Distance prompting method based on notebook light sensation

Abstract

The invention discloses a distance prompting method based on notebook light sensation. The method comprises the following steps: ultra-wide-angle light collection: arranging a multi-layer composite lens array containing a micro-lens unit to collect light; extracting depth information, calculating a distance by using light field imaging characteristics and a deep learning algorithm, and capturing iris and facial micro-expressions through an infrared camera to perform biological feature fusion distance measurement; judging a distance threshold value, and judging an interval to which the user distance belongs according to a preset graded distance interval threshold value; multi-mode prompting is carried out, visual, auditory and tactile prompts are provided according to distance intervals, and a privacy protection mechanism is set; adaptive adjustment is carried out, and a deep reinforcement learning framework is used to optimize an algorithm and a prompt mode; and environment compensation: correcting a measurement result in combination with environment illumination and temperature data. The method has the advantages that the distance measurement error is reduced through the special lens array, the distance measurement accuracy is improved through multi-modal fusion, and the individual requirements of different users can be met.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of notebook computer applications, and in particular to a distance prompting method based on notebook light sensing. BACKGROUND

[0002] With the widespread use of notebook computers, the damage to vision caused by users using notebook computers for a long time at close range has become a serious problem. Long-term close viewing distance can cause the ciliary muscle of the eye to be in a state of tension, which can easily lead to myopia, visual fatigue and other eye problems. According to statistics, the incidence of myopia in people who frequently use notebook computers is increasing year by year. However, traditional notebook computers do not have effective means to remind users to maintain an appropriate viewing distance.

[0003] In common notebook light sensing distance measurement methods, the light collection method is mostly narrow-angle light collection, which has the problem of large distance measurement error due to incomplete light collection. At the same time, the existing technology lacks means to improve the absorption and conversion efficiency of specific wavelength light to improve the collection accuracy in terms of light collection materials, resulting in a large influence of light during distance measurement. In terms of notebook light sensing distance prompting, the existing distance interval threshold settings are not flexible enough and cannot be well personalized according to the needs and usage habits of different users, which cannot meet the work requirements of notebook computer applications. Therefore, a distance prompting method based on notebook light sensing is proposed. SUMMARY

[0004] The present application provides the following technical solution: a distance prompting method based on notebook light sensing, comprising: S1 ultra-wide-angle light collection: A multi-layer composite lens array is set up to capture ambient light, the lens array contains a plurality of microlens units with a diameter of less than 0.1 mm, forming a wide-angle light capture area with a field of view angle of 90°, and a perovskite quantum dot material is coated on the surface of the microlens unit at the same time; S2 depth information extraction: The light field imaging characteristics of the microlens array in step S1 are utilized, and a deep learning algorithm is combined to calculate the real-time distance between the user and the screen. At the same time, the user's iris diameter change and facial micro-expression are captured by an infrared camera for biological feature fusion ranging, and a multi-modal biological ranging model is established in combination with the light field ranging data; S3 distance threshold judgment: A plurality of distance interval thresholds are preset to determine the interval to which the current user belongs; S4 multi-modal prompting: According to the distance interval, the corresponding visual, auditory and tactile prompting methods are selected, and a privacy protection mechanism is added at the same time; S5 adaptive adjustment: The adaptive adjustment algorithm optimizes the distance judgment algorithm and the prompting mode according to user feedback and usage habits; S6 Environment compensation: The distance measurement result is corrected in real time in combination with ambient light and temperature data.

[0005] Preferably, the multi-layer composite lens array in step S1 adopts a three-dimensional arrangement mode, and comprises a substrate layer, a microlens layer and an anti-reflection coating layer. The microlens layer is composed of 100-200 hemispherical microlenses, and the radius of curvature is 0.05-0.1 mm.

[0006] Preferably, the deep learning algorithm in step S2 adopts a 3D-ResNet network architecture, which comprises 5 residual modules and 3 spatio-temporal attention modules.

[0007] Preferably, the biometric feature fusion ranging in step S2 specifically comprises: capturing a user's iris image at a frame rate of 60 fps through an 850 nm near-infrared camera to measure the pupil diameter change curve; simultaneously tracking 68 feature points using a facial recognition algorithm to analyze the micro-expression change frequency; and finally fusing the light field depth data and the biometric feature data through an adaptive weighting algorithm.

