Children-oriented sunlight intensity measuring device and method

Abstract

The invention discloses a child-oriented sunlight intensity measurement method and device. The method comprises the following steps: acquiring an image information set and a sunlight intensity information set of a child; the image information set and the sunlight intensity information set are acquired at a plurality of same discrete moments; the sunlight intensity information set comprises sunlight intensity values acquired at a plurality of discrete moments; the image information set comprises child whole body images acquired at a plurality of discrete moments; performing skin identification processing on the image information set to obtain a child skin area information set; and performing weighted summation processing on the child skin area information set by using a sunlight intensity information set to obtain child sunlight intensity measurement information. According to the invention, through the skin image identification model and the collection of the corresponding sunlight intensity data, the sunlight intensity experienced by the child every day is accurately evaluated.

Description

Technical Field

[0001] The present invention relates to the fields of communication and artificial intelligence, and in particular to a sunlight intensity measurement device and method for children. Background Art

[0002] Children are at a critical stage of growth and development. The necessary amount of sunshine time and intensity every day is essential to ensure their growth and development. Children's skin can effectively promote the production of various vitamins by absorbing sunlight. However, in different seasons, due to factors such as clothing, the area of ​​children's skin exposed to the sun is different.

[0003] Therefore, how to accurately assess the intensity of sunlight that children are exposed to every day is an issue that needs to be addressed urgently. Summary of the invention

[0004] The present invention mainly solves the problem of how to accurately evaluate the sunlight intensity experienced by children every day. The present invention discloses a sunlight intensity measurement device and method for children.

[0005] In a first aspect of the embodiments of the present application, a method for measuring sunlight intensity for children is disclosed, comprising:

[0006] S1, collecting and obtaining a set of image information and a set of sunlight intensity information of a child; the set of image information and the set of sunlight intensity information are collected at the same number of discrete moments; the set of sunlight intensity information includes sunlight intensity values ​​collected at the same number of discrete moments; the set of image information includes full-body images of a child collected at the same number of discrete moments;

[0007] S2, performing skin recognition processing on the image information set to obtain a children's skin area information set;

[0008] S3, using the sunlight intensity information set, performing weighted summation processing on the children's skin area information set to obtain children's sunlight intensity measurement information.

[0009] The step of performing skin recognition processing on the image information set to obtain a child skin area information set includes:

[0010] S21, using a skin image recognition model, respectively performing recognition processing on each child's full-body image in the image information set to obtain corresponding skin image information;

[0011] S22, calculating the pixel area of ​​the skin image information corresponding to each child's whole body image to obtain the corresponding child's skin area information;

[0012] S23, fusing all the children's skin area information to obtain a children's skin area information set.

[0013] The skin image recognition model is obtained by training a preset skin image model;

[0014] The skin image model comprises: an encoding unit, a feature fusion unit and a prediction unit;

[0015] The encoding unit comprises a two-dimensional convolution module, an activation module and a quantization module; in the encoding unit, the two-dimensional convolution module, the activation module and the quantization module are connected in sequence;

[0016] The feature fusion unit is composed of a differential module, a cross-scale fusion attention module, a first convolutional layer, an upsampling layer, and a second convolutional layer. The output end of the cross-scale fusion attention module is connected to the input ends of the first convolutional layer and the upsampling layer, the output end of the upsampling layer is connected to the input end of the second convolutional layer, the output end of the first convolutional layer is connected to the first input end of the differential module, and the output end of the second convolutional layer is connected to the second input end of the differential module.

[0017] The input end of the feature fusion unit is connected to the output end of the quantization module;

[0018] The input end of the feature fusion unit is the input end of the cross-scale fusion attention module; the output end of the feature fusion unit is the output end of the difference module;

[0019] The prediction unit includes a fully connected network and an output end; the fully connected network and the output end are connected; the input end of the fully connected network is connected to the output end of the differential module; the output end of the prediction unit is used to output the recognition result of the skin image information.

[0020] The training process of the skin image model includes:

[0021] A training data set is obtained; the training data set includes training data and label information; the training data is children's image data; the label information is skin image information in the children's image data;

[0022] Initialize the number of training iterations;

[0023] Inputting the training data in the training data set as input data into the skin image model;

[0024] Using the skin image model, the input data is processed to obtain a recognition result of the skin image information;

[0025] Performing difference calculation processing on the obtained recognition result of the skin image information and the label information corresponding to the input data to obtain a difference value;

[0026] Determine whether the difference value satisfies a convergence condition, and obtain a first determination result;

[0027] When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result;

[0028] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0029] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0030] When the first judgment result is yes, determining that the model training state satisfies a training termination condition;

[0031] When the model training state does not meet the training termination condition, the skin image model is updated with the parameter update model, the training iteration number value is increased by 1, and the execution is triggered to input the training data in the training data set as input data into the skin image model;

[0032] When the model training state satisfies the training termination condition, the training process of the skin image model is completed, and the trained skin image model is used as a skin image recognition model.

