaning efficiency of the experimental material sample and a damage coefficient of the experimental material, and accurately measures the cleaning urgency of the regional surface material according to the first personalized threshold and the third home data. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 Fig. 1 is a structural schematic diagram of a home remote monitoring and early warning system based on a personalized threshold according to the application;

[0049] Figure 2 Fig. 2 is a flowchart of a home remote monitoring and early warning method based on a personalized threshold according to the application;

[0050] Figure 3 Fig. 3 is a carpet diagram according to the application. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the application will be apparently and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0052] Embodiment one

[0053] The application provides a home remote monitoring and early warning system based on a personalized threshold, as shown in Fig. 1, which comprises a first intelligent device control module, a first intelligent device early warning module, a second intelligent device control module and a second intelligent device early warning module. Figure 1

[0054] ​The first smart device control module controls the first smart device to monitor the target and obtain first home data and second home data of the target; the first home data is pet trajectory data, and the second home data is pet behavior data; the second home data is divided and identified according to the target behavior to obtain third home data, and the third home data includes target hair loss data and target influence range data.

[0055] The process of obtaining the third home data in the first smart device control module is as follows:

[0056] Obtain target behavior and target parameters, where the target parameters include target type, target age, target gender, and target health status. The target health status includes target physiological indicators, hair health data, and skin health data. The physiological indicators include heart rate data, body temperature data, blood indicators, etc. The hair health data includes hair data on the pet and hair loss data of the pet. The skin health data includes abnormal skin areas and a skin severity coefficient. The skin severity coefficient is determined based on the impact of abnormal skin areas on hair loss.

[0057] Splitting the second home data according to the target behavior to obtain the second home split data;

[0058] A target behavior recognition model is constructed to identify the second home splitting data and target parameters, and obtain target hair loss data and target impact range data.

[0059] The target behavior recognition model includes a hair loss data recognition layer and an impact range recognition layer, and its construction process is as follows:

[0060] Identify the target's historical behavior data to obtain the target's historical parameters, target's historical behavior, target's historical hair loss data, and target's historical impact range;

[0061] Combining target historical parameters, target historical behaviors, target historical hair loss data, and target historical influence range to obtain first historical training data and second historical training data; the first historical training data includes target historical parameters, target historical behaviors, and target historical hair loss data; the second historical training data includes target historical parameters, target historical behaviors, and target historical influence range;

[0062] The hair loss data recognition layer is obtained by training according to the first historical training data; and the impact range recognition layer is obtained by training according to the second historical training data.

[0063] In order to verify the recognition effect of the target behavior recognition model on the target hair loss data and the target influence range data, the data dimensions selected for the target health status are verified; including the first verification model, the second verification model and the third verification model.

[0064] The first verification model is the target behavior recognition model of the application, which recognizes target hair loss data and target influence range data through target physiological indicators, hair health data and skin health data.

[0065] The second verification model is based on the target behavior recognition model, and does not consider hair health data, but only recognizes target hair loss data and target influence range data through target physiological indicators and skin health data.

[0066] The third verification model is based on the target behavior recognition model, and does not consider skin health data, but only recognizes target hair loss data and target influence range data through target physiological indicators and hair health data.

[0067] According to the obtained target hair loss data and the actual hair loss data, the hair loss recognition accuracy is obtained; according to the target influence range data and the actual target influence range, the range recognition accuracy is obtained; according to the hair loss recognition accuracy and the range recognition accuracy, the model recognition accuracy is obtained by weighting calculation, wherein the weight is 0.5 by default; the data obtained after multiple verifications is shown in Table 1.

[0068] Table 1 Verification data table of target behavior recognition model

[0069] Model Fuzz recognition accuracy Range recognition accuracy Model recognition accuracy First verification model 0.9246 0.9048 0.9147 Second verification model 0.8624 0.8734 0.8679 Third verification model 0.8892 0.8624 0.8758

[0070] According to the data in Table 1, the recognition effect of the first verification model is the best.