[0008] Preferably, the distance interval threshold in step S3 is set in a hierarchical manner. The first-level warning interval of the distance interval threshold is 20-30 cm, triggering a high-level prompt. The second-level warning interval of the distance interval threshold is 30-45 cm, triggering a moderate-level prompt. The third-level safety interval of the distance interval threshold is 45-60 cm, only triggering a slight prompt.

[0009] Preferably, the visual prompt in step S4 adopts a dynamic light strip design at the edge of the screen. The light strip color is displayed in red, orange and green according to the distance interval. The auditory prompt is a 3D spatial sound effect with a frequency of 200-800 Hz. The tactile prompt is performed through a piezoelectric ceramic vibrator under the keyboard. The vibration frequency of the piezoelectric ceramic vibrator is 50-200 Hz.

[0010] Preferably, the adaptive adjustment algorithm in step S5 adopts a deep reinforcement learning framework. The deep reinforcement learning framework records the distance parameters, ambient light conditions and application scene types adjusted by the user each time, establishes a user preference model, and the preference model automatically updates the model parameters once a week.

[0011] Preferably, the privacy protection mechanism in step S4 automatically switches to a concealed prompt mode when sensitive scenes such as video conferences and online exams are detected. The visual prompt is changed to flashing icons in the corners of the screen, the auditory prompt is changed to bone conduction vibration only perceptible to the user, and all data uploading functions are temporarily suspended.

[0012] Preferably, in the step S2, a child special protection mechanism is provided when information is extracted, after the child user is recognized by face feature analysis and skeleton key point detection of the camera, a strict prompt strategy is automatically enabled, the strict prompt strategy includes increasing the warning distance threshold by 10cm, increasing the prompt frequency by 50%, and locking the system settings to prevent the child from turning off the prompt function.

[0013] Preferably, in the step S6, when considering the influence of temperature data on distance measurement results, a temperature-distance compensation model is established simultaneously, which associates temperature changes with changes in optical performance of the light field imaging device.