[0033] The parameter updating model is:

[0034]

[0035] θ←θ+v;

[0036] In the formula, x (i) is the i-th training data in the training data set, y (i) is the label information of the i-th training data in the training data set, is the loss function, v is the parameter update value, θ is the parameter of the module to be updated, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, It means to find the partial derivative of variable θ, f(x (i) ; θ) represents the predicted value of the skin image model for the i-th training data in the training data set, f(·) is the calculation function corresponding to the skin image model; exp represents the power operation of the constant e; η and α are preset values.

[0037] The skin image recognition model comprises:

[0038] S211, obtaining a standard child skin image;

[0039] S212, calculating and obtaining the first-order color distance A1, the second-order color distance A2, and the third-order color distance A3 of the standard child skin image;

[0040] S213, evenly dividing the child's full body image to obtain a plurality of sub-images;

[0041] S214, for each sub-image, calculating and obtaining the corresponding first-order color distance B1, second-order color distance B2 and third-order color distance B3;

[0042] S215, performing difference discrimination processing on the first-order color distance B1, the second-order color distance B2 and the third-order color distance B3 of each sub-image and the first-order color distance A1, the second-order color distance A2 and the third-order color distance A3 of the standard child skin image to obtain a difference discrimination result;

[0043] When the difference discrimination result is not obvious, the sub-image is determined to be a skin sub-image; when the difference discrimination result is obvious, the sub-image is not a skin sub-image;

[0044] S216, after executing S214 to S215 on all sub-images of the child's whole body image, all skin sub-images of the child's whole body image are obtained; all the skin sub-images are merged to obtain corresponding skin image information.

[0045] The method of using the sunlight intensity information set to perform weighted summation processing on the children's skin area information set to obtain the children's sunlight intensity measurement information includes:

[0046] Using the sunshine intensity value collected at each discrete moment, multiply the skin area information of the child at the same discrete moment to obtain the cumulative sunshine intensity value at the discrete moment;

[0047] The accumulated sunshine intensity values ​​at all discrete moments are accumulated to obtain the children's sunshine intensity measurement information.

[0048] In a second aspect of an embodiment of the present invention, a sunlight intensity measuring device for children is disclosed, the device comprising:

[0049] A memory storing executable program code;

[0050] a processor coupled to the memory;

[0051] The processor calls the executable program code stored in the memory to execute the sunlight intensity measurement method for children.

[0052] According to a third aspect of an embodiment of the present invention, a computer storable medium is disclosed, wherein the computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the sunlight intensity measurement method for children.

[0053] According to a fourth aspect of the embodiments of the present invention, an information data processing terminal is disclosed, and the information data processing terminal is used to implement the sunlight intensity measurement method for children.

[0054] The beneficial effects of the present invention are:

[0055] The present invention realizes accurate assessment of the sunlight intensity experienced by children every day through a skin image recognition model and collection of corresponding sunlight intensity data.

[0056] The present invention achieves accurate recognition of children's skin images in different seasons, different lighting conditions, different environments, and different wearing conditions by constructing a skin image recognition model, thereby improving the recognition accuracy and the adaptability of the recognition model to different scenarios. During the model training process, a special parameter update model is constructed to achieve accurate conversion from training difference to model update, ensuring rapid convergence of model training.

[0057] The present invention constructs a skin image recognition model based on image recognition and effectively suppresses noise and interference by extracting and processing the first-order color distance and the second-order color distance, thereby ensuring the accuracy of the skin recognition result. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart for implementing the method of the present invention. DETAILED DESCRIPTION

[0059] In order to better understand the content of the present invention, an embodiment is given here.

[0060] Figure 1 It is a flow chart for implementing the method of the present invention.

[0061] In order to solve the problem of how to accurately evaluate the sunlight intensity experienced by children every day, the present invention discloses a sunlight intensity measurement device and method for children.

