Visual early warning equipment and early warning method for hearing-impaired people

Abstract

The invention discloses a visual early warning device and early warning method for hearing-impaired people, and the method comprises the steps: presetting a detection range with a user as a center, recognizing a moving object through a millimeter wave radar, and obtaining the motion parameters of the moving object; acquiring a sound signal of a moving object through a microphone array; performing sound type recognition on the sound signal to obtain the sound type and the voice content of the sound signal; constructing a risk assessment model; acquiring a danger level and related information of the moving object; and converting the danger level and the related information into image information and displaying the image information in projection glasses of the user. According to the invention, real-time moving object detection is carried out on the blind area of the back of the hearing-impaired person through the wearable micro radar; and sound signals in the direction of the moving object are collected through the microphone. According to the invention, intelligent guarantee of travel safety of hearing-impaired people is realized, dangerous target information is directly visualized in the projection glasses, active early warning is realized, and the hearing-impaired people can be helped to avoid dangers in time.

Description

Technical Field

[0001] This application belongs to the field of electronic information equipment technology, specifically relating to a visual early warning device and early warning method for hearing-impaired individuals. Background Technology

[0002] Hearing loss is a widespread health problem caused by a variety of factors. The World Health Organization estimates that over 400 million people worldwide suffer from disabling hearing loss. The 2021 World Hearing Report predicts that, driven by global demographic trends and increasing risk factors, nearly a quarter of the world's population will have some degree of hearing loss by 2050. Currently, common assistive hearing devices include hearing aids and cochlear implants. Neither of these devices can guarantee complete hearing restoration for those with hearing impairments. This means they may not be able to promptly perceive traffic warning signals such as horns during daily activities, posing a significant safety hazard.

[0003] Currently, there are very few products focusing on the travel safety of hearing-impaired individuals. One national invention patent, "A Sound Recognition Method and Light-Indicating Glasses to Assist Deaf-Mute Travelers" (application number: 202211739426.X), proposes a sound recognition method for assisting deaf-mute individuals in traveling using light-indicating glasses, aiming to address the difficulties and safety concerns associated with travel for the hearing-impaired. Another national invention patent, "Assistive Method, Device, Wearable Device, and Computer-Readable Storage Medium" (application number: 202310290389.7), proposes an assistive method for wearable devices equipped with sound acquisition devices and displays, aiming to ensure that hearing-impaired individuals can accurately receive safety alarm information, thereby guaranteeing their personal safety and improving user experience. Existing technologies and products rely solely on sound signals for safety warnings, resulting in a high probability of false alarms and an inability to warn of dangers in scenarios without sound prompts. The lack of accurate distance information also fails to provide sufficient reaction time for hearing-impaired individuals. Summary of the Invention

[0004] This application provides a visual early warning device and method for hearing-impaired individuals, in order to solve the technical problem that existing technologies and products rely solely on sound signals for safety warnings, and the lack of accurate distance information results in insufficient reaction time for users.

[0005] To address the aforementioned technical problems, this application adopts the following technical solution: a visual early warning method for hearing-impaired individuals, comprising the following steps:

[0006] The detection range is preset with the user at the center. The moving object is identified by millimeter-wave radar and its motion parameters are obtained. The sound signal of the moving object is obtained by microphone array.

[0007] Based on the speech recognition model and the sound signal, the sound type of the sound signal is identified to obtain the sound type and speech content of the sound signal;

[0008] A hazard assessment model is constructed based on motion parameters, sound type, and speech content;

[0009] Based on the hazard assessment model, the hazard level and related information of moving objects are obtained;

[0010] The hazard level and related information are converted into image information and displayed in the user's projection glasses.

[0011] Furthermore, methods for obtaining motion parameters of a moving object include:

[0012] Millimeter-wave radar emits electromagnetic waves in all directions and detects the presence of targets by reflecting the waves.

[0013] In response to the detection of a target, the first distance between the target and the millimeter-wave radar is calculated;

[0014] Based on the millimeter-wave radar re-emitting electromagnetic waves toward the target, the second distance between the target and the millimeter-wave radar is calculated to determine whether the target is a moving object;

[0015] In response to the target being a moving object, the relative speed and relative distance between the moving object and the user are obtained through a first distance and a second distance.