[0071] According to the target behavior, the second home data is split to obtain second home split data; a target behavior recognition model is constructed to recognize the second home split data and target parameters; the hair loss data of the pet is obtained by recognizing the second home split data and target parameters according to the hair loss data recognition layer; the influence range data of the pet is obtained by recognizing the second home split data and target parameters according to the influence range recognition layer; the hair loss condition is accurately recognized according to the behavior data of the pet.

[0072] The first intelligent device warning module obtains the area surface material according to the first home data; the home sub-area is obtained by clustering and dividing according to the area surface material, the first home data and the third home data; the first personalized threshold is set according to the area surface material of the home sub-area, and the first home warning data is obtained according to the first personalized threshold and the third home data.

[0073] The first home warning data is the product of the first personalized threshold and the average hair distribution amount, and the average hair distribution amount of the home sub-area is obtained according to the target hair loss data in the third home data and the area of the home sub-area.

[0074] The division process of the home sub-area is:

[0075] According to the first home data and the third home data, fusion home data is obtained; the fusion home data includes pet trajectory data, target hair loss data corresponding to the pet trajectory data, and a target influence range;

[0076] The fusion home data is divided according to the area surface material, and single-material home data is obtained; the single-material home data includes the fusion home data under the same area surface material;

[0077] The single-material home data is clustered and divided according to a clustering algorithm, and a home sub-area is obtained.

[0078] According to the first home data and the third home data, fusion home data is obtained; the fusion home data includes pet trajectory data, target hair loss data corresponding to the pet trajectory data, and a target influence range; the fusion home data is divided according to the area surface material, and single-material home data is obtained; the fusion home data under the same area surface material is accurately obtained; and the single-material home data is clustered and divided according to a clustering algorithm, and a region affected by pet hair is accurately divided.

[0079] The first personalized threshold reflects the hair cleaning difficulty of the area surface material; the setting process is: an experimental material sample is established according to the area surface material, pet hair is uniformly placed on the experimental material sample, and a second intelligent device is selected to clean the experimental material sample, and the cleaning process is consistent;

[0080] The first personalized threshold is obtained according to the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; the hair cleaning efficiency is determined according to the weight of the cleaned pet hair, and the damage coefficient of the experimental material is obtained according to the difference between the cleaned area image and the standard image.

[0081] According to the area surface material, an experimental material sample is established, pet hair is uniformly placed on the experimental material sample, a second intelligent device is selected to clean the experimental material sample, and the first personalized threshold is obtained according to the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; according to the first personalized threshold and the third home data, the cleaning difficulty of the area surface material is accurately measured.

[0082] The second intelligent device control module determines a warning sub-area according to the first home warning data of the home sub-area, controls the second intelligent device to clean the warning sub-area, and adjusts the parameters of the second intelligent device according to the first home warning data.

[0083] The second intelligent device early warning module obtains sub-region data of the second intelligent device cleaning the early warning sub-region, obtains second home early warning data according to the third home data and the sub-region data, and adjusts and controls the second intelligent device according to the second home early warning data.

[0084] An initial weight before the second intelligent device cleans the early warning sub-region and a final weight after the second intelligent device cleans the early warning sub-region are obtained.

[0085] The sub-region cleaning weight is obtained according to the initial weight and the final weight.

[0086] A foreign matter image of foreign matters obtained by the second intelligent device cleaning the early warning sub-region is obtained, the distribution proportion of pet hair is obtained by identifying the foreign matter image, hair cleaning data and non-hair cleaning data are obtained according to the sub-region cleaning weight and the distribution proportion of pet hair, and the sub-region data is obtained according to the hair cleaning data and the non-hair cleaning data.

[0087] The second home early warning data is obtained according to the hair cleaning data, the non-hair cleaning data and target hair loss data in the third home data.

[0088] The working parameters of the second intelligent device are adjusted according to the second home early warning data.

[0089] The initial weight before the second intelligent device cleans the early warning sub-region and the final weight after the second intelligent device cleans the early warning sub-region are obtained, the sub-region cleaning weight is obtained according to the initial weight and the final weight, the foreign matter image of foreign matters obtained by the second intelligent device cleaning the early warning sub-region is obtained, the distribution proportion of pet hair is obtained, the hair cleaning data and the non-hair cleaning data are obtained according to the sub-region cleaning weight and the distribution proportion of pet hair, and the sub-region data is obtained according to the hair cleaning data and the non-hair cleaning data. The pet hair cleaning condition of the home sub-region is accurately obtained, and the second intelligent device can be accurately adjusted.