[0014] Compared with the prior art, the distance prompting method based on notebook light sensing provided by the present application has the following advantages: 1、The present application sets up a multi-layer composite lens array and contains a plurality of microlens units, forming a wide-angle light capture area with a 90° field of view, so that a wider range of ambient light can be collected, compared with the traditional narrow-angle light collection method, reducing the distance measurement error caused by incomplete light collection, and the perovskite quantum dot material coated on the surface of the microlens unit further enhances the light collection effect, as the perovskite quantum dot material has special optical properties, it can improve the absorption and conversion efficiency of specific wavelength light, making the light collection more accurate, which helps to improve the distance measurement accuracy based on the light field imaging characteristics; 2、The present application combines the light field imaging characteristics of the microlens array, the deep learning algorithm, and the user's iris diameter change and facial micro-expression captured by the infrared camera to establish a multi-modal biological ranging model, this multi-modal fusion method makes full use of the advantages of different data sources, that is, the light field imaging characteristics provide basic distance-related light information, the deep learning algorithm can efficiently analyze and process these information, and the iris diameter change and facial micro-expression add biological feature dimension information to distance measurement, thereby improving the accuracy of the entire distance measurement; 3、The present application predefines multiple distance interval thresholds, so that personalized distance interval settings can be made according to the needs and usage habits of different users, thereby meeting the individual needs of different users, and through visual, auditory and tactile prompting methods, it can adapt to the needs of different users in different scenarios, visual prompts can directly prompt users on the screen or through LED light strips at the edge of the screen, auditory prompts can be directed at users who are not constantly paying attention to the screen, and tactile prompts can privately remind users without affecting others, improving the comprehensiveness of user prompts. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a flow chart of the method of the present application. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0017] Please refer to Figure 1 The present application provides a technical solution, a distance prompting method based on notebook light sensing, comprising the following steps: S1 ultra-wide-angle light collection: A multi-layer composite lens array is arranged to capture ambient light. The lens array includes a plurality of microlens units with a diameter less than 0.1 mm, forming a wide-angle light capture area with a field of view of 90°. A perovskite quantum dot material is coated on the surface of the microlens units. The multi-layer composite lens array adopts a three-dimensional arrangement, including a base layer, a microlens layer, and an anti-reflection coating. The microlens layer is composed of 100-200 hemispherical microlenses with a radius of curvature of 0.05-0.1 mm. The specific implementation process of the above method is as follows: Construction of multi-layer composite lens array: First, determine the material suitable for the base layer. This material needs to have good optical performance, stability, and compatibility with other layers. For example, optical glass or transparent polymer materials can be selected. Then, the base layer is processed into a shape and size that meets the requirements of the internal space and structure of the notebook. According to the regulations, microlens units with a diameter less than 0.1 mm are designed. Each microlens unit is hemispherical, with a radius of curvature set in the range of 0.05 - 0.1 mm. High-precision molds or lithography techniques are used to manufacture these microlens units, ensuring the accuracy of their shape and size. 100 - 200 such manufactured hemispherical microlenses are arranged in a predetermined manner to form a microlens layer. After the microlens layer is completed, a physical vapor deposition (PVD) or chemical vapor deposition (CVD) technique is used to add an anti-reflection coating on the surface of the microlens layer. The material selection of the anti-reflection coating should be able to effectively reduce the reflection of light on the lens surface and improve the transmittance of light. For example, materials such as magnesium fluoride (MgF2) can be selected. The base layer, microlens layer, and anti-reflection coating are combined in a three-dimensional arrangement. This requires precise positioning and assembly processes to ensure that the layers are tightly combined and do not affect the propagation path of light. For example, special glue or clamps can be used to fix the layers, while avoiding the introduction of dust or impurities during the assembly process that may affect the propagation of light; Perovskite quantum dot material coating: First, synthesize perovskite quantum dot material, which requires strict control of reaction conditions such as temperature, reaction time, and raw material ratio to ensure that the synthesized quantum dots have the desired optical properties. For example, using the solution method, control the reaction temperature within a certain range, the reaction time is accurate to the minute level, and the raw materials are mixed according to a specific molar ratio. Use appropriate coating methods such as spin coating or spraying to uniformly coat the prepared perovskite quantum dot material on the surface of the microlens unit. When spin coating, parameters such as spin speed, time, and quantum dot solution concentration need to be controlled to obtain a uniform coating thickness; when spraying, the distance between the spray head and the microlens unit, the spraying pressure and speed, etc. need to be adjusted to ensure the integrity and uniformity of the coating; The wide-angle light capture area formed by the multi-layer composite lens array with a 90° field of view angle enables the collection of a wider range of ambient light. In notebook use scenarios, whether in a complex indoor light environment (such as multiple light sources or uneven light distribution) or in outdoor light from different directions, light information can be fully acquired. Compared with traditional narrow-angle light collection methods, distance measurement errors caused by incomplete light collection are greatly reduced. For example, in a location near a window indoors, there are both strong light from the window and light from other directions in the room. Wide-angle light capture can collect all these lights, providing a more sufficient data basis for accurate distance calculation; The perovskite quantum dot material coated on the surface of the microlens unit further enhances the effect of light collection. Perovskite quantum dot material has special optical properties that can improve the absorption and conversion efficiency of specific wavelength light. In ambient light, different wavelengths of light have different propagation and reflection characteristics. Through the high-efficiency absorption and conversion of perovskite quantum dot material for specific wavelength light, the information in the light can be more accurately captured, which helps to improve the distance measurement accuracy based on light field imaging characteristics. For example, certain specific wavelengths of light may contain key information related to distance measurement, and perovskite quantum dot material can better capture and utilize these lights, thereby improving the accuracy of distance measurement; S2 depth information extraction: Utilize the light field imaging characteristics of the microlens array in step S1 and combine the deep learning algorithm to calculate the real-time distance between the user and the screen, while capturing the user's iris diameter