[0016] Furthermore, methods for acquiring the sound types and speech content of sound signals include:

[0017] Based on the microphone arrays located directly in front of, to the left rear, and to the right rear of the user, and the relative direction of the moving object, the microphone array closest to the moving object is selected to identify the sound signal.

[0018] Determine whether the sound signal reaches the threshold value. If the sound signal reaches the threshold value, then use the speech recognition model to identify the sound type and speech content of the sound signal.

[0019] Furthermore, methods for constructing hazard assessment models include:

[0020] Determine four moving object feature data from the input data; among which, the moving object features include relative distance, relative speed, sound type, and speech content;

[0021] Cluster analysis is performed on the feature data of moving objects to construct a traffic scene hazard assessment model based on decision trees.

[0022] Furthermore, methods for constructing traffic scenario hazard assessment models based on decision trees include:

[0023] Based on relative velocity as the root node of the decision tree model, the Gini coefficients of three features—relative distance, relative velocity, sound type, and speech content—are calculated to determine the internal and leaf nodes of the decision tree model.

[0024] The Gini coefficient is shown in formula (1):

[0025]

[0026] Where p represents k The probability of occurrence of a certain category.

[0027] Furthermore, methods for clustering analysis of feature data of moving objects include:

[0028] Based on the relative distance, which includes three categories: near distance, medium distance, and long distance, the category of relative distance is obtained;

[0029] The relative speed is categorized into four types: slow, medium, fast, and ultra-fast.

[0030] Based on the sound types, including silence, human voice, bicycle bell, electric vehicle horn, and car horn, the sound types are categorized.

[0031] The voice content is categorized into two types: those with cue words and those without.

[0032] Furthermore, methods for converting hazard levels and related information into image information include:

[0033] If the hazard level is safe, no hazard warning will be included in the image information;

[0034] If the hazard level is low, the image information will display a blue hazard warning;

[0035] If the hazard level is medium, the image information will display a yellow hazard warning;

[0036] If the risk level is high, the image information will display a red danger warning.

[0037] One technical solution adopted in this application is: a visual early warning device for hearing-impaired individuals, comprising:

[0038] Projection glasses, designed to be worn in front of the user's eyes;

[0039] Millimeter-wave radar, located on the side of the projection glasses, is used to emit electromagnetic waves in all directions centered on the user and identify moving objects.

[0040] A microphone array is located inside the projection glasses; the microphone array is used to identify sound signals from directly in front of the user, to the left rear, or to the right rear.

[0041] The projection beam splitter is located on one side of the front of the projection glasses. It is used to project display information onto the human eye's retina and superimpose it with the real image, so as to issue a danger warning to the user without obstructing the user's view.

[0042] The main control component is connected to the millimeter-wave radar, microphone array, and projection beam splitter. The main control component is used to receive sound signals and reflected electromagnetic waves, calculate the danger level of moving objects and related information, and convert them into corresponding display information.

[0043] The power supply component is connected to the main control component, and is used to provide power to other components.

[0044] The microprocessor armband is worn on the user's arm and is used to store the main control components and power supply components.

[0045] Furthermore, the microphone array includes six microphones, with two microphones per group, and each group of microphones is positioned in front of the projection glasses, to the left rear, and to the right rear, respectively.

[0046] The beneficial effects of this application are as follows: This application uses wearable miniature radar to detect moving objects in the blind spot behind the hearing-impaired person in real time; it collects sound signals from the direction of the moving object through a microphone; it constructs a traffic scene hazard warning sound recognition network model to classify and recognize the collected sound signals; simultaneously, when human voices are recognized, it constructs a hazard warning word speech recognition model to recognize hazard warning words; it constructs a traffic scene hazard assessment model to assess the hazard level of moving objects; based on the assessment results, it issues an active warning to the hearing-impaired person through projection glasses when there is a hazard, and intuitively displays the specific hazard information to help them better avoid the hazard; it realizes intelligent protection of the travel safety of the hearing-impaired person, directly visualizing the hazard target information in the projection glasses, actively warning, and helping the hearing-impaired person to avoid the hazard in a timely and quick manner. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating an embodiment of the visual early warning method for hearing-impaired individuals according to this application.