[0090] The application further provides a home remote monitoring and early warning method based on a personalized threshold, and a flowchart thereof is shown in Figure 2 The first intelligent device is controlled to monitor a target, first home data and second home data of the target are obtained, the first home data is pet trajectory data, and the second home data is pet behavior data, the second home data is divided and identified according to target behavior, third home data is obtained, and the third home data includes target hair loss data and target influence range data.

[0091] According to the first home data acquisition area surface material; according to the area surface material, the first home data and the third home data, clustering division is carried out, and the home sub-area is obtained;The first personalized threshold is set according to the area surface material of the home sub-area, and the first home early warning data is obtained according to the first personalized threshold and the third home data;

[0092] According to the first home early warning data of the home sub-area, the early warning sub-area is determined, the second intelligent device is controlled to clean the early warning sub-area, and the parameters of the second intelligent device are adjusted according to the first home early warning data;

[0093] Obtain the sub-area data of the second intelligent device cleaning the early warning sub-area, obtain the second home early warning data according to the third home data and the sub-area data, and adjust and control the second intelligent device according to the second home early warning data.

[0094] The application obtains the first home data and the second home data of the target;According to the target behavior, the third home data is obtained by dividing the second home data;According to the first home data, the area surface material is obtained;According to the area surface material, the first home data and the third home data, the home sub-area is obtained;The first personalized threshold of the home sub-area is obtained, and the first home early warning data is obtained in combination with the third home data;According to the first home early warning data, the early warning sub-area is determined;Obtain the sub-area data of the second intelligent device cleaning the early warning sub-area, obtain the second home early warning data according to the third home data and the sub-area data, and adjust and control the second intelligent device according to the second home early warning data.The application realizes the cleaning of pet hair accurately through the first home early warning data and the second home early warning data.

[0095] Example two.

[0096] Based on the actual scene of pet dog, a kind of home remote monitoring and early warning system based on personalized threshold of the application is verified;The structure of the kind of home remote monitoring and early warning system based on personalized threshold is as shown in Figure 1 As shown, it comprises:

[0097] The first intelligent device control module controls the first intelligent device to monitor the target, obtains the first home data and the second home data of the target;The first home data is pet trajectory data, and the second home data is pet behavior data;According to the target behavior, the second home data is divided and identified, and the third home data is obtained, the third home data includes target hair loss data and target influence range data.

[0098] The third home data acquisition process is:

[0099] The target behavior and the target parameter are obtained, the target parameter includes target type, target age, target gender and target health status;

[0100] Splitting the second home data according to the target behavior to obtain the second home split data;

[0101] A target behavior recognition model is constructed to identify the second home splitting data and target parameters, and obtain target hair loss data and target impact range data.

[0102] The target parameters of the pet dog are collected, as shown in Table 2.

[0103] Table 2 Target parameter data table for pet dogs

[0104] Pet number Target species Target age (months) Target gender 01 Labrador 25 0 02 Golden retriever 12 1 03 Poodle 18 1 04 Chihuahua 34 0 05 Akita 23 1

[0105] Among them, in the target gender column in Table 2, "1" represents male and "0" represents female; and the age of the pet is represented by the month, which can also reflect the impact of the month and season on the pet's hair loss.

[0106] The target behavior recognition model includes a hair loss data recognition layer and an impact range recognition layer, and its construction process is as follows:

[0107] Identify the target's historical behavior data to obtain the target's historical parameters, target's historical behavior, target's historical hair loss data, and target's historical impact range;

[0108] Combining target historical parameters, target historical behaviors, target historical hair loss data, and target historical influence range to obtain first historical training data and second historical training data; the first historical training data includes target historical parameters, target historical behaviors, and target historical hair loss data; the second historical training data includes target historical parameters, target historical behaviors, and target historical influence range;

[0109] The hair loss data recognition layer is obtained by training according to the first historical training data; and the impact range recognition layer is obtained by training according to the second historical training data.