[0062] Embodiment 1:

[0063] In a first aspect of the embodiments of the present application, a method for measuring sunlight intensity for children is disclosed, comprising:

[0064] S1, collecting and obtaining a set of image information and a set of sunlight intensity information of a child; the set of image information and the set of sunlight intensity information are collected at the same number of discrete moments; the set of sunlight intensity information includes sunlight intensity values ​​collected at the same number of discrete moments; the set of image information includes full-body images of a child collected at the same number of discrete moments;

[0065] S2, performing skin recognition processing on the image information set to obtain a children's skin area information set;

[0066] S3, using the sunlight intensity information set, performing weighted summation processing on the children's skin area information set to obtain the children's sunlight intensity measurement information;

[0067] The step of performing skin recognition processing on the image information set to obtain a child skin area information set includes:

[0068] S21, using a skin image recognition model, respectively performing recognition processing on each child's full-body image in the image information set to obtain corresponding skin image information;

[0069] S22, calculating the pixel area of ​​the skin image information corresponding to each child's whole body image to obtain the corresponding child's skin area information;

[0070] S23, fusing all the children's skin area information to obtain a children's skin area information set;

[0071] The skin image recognition model is obtained by training a preset skin image model;

[0072] The skin image model comprises: an encoding unit, a feature fusion unit and a prediction unit;

[0073] The encoding unit comprises a two-dimensional convolution module, an activation module and a quantization module; in the encoding unit, the two-dimensional convolution module, the activation module and the quantization module are connected in sequence;

[0074] The feature fusion unit is composed of a differential module, a cross-scale fusion attention module, a first convolutional layer, an upsampling layer, and a second convolutional layer. The output end of the cross-scale fusion attention module is connected to the input ends of the first convolutional layer and the upsampling layer, the output end of the upsampling layer is connected to the input end of the second convolutional layer, the output end of the first convolutional layer is connected to the first input end of the differential module, and the output end of the second convolutional layer is connected to the second input end of the differential module.

[0075] The input end of the feature fusion unit is connected to the output end of the quantization module;

[0076] The input end of the feature fusion unit is the input end of the cross-scale fusion attention module; the output end of the feature fusion unit is the output end of the difference module.

[0077] The prediction unit includes a fully connected network and an output end; the input end of the fully connected network is connected to the output end of the differential module; the output end of the prediction unit is used to output the recognition result of the skin image information.

[0078] The training process of the skin image model includes:

[0079] A training data set is obtained; the training data set includes training data and label information; the training data is children's image data; the label information is skin image information in the children's image data;

[0080] Initialize the number of training iterations;

[0081] Inputting the training data in the training data set as input data into the skin image model;

[0082] Using the skin image model, the input data is processed to obtain a recognition result of the skin image information;

[0083] Performing difference calculation processing on the obtained recognition result of the skin image information and the label information corresponding to the input data to obtain a difference value;

[0084] Determine whether the difference value satisfies a convergence condition, and obtain a first determination result;

[0085] When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result;

[0086] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0087] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0088] When the first judgment result is yes, determining that the model training state satisfies a training termination condition;

[0089] When the model training state does not meet the training termination condition, the skin image model is updated with the parameter update model, the training iteration number value is increased by 1, and the execution is triggered to input the training data in the training data set as input data into the skin image model;

[0090] When the model training state satisfies the training termination condition, the training process of the skin image model is completed, and the trained skin image model is used as a skin image recognition model.

[0091] The parameter updating model is:

[0092]

[0093] θ←θ+v;

[0094] In the formula, x (i) is the i-th training data in the training data set, y (i) is the label information of the i-th training data in the training data set, is the loss function, v is the parameter update value, θ is the parameter of the module to be updated, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, It means to find the partial derivative of variable θ, f(x (i) ; θ) represents the predicted value of the skin image model for the i-th training data in the training data set, f(·) is the calculation function corresponding to the skin image model; exp represents the power operation of the constant e; η and α are preset values.

[0095] The skin image recognition model comprises:

[0096] S211, obtaining a standard child skin image;

[0097] S212, calculating and obtaining the first-order color distance A1, the second-order color distance A2, and the third-order color distance A3 of the standard child skin image;

[0098] S213, evenly dividing the child's full body image to obtain a plurality of sub-images;

[0099] S214, for each sub-image, calculating and obtaining the corresponding first-order color distance B1, second-order color distance B2 and third-order color distance B3;

[0100] S215, performing difference discrimination processing on the first-order color distance B1, the second-order color distance B2 and the third-order color distance B3 of each sub-image and the first-order color distance A1, the second-order color distance A2 and the third-order color distance A3 of the standard child skin image to obtain a difference discrimination result;

[0101] When the difference discrimination result is not obvious, the sub-image is determined to be a skin sub-image; when the difference discrimination result is obvious, the sub-image is not a skin sub-image;

[0102] S216, after executing S214 to S215 on all sub-images of the child's whole body image, all skin sub-images of the child's whole body image are obtained; all the skin sub-images are merged to obtain corresponding skin image information.