[0048] Figure 2 yes Figure 1 A flowchart illustrating step S1 of an embodiment;

[0049] Figure 3 yes Figure 1 A flowchart illustrating step S2 of an embodiment;

[0050] Figure 4 yes Figure 1 A flowchart illustrating step S3 in one embodiment;

[0051] Figure 5 This is a schematic diagram of a traffic scene hazard level evaluation decision tree structure, which is an embodiment of the visual early warning method for hearing-impaired persons according to this application.

[0052] Figure 6 This is a schematic diagram of the structure of an embodiment of the visual early warning device for hearing-impaired persons according to this application;

[0053] Figure 7 This is a schematic diagram illustrating the dangerous target identification and division area and the sound signal collection of an embodiment of the visual early warning device for hearing-impaired persons according to this application;

[0054] Figure 8 yes Figure 6 A schematic diagram of the structure of one embodiment of the projection glasses;

[0055] Figure 9 yes Figure 6 A schematic diagram of the structure of a microprocessor arm package in one embodiment. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments.

[0057] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0058] See Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the visual early warning method for hearing-impaired individuals according to this application. The method includes the following steps:

[0059] Step S1. Preset the detection range centered on the user, identify moving objects using millimeter-wave radar, and obtain the motion parameters of the moving objects; obtain the sound signals of the moving objects using a microphone array.

[0060] For details, please refer to Figure 2 , Figure 2 yes Figure 1 A flowchart illustrating an embodiment of step S1. Step S1 includes:

[0061] Step S11. Based on the millimeter-wave radar, electromagnetic waves are emitted in all directions, and the presence of a target is detected by the reflected waves.

[0062] Step S12. In response to the detection of a target, calculate the first distance between the target and the millimeter-wave radar.

[0063] Specifically, after detecting the reflected wave, based on the Time-of-Flight (TOF) principle, the first distance between the target and the user is calculated using the time difference between the reflected and transmitted waves. The distance calculation formula is as follows:

[0064] s = c × t / 2;

[0065] Where s is the distance between the target and the user, t is the time it takes for the electromagnetic wave to travel from the radar emission point to the target echo reception point, and c is the speed of light.

[0066] Step S13. Based on the millimeter-wave radar, transmit electromagnetic waves to the target again, calculate the second distance between the target and the millimeter-wave radar, and determine whether the target is a moving object;

[0067] Specifically, the distance calculated from two consecutive electromagnetic wave transmissions determines whether the target object has moved. If it is moving, the second distance calculated from the second transmitted wave is transmitted to the processor. After electromagnetic waves are emitted by the millimeter-wave radar's transmitting antenna, they are reflected by the parallel receiving antennas of the millimeter-wave radar upon encountering a moving object. By calculating the phase difference between the reflected electromagnetic waves from the same target, the azimuth angle of the moving object is calculated and transmitted to the terminal. The relationship between phase change and distance is as follows:

[0068]

[0069] Where Δφ is the phase difference, Δd is the differential distance from the target to each antenna, and λ is the millimeter-wave radar wavelength. Assuming the distance between antennas is d, then: Δd = l sin(θ), therefore the angle of arrival θ can be estimated as:

[0070]

[0071] Step S14. In response to the target being a moving object, obtain the relative speed and relative distance between the moving object and the user through a first distance and a second distance.

[0072] Specifically, based on the Doppler principle, when there is relative movement between the emitted electromagnetic wave and the detected moving target, the frequency of the emitted wave will differ from that of the reflected wave. By calculating the frequency difference between the reflected and emitted waves, the radial velocity of the moving object relative to the millimeter-wave radar can be measured and transmitted to the terminal. The velocity calculation formula is as follows:

[0073]

[0074] Where v represents the speed of the moving object; f r It is the received echo frequency, reflecting the original pulse frequency emitted by the millimeter-wave radar; fs is the center frequency of the millimeter-wave radar transmission, and c is the speed of light.

[0075] Step S2. Based on the speech recognition model and the sound signal, identify the sound type of the sound signal to obtain the sound type and speech content of the sound signal.

[0076] For details, please refer to Figure 3 , Figure 3 yes Figure 1 A flowchart illustrating step S2 of an embodiment. Step S2 includes:

[0077] Step S21. Based on the microphone arrays located directly in front of, to the left rear, and to the right rear of the user and the relative direction of the moving object, select the microphone array closest to the moving object to identify the sound signal.