[0110] The present invention splits the second household data according to the target behavior to obtain the second household split data; constructs a target behavior recognition model to identify the second household split data and the target parameters; identifies the second household split data and the target parameters according to the hair loss data recognition layer to obtain the pet's hair loss data; identifies the second household split data and the target parameters according to the influence range recognition layer to obtain the pet's influence range data; and accurately identifies the hair loss situation based on the pet's behavior data.

[0111] The first intelligent device early warning module obtains a region surface material according to the first home data; a home sub-region is obtained by clustering and dividing according to the region surface material, the first home data and the third home data; a first personalized threshold is set according to the region surface material of the home sub-region, and the first home early warning data is obtained according to the first personalized threshold and the third home data.

[0112] According to the first home data and the third home data, fusion home data is obtained; the fusion home data includes pet trajectory data, target hair loss data corresponding to the pet trajectory data and a target influence range;

[0113] The fusion home data is divided according to the region surface material to obtain single-material home data; the single-material home data includes the fusion home data under the same region surface material;

[0114] The single-material home data is clustered and divided according to a clustering algorithm to obtain a home sub-region.

[0115] The first personalized threshold reflects the hair cleaning difficulty of the region surface material; the setting process is: an experimental material sample is established according to the region surface material, pet hair is uniformly placed on the experimental material sample, and a second intelligent device is selected to clean the experimental material sample, and the cleaning process is consistent;

[0116] The first personalized threshold is obtained according to the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; the hair cleaning efficiency is determined according to the weight of the cleaned pet hair, and the damage coefficient of the experimental material is obtained according to the difference between the cleaned image and the standard image of the cleaned region.

[0117] The acquisition process of the first personalized threshold is shown in Table 3, wherein the selected experimental material sample is a square with a side length of 1 meter.

[0118] Table 3: Acquisition data table of the first personalized threshold

[0119] Regional surface material Pet hair placement weight Pet hair cleaning weight Damage coefficient of experimental material Floor tile 30.0g 28.1g 0.98 Carpet 30.0g 22.6 0.83 Sofa 30.0g 24.2 0.87 Wooden dining table 30.0g 27.3 0.97 Curtain 30.0g 18.4 0.92

[0120] Wherein, the scene of the carpet is as shown in Figure 3

[0121] According to the first home data and the third home data, fusion home data is obtained; the fusion home data includes pet trajectory data, target hair loss data corresponding to the pet trajectory data and a target influence range; the fusion home data under the same region surface material is accurately obtained by dividing the fusion home data according to the region surface material; and the region affected by the pet hair is accurately divided by clustering and dividing the single-material home data according to a clustering algorithm. ​

[0122] The first personalized threshold reflects the difficulty of hair cleaning of the region surface material; the setting process is: establishing an experimental material sample according to the region surface material, uniformly placing pet hair on the experimental material sample, and selecting a second intelligent device to clean the experimental material sample, with consistent cleaning process;

[0123] The first personalized threshold is obtained according to the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; the hair cleaning efficiency is determined according to the weight of the cleaned pet hair placed; and the damage coefficient of the experimental material is obtained according to the difference between the cleaned region image and the standard image.

[0124] According to the experimental material sample, an experimental material sample is established, pet hair is uniformly placed on the experimental material sample, and a second intelligent device is selected to clean the experimental material sample; the first personalized threshold is obtained according to the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; the difficulty of cleaning the region surface material is accurately measured.

[0125] The second intelligent device control module determines a warning sub-region according to the first home warning data of the home sub-region, controls the second intelligent device to clean the warning sub-region, and adjusts the parameters of the second intelligent device according to the first home warning data.

[0126] The second intelligent device warning module obtains sub-region data of the second intelligent device cleaning the warning sub-region, obtains second home warning data according to the third home data and the sub-region data, and adjusts and controls the second intelligent device according to the second home warning data.