change and facial micro-expression through the infrared camera for biometric fusion ranging. Combine the light field ranging data to establish a multi-modal biometric ranging model. The deep learning algorithm uses a 3D-ResNet network architecture, which includes 5 residual modules and 3 spatio-temporal attention modules. The implementation process of the above method is as follows: Network architecture initialization: According to the requirements of the 3D-ResNet network architecture, it is determined that the entire network will contain 5 residual modules and 3 spatio-temporal attention modules. The design of this structure is to effectively extract features and improve the accuracy of distance calculation when processing distance prompt related data based on notebook light sensing. The input data will be obtained from the light ray data acquired by the light field imaging characteristics of the micro-lens array in the ultra-wide angle light collection step (S1), as well as the user's iris diameter change and facial micro-expression data captured by the infrared camera. These data need to be organized in a specific format to adapt to the input requirements of the 3D-ResNet network. For example, the light ray data may be three-dimensional image data (containing spatial and temporal dimensions, as it may be continuous frame data), while the iris and facial expression data also need to be formatted accordingly, which may be converted into a tensor form matching the dimensions of the light ray data; Residual module construction and connection: For each residual module, first construct its basic convolutional layer structure. Use 3D convolution kernels for convolution operations to process data with spatio-temporal characteristics. For example, the size of the convolution kernel can be set according to the characteristics of the input data and the desired feature extraction effect, such as setting the convolution kernel size to 3x3x3. After the convolution layer, add a batch normalization (Batch Normalization) layer. Batch normalization helps to speed up the training process of the network, improve the convergence speed of the network, and reduce the problem of gradient vanishing or explosion. It will normalize the data of each batch, making the distribution of the data more stable, and then apply a nonlinear activation function such as the ReLU (Rectified Linear Unit) function. The ReLU function can increase the non-linear expression ability of the network, set the negative input value to 0, and only retain the positive value part, which helps the network to learn more complex feature relationships. The construction of the residual connection part is the key feature of the residual module. By connecting the input directly to the output part of the module (with appropriate dimension adjustment if necessary), the network can more easily learn the identity mapping, which helps to alleviate the degradation problem in deep network training. The constructed 5 residual modules are connected in order. The output of the previous residual module is used as the input of the next residual module, which can gradually extract high-level features from the data. During the connection process, it is necessary to ensure that the data dimensions between each module match, including spatial dimensions, temporal dimensions, and channel dimensions, etc. Space-time attention module integration: in the space-time attention module, first, the space attention submodule and the time attention submodule are constructed respectively. For the space attention submodule, it will perform attention calculation on the spatial dimension of the input data (such as the width and height dimension of the image). Convolutional layers can be used to calculate the importance weight of each spatial position, and then these weights are applied to the input data, so that the network can pay more attention to the key spatial area related to the distance prompt. For the time attention submodule, attention calculation is performed on the time dimension of the input data (such as the order of consecutive frames). The idea of recurrent neural network (RNN) or self-attention mechanism (Self-Attention) can be used to calculate the importance weight of each time step, so that the network can focus on the key information related to distance calculation at different time points. The outputs of the space attention submodule and the time attention submodule are fused to obtain the final output of the space-time attention module. The constructed three space-time attention modules are inserted into the network structure composed of five residual modules. The specific insertion position can be determined according to the experiment and the understanding of the data characteristics, for example, a space-time attention module can be inserted every certain number of residual modules. During the integration process, it is necessary to ensure that the input and output data dimensions of the space-time attention module match the surrounding residual modules, so that the data can flow smoothly in the whole network; The biological feature fusion ranging specifically includes: capturing the user's iris image through an 850nm near-infrared camera at a frame rate of 60fps to measure the pupil diameter change curve; simultaneously tracking 68 feature points using a face recognition algorithm to analyze the micro-expression change frequency; finally, the light field depth data and the biological feature data are fused through an adaptive weighting algorithm, and the specific process of the above method is as follows: Iris image capture and pupil diameter change curve measurement: An 850nm near-infrared camera is used, as this wavelength is chosen for its good penetration and safety for the eye in iris recognition. The camera is mounted in a suitable position on the notebook, ensuring that it can clearly capture the user's eye area, and the camera's frame rate is set to 60fps. This frame rate ensures that sufficient image data is obtained without excessive system resource consumption. The user's eye images are captured at a frequency of 60 times per second, ensuring that dynamic changes in the eye can be captured. Each captured iris image is pre-processed. First, grayscale processing is performed to convert the color image to a grayscale image, reducing the data volume while highlighting the main features of the image. Then, noise reduction processing is performed using methods such as Gaussian filtering to remove noise in the image and improve image clarity. Pupil positioning techniques in iris recognition algorithms, such as the circular detection-based method, are used to accurately locate the pupil in the pre-processed image. The pupil is located by detecting the gray level changes in the pupil edge or using the circular feature of the pupil. Once the pupil position is determined, the pupil diameter can be measured. The diameter value can be obtained by calculating the distance between two points on the pupil edge. This measurement is performed for each image frame, resulting in a series of pupil diameter values over time, forming a pupil diameter change curve; Facial feature point tracking and micro-expression change frequency analysis: A pre-trained facial recognition algorithm is used, which can recognize and track 68 facial feature points. These feature points are distributed in key locations on the face, such as the eyes, eyebrows, nose, and mouth. The facial images captured by the camera are input into the facial recognition algorithm. The algorithm automatically locates the initial positions of these 68 feature points in the image. In subsequent image frames, the algorithm continuously tracks the position changes of these feature points based on their characteristics and the surrounding image information, calculating the displacement of each feature point between adjacent frames. By comparing the coordinate positions of the same feature point in different frames, the displacement of the feature point in the horizontal and vertical directions can be obtained. According to the displacement of the feature points and specific pattern recognition methods, micro-expressions