[0103] The method of using the sunlight intensity information set to perform weighted summation processing on the children's skin area information set to obtain the children's sunlight intensity measurement information includes:

[0104] Using the sunshine intensity value collected at each discrete moment, multiply the skin area information of the child at the same discrete moment to obtain the cumulative sunshine intensity value at the discrete moment;

[0105] The accumulated sunshine intensity values ​​at all discrete moments are accumulated to obtain the children's sunshine intensity measurement information.

[0106] The difference value satisfies the convergence condition, which means that the difference value is less than a preset convergence threshold; the difference value does not satisfy the convergence condition, which means that the difference value is not less than the preset convergence threshold.

[0107] The difference calculation process may be implemented using a loss function.

[0108] The loss function may adopt a cross entropy loss function.

[0109] The cross-scale fusion attention module is implemented using the multi-head attention model in the Transfromer model.

[0110] The activation module is used to implement an activation function; the activation function may be a sigmoid function;

[0111] The calculation formula of the quantization module is:

[0112]

[0113] Wherein, X is the input of the quantization module.

[0114] The difference discrimination process comprises:

[0115] The total discriminant is calculated, and its calculation expression is:

[0116] a1=|A1-B1| / |A1+B1|+|1-A2 / B2|+|exp(A3 / B3)-p| / |A3+B3|, where a1 is the total discriminant and p is the preset calculation factor; the operator || represents the absolute value for scalar values ​​and the Euclidean distance for vector values;

[0117] Determine whether a1 is greater than the set discrimination factor. If so, the difference discrimination result is obvious; if not, the difference discrimination result is not obvious.

[0118] The pixel area calculation is to obtain the total number of pixels included by counting, and the total number is determined as the corresponding child skin area information.

[0119] Embodiment 2:

[0120] This embodiment discloses a system and method for monitoring the sunlight intensity of children based on smart wearable devices. The system includes a smart probe worn on the child's head, which is used to detect the duration of exposure to sunlight and the intensity of ultraviolet rays, and transmit the relevant data to the corresponding mobile phone software wirelessly. Functions of the mobile phone software:

[0121] S01, select the current child's clothing in the form of a schematic diagram, whether it is long pants or shorts, long sleeves or short sleeves, and whether it is wearing a hat, to identify which parts of the current child are exposed.

[0122] S02, calculating the body surface area exposed to sunlight according to the child's age and weight and the skin exposed parts in step S01.

[0123] S03, calculating the exposure intensity of the current period according to the area of ​​S02, the ultraviolet intensity of the current period, and the exposure duration.

[0124] S04. Record your exposure every time you go out. You can calculate your exposure for a day, a month, etc. By comparing it with the guidelines, you can determine whether you have reached the sunlight intensity required for the healthy growth of children.

[0125] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for measuring sunlight intensity for children, characterized in that: include: S1, collecting and obtaining a set of image information and a set of sunlight intensity information of a child; the set of image information and the set of sunlight intensity information are collected at the same number of discrete moments; the set of sunlight intensity information includes sunlight intensity values ​​collected at the same number of discrete moments; the set of image information includes full-body images of a child collected at the same number of discrete moments; S2, performing skin recognition processing on the image information set to obtain a children's skin area information set; S3, using the sunlight intensity information set, performing weighted summation processing on the children's skin area information set to obtain children's sunlight intensity measurement information.

2. The method for measuring sunlight intensity for children according to claim 1, characterized in that: The step of performing skin recognition processing on the image information set to obtain a child skin area information set includes: S21, using a skin image recognition model, respectively performing recognition processing on each child's full-body image in the image information set to obtain corresponding skin image information; S22, calculating the pixel area of ​​the skin image information corresponding to each child's whole body image to obtain the corresponding child's skin area information; S23, fusing all the children's skin area information to obtain a children's skin area information set.