[0127] The initial weight before the second intelligent device cleans the warning sub-region and the final weight after the second intelligent device cleans the warning sub-region are obtained;

[0128] The sub-region cleaning weight is obtained according to the initial weight and the final weight;

[0129] The impurity image of the impurities obtained by the second intelligent device cleaning the warning sub-region is obtained, the impurity image is identified, the distribution proportion of pet hair is obtained, the hair cleaning data and non-hair cleaning data are obtained according to the sub-region cleaning weight and the distribution proportion of pet hair, and the sub-region data is obtained according to the hair cleaning data and the non-hair cleaning data;

[0130] The second home warning data is obtained based on the hair cleaning data, non-hair cleaning data and target hair loss data in the third home data; the working parameters of the second smart device are adjusted according to the second home warning data. If it is determined that the area has been cleaned, the next home sub-area is cleaned; if it is determined that the area has not been cleaned, it is cleaned again; during the re-cleaning process, its parameters can be adjusted to improve the cleaning effect.

[0131] The present invention obtains the initial weight of the warning sub-area before the second smart device cleans the warning sub-area, and the final weight after the second smart device cleans the warning sub-area; obtains the sub-area cleansing weight based on the initial weight and final weight; obtains an impurity image of the impurities obtained by the second smart device in cleaning the warning sub-area, and obtains the distribution ratio of pet hair; obtains hair cleaning data and non-hair cleaning data based on the sub-area cleansing weight and the distribution ratio of pet hair; and obtains sub-area data based on the hair cleaning data and non-hair cleaning data. Accurately obtaining the pet hair cleaning status of the home sub-area allows accurate adjustment of the second smart device.

[0132] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A home remote monitoring and early warning system based on personalized thresholds, characterized in that: include: a first smart device control module, controlling the first smart device to monitor the target and obtain first home data and second home data of the target; The first home data is pet trajectory data, and the second home data is pet behavior data; the second home data is divided and identified according to the target behavior to obtain third home data, and the third home data includes target hair loss data and target influence range data; A first smart device warning module obtains a surface material of a region based on the first home data; performs clustering based on the surface material of the region, the first home data, and the third home data to obtain home sub-regions; obtains a first personalized threshold based on the surface material of the home sub-regions; and obtains first home warning data based on the first personalized threshold and the third home data; a second smart device control module, which determines a warning sub-area based on the first home warning data of the home sub-area, controls the second smart device to clean the warning sub-area, and adjusts parameters of the second smart device based on the first home warning data; The second smart device early warning module obtains the sub-area data of the early warning sub-area cleared by the second smart device, obtains the second home early warning data according to the third home data and the sub-area data, and adjusts and controls the second smart device according to the second home early warning data.

2. The home remote monitoring and early warning system based on personalized threshold according to claim 1, characterized in that: The process of obtaining the third home data in the first smart device control module is as follows: Obtaining target behavior and target parameters, wherein the target parameters include target type, target age, target gender, and target health status, wherein the target health status includes target physiological indicators, hair health data, and skin health data; the physiological indicators include heart rate data, body temperature data, and blood indicators; the hair health data includes pet hair data and pet hair loss data; the skin health data includes skin abnormality area and skin severity coefficient; the skin severity coefficient is determined based on the impact of the skin abnormality area on hair loss; Splitting the second home data according to the target behavior to obtain the second home split data; A target behavior recognition model is constructed to identify the second home splitting data and target parameters, and obtain target hair loss data and target impact range data.

3. The home remote monitoring and early warning system based on personalized threshold according to claim 2, characterized in that: The target behavior recognition model includes a hair loss data recognition layer and an impact range recognition layer, and its construction process is as follows: Identify the target's historical behavior data to obtain the target's historical parameters, target's historical behavior, target's historical hair loss data, and target's historical impact range; Combining target historical parameters, target historical behaviors, target historical hair loss data, and target historical impact range to obtain first historical training data and second historical training data; the first historical training data includes target historical parameters, target historical behaviors, and target historical hair loss data; The second historical training data includes target historical parameters, target historical behaviors, and target historical influence range; The hair loss data recognition layer is obtained by training according to the first historical training data; and the impact range recognition layer is obtained by training according to the second historical training data.