can be recognized. For example, the upward movement of the mouth corners and the wrinkling of the eyebrows will cause specific displacement patterns of related feature points. By analyzing the change frequency of these displacement patterns, the change frequency of micro-expressions can be understood; Fusion of light field depth data and biometric data: Light field depth data is obtained from the micro-lens array light field imaging characteristics in the ultra-wide-angle light ray collection step (S1). The light field depth data reflects the preliminary distance information between the user and the screen, and organizes the previously obtained biometric data, i.e. the pupil diameter change curve and the facial micro-expression change frequency data. These data are standardized to make them within the same numerical range, facilitating subsequent fusion operations. An adaptive weighting algorithm determines the respective weights of the light field depth data and the biometric data according to the importance and reliability of the data. For example, if the accuracy of the light field depth data is higher in a certain environment, it may be given a higher weight; if the biometric data performs better in distinguishing user states, its weight will also be adjusted accordingly. The light field depth data is multiplied by its corresponding weight, and the biometric data is multiplied by its corresponding weight, then the two are added to obtain the fused distance data. This fused data considers the information from both light field imaging and biometrics, thereby improving the accuracy of the entire distance measurement; A child special protection mechanism is set when information is extracted. After the child special protection mechanism recognizes a child user through face feature analysis and skeleton key point detection, a strict prompt strategy is automatically enabled. The strict prompt strategy includes increasing the warning distance threshold by 10 cm, increasing the prompt frequency by 50%, and locking the system settings to prevent the child from turning off the prompt function. The specific process of the above method is as follows: Child user identification: Obtain the user's facial image using the notebook's built-in camera. The camera continuously captures image frames at a certain frame rate (e.g., compatible with other functions), and feature extraction is performed on the obtained facial image. This process can use a deep learning-based facial recognition algorithm that can identify various features of the face, such as the shape, position, and proportion of the facial features. For example, the image is processed through a convolutional neural network (CNN) to extract facial feature representations at different levels, and the extracted facial features are compared with a pre-constructed child facial feature model. This child facial feature model is trained using a large amount of child facial image data and contains typical features of children's faces at different ages, genders, and ethnicities. By calculating the similarity between the features, such as the Euclidean distance or cosine similarity, it is determined whether the current user is likely to be a child. The facial image is subjected to skeletal key point detection. A specialized skeletal key point detection algorithm, such as one based on human pose estimation technology, is used to locate key skeletal points on the face, such as the eye sockets, cheekbones, and jawbones. These key points reflect the skeletal structure of the face, and the structure of the located skeletal key points is analyzed. The facial skeletal structure of children is significantly different from that of adults, such as the facial proportions and skeletal development. By comparing the skeletal structure features with the pre-set child skeletal structure feature range, the user is further determined to be a child. If the facial feature analysis and skeletal key point detection results both indicate that the user has a high probability of being a child, the current user is determined to be a child user; Strict prompt policy activation: Obtain the current distance interval threshold from the distance threshold determination module. For example, the first warning interval is 20-30 cm, the second reminder interval is 30-45 cm, and the third safety interval is 45-60 cm. Increase the lower limit of the first warning interval by 10 cm. That is, the new first warning interval becomes 30-30 cm (the lower limit increases by 10 cm and coincides with the upper limit, which can be further adjusted according to actual needs), and the second reminder interval and the third safety interval can be adjusted accordingly (such as maintaining the interval length) based on the new first warning interval. Determine the prompt frequency in each distance interval. For example, in different distance intervals, visual, auditory, and tactile prompts may have their own default prompt frequencies. Increase the prompt frequency of each distance interval by 50%. If the original prompt frequency is once every 5 minutes, it increases to once every 3.33 minutes after a 50% increase (calculation method: 5 ÷ (1 + 50%)). Enable the system setting lock function to prevent the child user from turning off the prompt function. This can be achieved through the operating system's permission management mechanism, such as setting the system setting options related to the distance prompt function to an unmodifiable state after detecting a child user, and only unlocking them for modification after passing a specific parent or administrator authentication (such as password verification). S3 Distance threshold judgment: A plurality of distance interval thresholds are preset to determine the distance of the current user from the interval. The distance interval threshold is set in stages. The first warning interval of the distance interval threshold is 20-30 cm, triggering a high-level prompt. The second warning interval of the distance interval threshold is 30-45 cm, triggering a moderate prompt. The third safety interval of the distance interval threshold is 45-60 cm, only triggering a slight prompt. S4 Multi-modal prompt: According to the distance interval, the corresponding visual, auditory and tactile prompt modes are selected, and a privacy protection mechanism is added synchronously. The visual prompt adopts a dynamic light strip design at the edge of the screen. The light strip color is displayed in red, orange and green according to the distance interval. The auditory prompt is a 3D spatial sound effect with a frequency of 200-800 Hz. The tactile prompt is performed through a piezoelectric ceramic vibrator below the keyboard. The vibration frequency of the piezoelectric ceramic vibrator is 50-200 Hz. The privacy protection mechanism automatically switches to a concealed prompt mode when detecting sensitive scenarios such as video conferences and online exams. The visual prompt is changed to an inconspicuous icon flashing in the corner of the screen. The auditory prompt is changed to a bone conduction vibration that only the user can perceive. At the same time, all data upload functions are suspended. S5 Self-adaptive adjustment: An adaptive adjustment algorithm is used to optimize the distance judgment algorithm and prompt mode according to user feedback and usage habits. The adaptive adjustment algorithm uses a deep reinforcement learning framework. The deep reinforcement learning framework records the distance parameters, ambient light conditions and application scenario types adjusted manually by the user each time, establishes a user preference model, and updates the model parameters of the preference model automatically once a week. S6 Environmental compensation: The distance measurement results are corrected in real time in combination with environmental light and temperature data. When considering the influence of temperature data on distance measurement results, a temperature-distance compensation model is established synchronously. The temperature-distance compensation model relates temperature changes to changes in the optical performance of the light field imaging device.