3. The method for measuring sunlight intensity for children according to claim 2, characterized in that: The skin image recognition model is obtained by training a preset skin image model; The skin image model comprises: an encoding unit, a feature fusion unit and a prediction unit; The encoding unit comprises a two-dimensional convolution module, an activation module and a quantization module; in the encoding unit, the two-dimensional convolution module, the activation module and the quantization module are connected in sequence; The feature fusion unit is composed of a differential module, a cross-scale fusion attention module, a first convolutional layer, an upsampling layer, and a second convolutional layer. The output end of the cross-scale fusion attention module is connected to the input ends of the first convolutional layer and the upsampling layer, the output end of the upsampling layer is connected to the input end of the second convolutional layer, the output end of the first convolutional layer is connected to the first input end of the differential module, and the output end of the second convolutional layer is connected to the second input end of the differential module. The input end of the feature fusion unit is connected to the output end of the quantization module; The input end of the feature fusion unit is the input end of the cross-scale fusion attention module; the output end of the feature fusion unit is the output end of the difference module; The prediction unit includes a fully connected network and an output end; the fully connected network and the output end are connected; the input end of the fully connected network is connected to the output end of the differential module; the output end of the prediction unit is used to output the recognition result of the skin image information.

4. The method for measuring sunlight intensity for children as claimed in claim 3, characterized in that: The training process of the skin image model includes: A training data set is obtained; the training data set includes training data and label information; the training data is children's image data; the label information is skin image information in the children's image data; Initialize the number of training iterations; Inputting the training data in the training data set as input data into the skin image model; Using the skin image model, the input data is processed to obtain a recognition result of the skin image information; Performing difference calculation processing on the obtained recognition result of the skin image information and the label information corresponding to the input data to obtain a difference value; Determine whether the difference value satisfies a convergence condition, and obtain a first determination result; When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result; When the second judgment result is no, determining that the model training state does not meet the training termination condition; When the second judgment result is yes, determining that the model training state satisfies the training termination condition; When the first judgment result is yes, determining that the model training state satisfies a training termination condition; When the model training state does not meet the training termination condition, the skin image model is updated with the parameter update model, the training iteration number value is increased by 1, and the execution is triggered to input the training data in the training data set as input data into the skin image model; When the model training state satisfies the training termination condition, the training process of the skin image model is completed, and the trained skin image model is used as a skin image recognition model.

5. The method for measuring sunlight intensity for children according to claim 4, characterized in that: The parameter updating model is: θ←θ+v; In the formula, x (i) is the i-th training data in the training data set, y (i) is the label information of the i-th training data in the training data set, is the loss function, v is the parameter update value, θ is the parameter of the module to be updated, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, It means to find the partial derivative of variable θ, f(x (i) ; θ) represents the predicted value obtained by the skin image model for the i-th training data in the training data set, f(·) is the calculation function corresponding to the skin image model; exp represents the power operation of the constant e; η and α are preset values.

6. The method for measuring sunlight intensity for children according to claim 2, characterized in that: The skin image recognition model comprises: S211, obtaining a standard child skin image; S212, calculating and obtaining the first-order color distance A1, the second-order color distance A2, and the third-order color distance A3 of the standard child skin image; S213, evenly dividing the child's full body image to obtain a plurality of sub-images; S214, for each sub-image, calculating and obtaining the corresponding first-order color distance B1, second-order color distance B2 and third-order color distance B3; S215, performing difference discrimination processing on the first-order color distance B1, the second-order color distance B2 and the third-order color distance B3 of each sub-image and the first-order color distance A1, the second-order color distance A2 and the third-order color distance A3 of the standard child skin image to obtain a difference discrimination result; When the difference discrimination result is not obvious, the sub-image is determined to be a skin sub-image; when the difference discrimination result is obvious, the sub-image is not a skin sub-image; S216, after executing S214 to S215 on all sub-images of the child's whole body image, all skin sub-images of the child's whole body image are obtained; all the skin sub-images are merged to obtain corresponding skin image information.

7. The method for measuring sunlight intensity for children according to claim 2, characterized in that: The method of using the sunlight intensity information set to perform weighted summation processing on the children's skin area information set to obtain the children's sunlight intensity measurement information includes: Using the sunshine intensity value collected at each discrete moment, multiply the skin area information of the child at the same discrete moment to obtain the cumulative sunshine intensity value at the discrete moment; The accumulated sunshine intensity values ​​at all discrete moments are accumulated to obtain the children's sunshine intensity measurement information.

8. A sunlight intensity measuring device for children, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the sunlight intensity measurement method for children according to any one of claims 1 to 7.

9. A computer storable medium, characterized in that: The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the sunlight intensity measurement method for children according to any one of claims 1 to 7.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the sunlight intensity measurement method for children according to any one of claims 1 to 7.

Images