[0078] Specifically, based on the calculated azimuth angle of the moving target, it is determined that the moving target appears to the user's left or right rear. Based on this location information, the microphone array at the corresponding location is controlled to collect sound signals.

[0079] Step S22. Determine whether the sound signal reaches a threshold value. If the sound signal reaches the threshold value, then identify the sound type and speech content of the sound signal through a speech recognition model.

[0080] Specifically, if no moving target is detected in the area behind the user, the system determines whether the warning sound should be input into the speech recognition model by judging whether the sound collected by the microphones in front exceeds a given threshold, thereby realizing microphone array group control based on millimeter-wave radar detection results.

[0081] Among them, the speech recognition models include a traffic scene hazard warning sound recognition network model and a hazard prompt word speech recognition model.

[0082] The traffic scene hazard warning sound recognition network model is a YAMNet deep network, which adopts the Mobilenet_v1 deep separable convolutional architecture.

[0083] Mobilenet_v1 is a lightweight deep learning network with 28 layers in total, excluding average pooling layers and softmax output layers. It reduces the number of model parameters and computational cost through depthwise separable convolutions. The specific network structure includes: one input layer (224×224×3) with standard convolutions, 13 groups of depthwise separable convolutions totaling 26 layers, one average pooling layer (7×7×1024), and one fully connected layer plus a softmax layer for output.

[0084] By retrieving and collecting datasets of car horns, electric bicycle horns, bicycle bells, and human voices, YAMNet was used for transfer learning to train a model to recognize specific types of audio.

[0085] A speech recognition model for danger warning words was constructed using Tensorflow. Sound datasets of warning words in different environments were collected, and the waveforms in the dataset were represented in the time domain. Then, a short-time Fourier transform was performed to convert the waveforms from time-domain signals to time-frequency-domain signals, converting the waveforms into a spectrogram to display frequency changes over time, which can be represented as a two-dimensional image. The spectrogram image was input into a convolutional neural network to train the model. A Keras preprocessing layer was used to sample the input to accelerate model training, and each pixel in the image was normalized according to the mean and standard deviation of the image. The trained model for danger warning word speech recognition was then validated using a validation set.

[0086] Step S3. Construct a hazard assessment model based on motion parameters, sound type, and speech content.

[0087] For details, please refer to Figure 4 , Figure 4 yes Figure 1 A flowchart of an embodiment of step S3.

[0088] Step S3 includes:

[0089] Step S31. Determine the four moving object feature data of the input data; wherein, the moving object features include relative distance, relative speed, sound type and speech content.

[0090] Step S32. Perform cluster analysis on the feature data of moving objects to construct a traffic scene hazard assessment model based on decision tree.

[0091] Specifically, the relative distance L between the moving object and the user is divided into three categories: short distance (L<5 meters), medium distance (5≤L≤15 meters), and long distance (L>15 meters). The category of the relative distance to the user is determined based on the relative distance measured by the radar.

[0092] The relative speed between a moving object and a user is classified into four categories: slow (<15 km / h), medium (15-25 km / h), fast (25-35 km / h), and ultra-fast (>35 km / h). The user's relative speed category is determined based on the relative speed measured by radar.

[0093] The sound types are divided into five categories: none, human voice, bicycle bell, electric vehicle horn, and car horn. The sound type of moving object is identified based on the speech recognition model.

[0094] Speech recognition content is divided into two categories: with and without prompts for avoiding words; the speech recognition model identifies whether the speech content contains prompts for or without prompts for avoiding words.

[0095] See Figure 5 The hazard assessment model is a decision tree-based traffic scenario hazard assessment model, which includes:

[0096] The decision tree model is based on relative speed as the root node, which has the highest classification ability in comprehensively evaluating the risk as best as possible. The Gini coefficients of three features, namely relative distance, relative speed, sound type and speech content, are calculated to determine the internal and leaf nodes of the decision tree model.

[0097] The Gini coefficient is shown in formula (1):

[0098]

[0099] Where p represents k The probability of occurrence of a certain category.

[0100] Step S4. Based on the hazard assessment model, obtain the hazard level and related information of the moving object.

[0101] Specifically, the system uses a hazard assessment model to determine the hazard level of the user's current traffic situation. The hazard levels are categorized into four levels: safe, low-risk, medium-risk, and high-risk. Each hazard level has a corresponding electronic graphic, symbol, and warning text.