[0018] The present scheme forms a wide-angle light ray capture area with a 90° field of view by setting a multi-layer composite lens array containing multiple microlens units, thereby being able to collect a wider range of environmental light. Compared with the traditional narrow-angle light collection method, the distance measurement error caused by incomplete light collection is reduced. The effect of light collection is further enhanced by the perovskite quantum dot material coated on the surface of the microlens unit. Since the perovskite quantum dot material has special optical properties, it can improve the absorption and conversion efficiency of specific wavelength light, making light collection more accurate and helping to improve the distance measurement precision based on the light field imaging characteristics.

[0019] The scheme combines the light field imaging characteristics of the microlens array, the deep learning algorithm, and the changes in the user's iris diameter and facial micro-expression captured by the infrared camera to establish a multi-modal biometric ranging model. This multi-modal fusion method takes full advantage of the strengths of different data sources, i.e., the light field imaging characteristics provide basic distance-related light information, the deep learning algorithm can efficiently analyze and process this information, and the changes in the iris diameter and facial micro-expression add biological feature dimension information to distance measurement, thereby improving the accuracy of the entire distance measurement.

[0020] The scheme presets multiple distance interval thresholds, allowing personalized distance interval setting according to the needs and usage habits of different users, thereby meeting the individual needs of different users. The visual, auditory, and tactile prompting methods can adapt to the needs of different users in different scenarios. Visual prompts can directly provide intuitive prompts to users on the screen or through screen edge LED light strips, auditory prompts can be provided to users who are not constantly focused on the screen, and tactile prompts can provide private reminders to users without affecting others, thereby improving the comprehensiveness of user prompts.

