Fall prevention warning method, system and terminal device based on environmental visual recognition

Abstract

The application discloses a kind of based on environmental visual identification's anti-falling water early warning method, system and terminal equipment, belong to terminal equipment safety protection technical field.The method includes: when terminal equipment is in use state, the visual information of surrounding environment is collected by low-power mode;The visual information is input to the lightweight artificial intelligence model built-in to equipment, to identify whether there is water risk area near the equipment;When the water risk area is identified and combined with the device state to determine that there is a risk of falling into water, before the terminal equipment and water occur physical contact, trigger strong interruptive warning operation to prompt user.The application changes passive waterproof to active early warning, through environmental visual identification and end side AI decision, significantly advance safety protection node, effectively solve the problem that prior art cannot intervene before device falls into water, with early warning timely, high accuracy, low power consumption and low implementation cost advantage.

Description

Technical Field

[0001] This invention relates to the field of terminal device security protection technology, specifically to a method, system, and terminal device for preventing water fallout warning based on environmental visual recognition. Background Technology

[0002] Portable devices such as smartphones and tablets have become deeply integrated into daily life and work. However, it is very common for them to be damaged by accidentally slipping into water (such as toilets, basins, pools, ponds, etc.), which brings economic losses and data loss risks to users.

[0003] Currently, protection solutions for terminal equipment damaged by water fall into two main categories:

[0004] The first category is passive protection solutions, which mainly refer to improving the waterproof performance of the equipment itself through physical means, such as achieving IP67 or IP68 dustproof and waterproof standards. This type of solution is essentially a "remedial" measure; its protective mechanism only works when the equipment has already come into contact with water, and it cannot prevent the water-falling event itself. Furthermore, high-level waterproof designs usually come with higher hardware costs, and their sealing performance may gradually decline with increased equipment usage time, component aging, or external impacts, resulting in a decrease in reliability over time.

[0005] The second type is post-contact warning solutions, such as integrating liquid contact sensors (e.g., humidity sensor contacts) into the device. This type of solution can trigger an alarm when it detects liquid contact with the device. However, its warning point begins the moment the device is "waterlogged," making it a typical "post-incident alarm." For most precision electronic devices, once water ingress occurs, damage such as short circuits can happen in a very short time. At this point, the alarm is often too late to prevent further damage.

[0006] In summary, the core concepts of existing technical solutions all focus on protection or alarms "after the equipment comes into contact with water," which is a passive response. The industry lacks a proactive intervention solution that can actively identify the risk of falling into water and provide effective early warnings before the terminal equipment even makes physical contact with the water. Therefore, there is an urgent need for a technical solution that can move the safety protection point forward, achieving true "fall-in-water prevention" through risk prediction, to fill this technological gap. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention discloses a method, system, and terminal device for preventing water damage based on environmental visual recognition, which solves the problems of passive and delayed protection schemes for terminal devices against water damage in existing technologies.

[0008] This invention is achieved through the following technical solution:

[0009] To achieve the above objectives, the present invention first provides a method for preventing water fall into a terminal device, characterized in that the method includes the following steps: acquiring environmental visual information around the terminal device; identifying a water body area near the terminal device based on the environmental visual information; and triggering an early warning operation before the terminal device makes physical contact with the water body when the water body area is identified and it is determined that there is a risk of water fall into the water.

[0010] The core concept of this invention lies in proactively "anticipating" dangerous environments (i.e., water areas) that may cause drowning accidents through the terminal's visual perception capabilities, and taking high-priority intervention measures to warn users before the risk is about to turn into actual loss (i.e., physical contact). This process changes the passive protection logic that relies on physical seals or liquid contact detection.

[0011] In a preferred embodiment of this method, the environmental visual information can be acquired through the camera of the terminal device. To ensure the feasibility of continuous or frequent monitoring, the acquisition process can operate in a low-power mode, for example, by acquiring low-resolution images, obtaining image thumbnails, reducing the acquisition frame rate, or utilizing frame data from the camera preview stream, thereby effectively mitigating the impact on device battery life.