[0102] Step S5. Convert the hazard level and related information into image information and display it in the user's projection glasses.

[0103] Specifically, if the danger level is "safe", no danger warning will be given;

[0104] If the traffic hazard level is "low risk" or above, the image information will display information such as the type of item, voice prompts, and the direction of oncoming traffic.

[0105] Specifically, different colored exclamation mark graphics are used to display different levels of danger: blue for "low risk," yellow for "medium risk," and red for "high risk." These colors with varying contrasts provide different levels of visual stimulation, with stronger stimulation at higher danger levels, thus serving as an early warning system for different levels of danger.

[0106] See Figure 6 , Figure 6 This application discloses a visual early warning device for hearing-impaired individuals, comprising: projection glasses 1, millimeter-wave radar 2, microphone array 3, projection beam splitter 4, main control component 5, power supply component 6, and microprocessor arm bag 7.

[0107] See Figure 8 The projection glasses 1 are worn in front of the user's eyes to project dangerous information of the current traffic scene onto the lens, thus making it easier for people with hearing impairments to perceive the traffic environment on site through vision.

[0108] The millimeter-wave radar 2 is located on the side of the projection glasses 1. The millimeter-wave radar 2 is used to emit electromagnetic waves in all directions centered on the user and to identify moving objects.

[0109] Microphone array 3 is located within projection glasses 1; wherein, microphone array 3 is used to identify sound signals from directly in front of, to the left rear, or to the right rear of the user. In this embodiment, microphone array 3 comprises six microphones. The grouping and embedding positions of microphone array 3 are as follows... Figure 7 As shown. There are six microphones, divided into three groups of two. Figure 7 As shown, the first group (a1, a2) is embedded in the slot at the upper front edge of the projection glasses 1, responsible for collecting sound signals in front of the user. The second group (c1, c2) is embedded in the slot slightly behind the left side of the projection glasses 1, responsible for collecting sound signals from the user's left rear. The third group (b1, b2) is embedded in the slot slightly behind the right side of the projection glasses 1, responsible for collecting sound information from the user's right rear.

[0110] The projection beam splitter 4 is located on one side of the front of the projection glasses 1. The projection beam splitter 4 projects display information onto the retina of the human eye, superimposing it with the real image to issue a danger warning to the user without obstructing their view. The projection beam splitter 4 is located in the projection device housing in front of the glasses. When worn, a square beam splitter prism is fixed directly in front of the eyes for imaging. A hollow square box is connected to the right side of the prism, with a slot inside to fix the display screen. The display screen is connected to the main control component 5 to receive and transmit image information. The image transmitted from the main control board is projected onto the beam splitter prism in a miniature projector manner for easy viewing by the user.

[0111] The main control component 5 is connected to the millimeter-wave radar 2, the microphone array 3, and the projection beam splitter 4. The main control component 5 is used to receive sound signals and reflected electromagnetic waves, calculate the danger level of moving objects and related information, and convert them into corresponding display information.

[0112] The power supply component 6 is connected to the main control component 5, and the power supply component 6 is used to provide power to other components.

[0113] See Figure 9 The microprocessor armband 7 is worn on the user's arm and houses the main control component 5 and the power supply component 6. The main control component 5 in the microprocessor armband 7 is connected to the various components in the projection glasses 1 via a communication cable.

[0114] This application utilizes wearable miniature radar to detect moving objects in the blind spot behind hearing-impaired individuals in real time; collects sound signals from the direction of moving objects via a microphone; constructs a traffic scene hazard warning sound recognition network model to classify and recognize the collected sound signals; simultaneously, when human voices are detected, it constructs a hazard warning word speech recognition model to identify hazard warning words; constructs a traffic scene hazard assessment model to assess the hazard level of moving objects; based on the assessment results, it issues an active warning to the hearing-impaired individual through projection glasses 1 when danger exists, and intuitively displays specific hazard information to help them better avoid danger; thus achieving intelligent protection of the travel safety of hearing-impaired individuals, directly visualizing hazard target information in projection glasses 1, providing active warnings, and helping hearing-impaired individuals to avoid danger in a timely and rapid manner.