[0021] It should be noted that, in this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0022] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements, and variations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A distance cues method based on a laptop's optical sensor, characterized in that, Includes the following steps: S1 Ultra Wide Angle Light Acquisition: A multi-layer composite lens array is configured to capture ambient light. The lens array contains multiple microlens units with a diameter of less than 0.1 mm, forming a wide-angle light-capturing area with a 90° field of view. Perovskite quantum dot material is simultaneously coated on the surface of the microlens units. S2 depth information extraction: Utilizing the light field imaging characteristics of the microlens array in step S1, and combining it with a deep learning algorithm, the real-time distance between the user and the screen is calculated. At the same time, the user's iris diameter changes and facial micro-expressions are captured by an infrared camera to perform biometric fusion ranging. Furthermore, a multimodal biometric ranging model is established by combining the light field ranging data. S3 Distance Threshold Judgment: Multiple distance range thresholds are preset to determine the current user's distance range. S4 Multimodal Hints: Select the corresponding visual, auditory, and tactile cues based on the distance range, and simultaneously add a privacy protection mechanism; S5 adaptive adjustment: The adaptive adjustment algorithm optimizes the distance judgment algorithm and prompting method based on user feedback and usage habits; S6 Environmental Compensation: The distance measurement results are corrected in real time by combining ambient light and temperature data.

2. The distance cues method based on laptop light sensing according to claim 1, characterized in that: The multilayer composite lens array in step S1 adopts a three-dimensional arrangement and includes a base layer, a microlens layer and an anti-reflective coating. The microlens layer is composed of 100-200 hemispherical microlenses with a radius of curvature of 0.05-0.1 mm.

3. The distance cues method based on laptop light sensing according to claim 1, characterized in that: The deep learning algorithm in step S2 adopts a 3D-ResNet network architecture, which includes 5 residual modules and 3 spatiotemporal attention modules.

4. The distance cues method based on laptop light sensing according to claim 1, characterized in that: The biometric fusion ranging in step S2 includes: capturing the user's iris image at 60fps using an 850nm near-infrared camera and measuring the pupil diameter change curve; simultaneously tracking 68 feature points using a facial recognition algorithm and analyzing the frequency of micro-expression changes; and finally fusing the light field depth data and biometric data using an adaptive weighting algorithm.

5. The distance cues method based on laptop light sensing according to claim 1, characterized in that: In step S3, the distance interval threshold is set in a tiered manner. The first-level warning interval of the distance interval threshold is 20-30cm, which triggers a high-level warning. The second-level reminder interval of the distance interval threshold is 30-45cm, which triggers a moderate warning. The third-level safety interval of the distance interval threshold is 45-60cm, which only provides a slight warning.

6. The distance cues method based on laptop light sensing according to claim 1, characterized in that: The visual cues in step S4 use a dynamic light strip design at the edge of the screen. The color of the light strip is red, orange, and green depending on the distance range. The auditory cues are 3D spatial sound effects with a frequency of 200-800Hz. The tactile cues are provided by a piezoelectric ceramic vibrator below the keyboard, with a vibration frequency of 50-200Hz.

7. The distance cues method based on laptop light sensing according to claim 1, characterized in that: The adaptive adjustment algorithm in step S5 adopts a deep reinforcement learning framework. The deep reinforcement learning framework establishes a user preference model by recording the distance parameters, ambient lighting conditions and application scenario types that the user manually adjusts each time. The preference model automatically updates the model parameters once a week.

8. The distance cues method based on laptop light sensing according to claim 1, characterized in that: When the privacy protection mechanism in step S4 detects sensitive scenarios such as video conferencing or online examinations, it automatically switches to a hidden prompt mode, changing the visual prompt to an inconspicuous flashing icon in the corner of the screen and the auditory prompt to bone conduction vibration that can only be perceived by the user, while suspending all data upload functions.

9. The distance cues method based on laptop light sensing according to claim 1, characterized in that: In step S2, a child protection mechanism is set up during information extraction. After identifying a child user through facial feature analysis and skeletal key point detection by the camera, the child protection mechanism automatically activates a strict prompting strategy. The strict prompting strategy includes increasing the warning distance threshold by 10cm and the prompting frequency by 50%, and locking the system settings to prevent children from turning off the prompting function on their own.

10. The distance cues method based on laptop light sensing according to claim 1, characterized in that: In the environmental compensation step of step S6, when considering the influence of temperature data on the distance measurement results, a temperature-distance compensation model is established simultaneously. The temperature-distance compensation model correlates temperature changes with changes in the optical performance of the light field imaging device.

Images