4. The home remote monitoring and early warning system based on personalized thresholds according to claim 1, characterized in that: The process of dividing the home sub-areas is as follows: Obtaining fused home data based on the first home data and the third home data; the fused home data includes pet trajectory data, and target hair loss data and target impact range corresponding to the pet trajectory data; The fused home data is divided according to the surface material of the region to obtain single-material home data; the single-material home data includes the fused home data under the same surface material of the region; The single-material home furnishing data is clustered and divided according to the clustering algorithm to obtain home furnishing sub-areas.

5. The home remote monitoring and early warning system based on personalized threshold according to claim 1, characterized in that: The first personalized threshold reflects the difficulty of hair cleaning of the surface material of the area; the setting process is: creating an experimental material sample based on the surface material of the area, evenly placing pet hair on the experimental material sample, and selecting a second smart device to clean the experimental material sample, and the cleaning process is consistent; A first personalized threshold is obtained based on the hair cleaning efficiency of the experimental material sample and the damage coefficient of the experimental material; the hair cleaning efficiency is determined based on the weight of the cleaned pet hair placed; and the damage coefficient of the experimental material is obtained based on the difference between the cleaned image and the standard image of the cleaned area.

6. The home remote monitoring and early warning system based on personalized threshold according to claim 1, characterized in that: Obtaining an initial weight before the second intelligent device cleans the warning sub-area, and an ending weight after the second intelligent device cleans the warning sub-area; Obtaining a sub-region cleanup weight based on the initial weight and the final weight; Obtaining an impurity image of impurities obtained by cleaning the warning sub-area using the second intelligent device, identifying the impurity image, and obtaining a distribution ratio of pet hair; obtaining hair cleaning data and non-hair cleaning data based on the sub-area cleaning weight and the distribution ratio of pet hair; and obtaining sub-area data based on the hair cleaning data and the non-hair cleaning data; Obtaining second household warning data according to the hair cleaning data, the non-hair cleaning data, and the target hair loss data in the third household data; The operating parameters of the second smart device are adjusted according to the second home early warning data.

7. A home remote monitoring and early warning method based on personalized thresholds, characterized in that: include: Controlling the first smart device to monitor the target and obtain first home data and second home data of the target; The first home data is pet trajectory data, and the second home data is pet behavior data; the second home data is divided and identified according to the target behavior to obtain third home data, and the third home data includes target hair loss data and target influence range data; Obtaining a surface material of the area based on the first home data; performing clustering based on the surface material of the area, the first home data, and the third home data to obtain home sub-areas; setting a first personalized threshold based on the surface material of the home sub-areas, and obtaining first home warning data based on the first personalized threshold and the third home data; determining a warning sub-area based on the first home warning data of the home sub-area, controlling the second smart device to clean the warning sub-area, and adjusting parameters of the second smart device based on the first home warning data; Obtain sub-area data of the warning sub-area cleared by the second smart device, obtain second home warning data based on the third home data and the sub-area data, and adjust and control the second smart device based on the second home warning data.

8. The home remote monitoring and early warning method based on personalized threshold according to claim 7, characterized in that: The process of obtaining the third household data is as follows: Obtaining target behavior and target parameters, wherein the target parameters include target type, target age, target gender, and target health status; the target health status includes target physiological indicators, hair health data, and skin health data; Splitting the second home data according to the target behavior to obtain the second home split data; A target behavior recognition model is constructed to identify the second home splitting data and target parameters, and obtain target hair loss data and target impact range data.

9. The method for remote home monitoring and early warning based on personalized thresholds according to claim 8, characterized in that: The target behavior recognition model includes a hair loss data recognition layer and an impact range recognition layer, and its construction process is as follows: Identify the target's historical behavior data to obtain the target's historical parameters, target's historical behavior, target's historical hair loss data, and target's historical impact range; Combining target historical parameters, target historical behaviors, target historical hair loss data, and target historical impact range to obtain first historical training data and second historical training data; the first historical training data includes target historical parameters, target historical behaviors, and target historical hair loss data; The second historical training data includes target historical parameters, target historical behaviors, and target historical influence range; The hair loss data recognition layer is obtained by training according to the first historical training data; and the impact range recognition layer is obtained by training according to the second historical training data.

Images