[0115] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A visual early warning method for hearing-impaired people, characterized in that: The following steps are involved: The detection range is preset with the user as the center, and the moving object is identified by the millimeter wave radar to obtain the motion parameters of the moving object; Acquiring a sound signal of the moving object through a microphone array; Based on the speech recognition model and the sound signal, identify the sound type of the sound signal to obtain the sound type and speech content of the sound signal; constructing a risk assessment model based on the motion parameters, the sound type, and the speech content; Based on the risk assessment model, obtaining the risk level and related information of the moving object; The danger level and the related information are converted into image information and displayed in the projection glasses of the user.

2. The method according to claim 1, characterized in that The method for obtaining the motion parameters of the moving object comprises: Based on the millimeter wave radar, electromagnetic waves are emitted to the surroundings, and the reflected waves are used to detect whether there is a target; In response to detecting the target, calculating a first distance between the target and the millimeter-wave radar; Based on the millimeter-wave radar transmitting electromagnetic waves to the target again, calculating a second distance between the target and the millimeter-wave radar, and determining whether the target is a moving object; In response to the target being a moving object, a relative speed and a relative distance between the moving object and the user are acquired through the first distance and the second distance.

3. The method according to claim 2, characterized in that The method for obtaining the sound type and voice content of the sound signal includes: Based on the microphone arrays located in front of, behind the left and behind the right of the user and the relative direction of the moving object, selecting the microphone array closest to the moving object to recognize the sound signal; It is determined whether the sound signal reaches a threshold value. If the sound signal reaches the threshold value, the sound type and speech content of the sound signal are identified through a speech recognition model.

4. The method according to claim 3, characterized in that: Methods for constructing risk assessment models include: Determine four moving object feature data of the input data; wherein the moving object features include the relative distance, the relative speed, the sound type and the voice content; Cluster analysis is performed on the moving object feature data to construct a traffic scene hazard assessment model based on a decision tree.

5. The method according to claim 4, characterized in that The method for constructing a traffic scene hazard assessment model based on a decision tree includes: Based on the relative speed being the root node of the decision tree model, the Gini coefficients of the three features of the relative distance, the relative speed, the sound type and the voice content are calculated to determine the internal nodes and leaf nodes of the decision tree model; The Gini coefficient is shown in formula (1): Among them, it means p k The probability of a certain category occurring.

6. The method according to claim 4, characterized in that The method for clustering analysis of the moving object feature data comprises: Based on the fact that the relative distance includes three categories: short distance, medium distance and long distance, obtaining the category of the relative distance; Based on the fact that the relative speed includes four categories: slow, medium, fast and super fast, obtaining the category of the relative speed; Based on the fact that the sound types include silence, human voice, bicycle bell, battery vehicle horn and car horn, obtaining the category of the sound type; Based on the fact that the voice content includes two categories, namely, prompt avoidance words and non-prompt avoidance words, the category of the voice content is acquired.

7. The method according to claim 4, characterized in that The method of converting the danger level and the related information into image information comprises: If the danger level is safe, no danger prompt is given in the image information; If the danger level is low, the image information displays a blue danger prompt; If the danger level is medium, the image information displays a yellow danger warning; If the danger level is high, the image information displays a red danger prompt.

8. A visual warning device for the hearing-impaired, characterized in that: include: Projection glasses, which are worn in front of the user's eyes; A millimeter wave radar is located at the side of the projection glasses, wherein the millimeter wave radar is used to emit electromagnetic waves around the user and identify moving objects; A microphone array is located in the projection glasses; wherein the microphone array is used to identify sound signals in front of, behind the left or behind the right of the user; A projection spectroscope, located in front of one side of the projection glasses, wherein the projection spectroscope is used to project the display information to be directly reflected into the retina of the human eye, superimposed with the real image, and issue a danger warning to the user without blocking the user's line of sight; A main control component is connected to the millimeter wave radar, the microphone array and the projection spectroscope, wherein the main control component is used to receive the sound signal and the reflected electromagnetic wave, calculate the danger level and related information of the moving object and convert them into the corresponding display information; A power supply component, connected to the main control component, wherein the power supply component is used to provide power to other components; The microprocessor arm bag is worn on the user's arm, wherein the microprocessor arm bag is used to store the main control component and the power supply component.

9. The early warning device according to claim 8, characterized in that: The microphone array includes six microphones, wherein two microphones form a group, and each group of microphones is respectively arranged in front of, behind and behind the projection glasses.

Images