[0012] In a preferred embodiment of this method, the process of identifying water bodies can be accomplished using an artificial intelligence recognition model running on the device. This model is configured to analyze the input environmental visual information and output a judgment result indicating whether a water body exists and its type. The water body particularly includes common, high-risk micro-scenes in daily life, such as toilets, basins, pools, and sinks.

[0013] As a further optimization of this method, when assessing the risk of falling into water, data from other sensors besides visual information can be incorporated, such as the motion posture information of the terminal device (obtained through an accelerometer and gyroscope) and / or environmental humidity information. By fusing multi-dimensional information for comprehensive judgment, the accuracy of risk assessment can be effectively improved, and the false alarm rate reduced. For example, when a water area is visually detected, and the motion posture information indicates that the device is undergoing rapid displacement or tilting at a specific angle, it is determined to be a high-risk state.

[0014] In a preferred embodiment of this method, the triggering of the warning operation refers to activating a high-priority prompt that strongly attracts the user's attention and may interrupt the current operation. Specifically, it may include at least one of the following methods: controlling the terminal device to play a rapid alarm sound at maximum or near maximum volume; triggering a high-intensity, long-duration vibration; or displaying a visual warning pop-up window or full-screen warning interface with a large coverage area and bright colors on the top layer of the screen.

[0015] Corresponding to the above method, the present invention also provides a water-prevention early warning system for a terminal device, which is integrated into the terminal device. The system includes an environmental perception module, a risk decision module, and an early warning execution module. The environmental perception module is used to acquire environmental visual information; the risk decision module is used to identify water areas based on the environmental visual information and perform risk assessment, generating an early warning command when it determines that there is a risk of falling into the water; the early warning execution module is used to respond to the early warning command and execute the early warning operation.

[0016] The present invention also provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the terminal device anti-drowning warning method as described above.

[0017] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the terminal device anti-drowning warning method as described above.

[0018] The beneficial effects of this invention are as follows:

[0019] This invention revolutionizes the concept of protection, transforming the traditional "waterproof" (resisting water ingress) and "water ingress alarm" (responding after contact) into "water-proof" (intervening before contact), eliminating potential safety hazards before they occur through proactive environmental perception.

[0020] The invention provides a fundamentally earlier warning time, triggering the warning at the moment of risk identification before the equipment comes into physical contact with the water, thus giving users critical reaction time and effectively preventing accidents.

[0021] This invention balances accuracy and practicality, utilizing artificial intelligence for visual recognition, achieving high accuracy in identifying water bodies, especially small indoor water bodies; and through a low-power design strategy, this active monitoring function can operate sustainably on ordinary mobile devices for extended periods.

[0022] This invention provides a direct and effective user experience by employing the highest priority strong reminder method to ensure that the warning is immediately perceived by the user and that the intervention measures are decisive and powerful. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating the overall process of the terminal device anti-water-falling early warning method provided in an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram of the module structure of the terminal device anti-water-falling early warning system provided in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] In one embodiment, a method for providing anti-water-falling early warning for a terminal device is provided, the overall process of which can be referred to Figure 1 As shown. This method operates on mobile devices such as smartphones and tablets, and its core lies in using the device's own visual capabilities to proactively "anticipate" dangers.

[0028] Step S1: Service activation and status monitoring.

[0029] Once the terminal device (such as a mobile phone) is powered on, the system automatically starts a data collection service in the background responsible for preventing water damage and providing early warnings. This service continues to run in the background and monitors the real-time usage status of the device.

[0030] When the device is in a "used" state (screen on and user interaction present, such as touch operation or an application running in the foreground), it is assumed that the user is holding or operating the device. In this state, the risk of accidental drop is high, so the process proceeds to the next step (S2) to begin data collection. Conversely, when the device is in a "standby" state (screen locked or no operation), it is assumed that the device is properly placed (e.g., on a table or in a pocket), and the risk of water damage is extremely low. To conserve power, the service pauses data collection until the device re-enters the "used" state. This dynamic control based on usage status is one of the key design features of this solution for achieving low-power operation.

[0031] Step S2: Environmental data collection.

[0032] When the device is in "in use" mode, the system begins silently collecting multimodal data about the surrounding environment. The entire acquisition process prioritizes low power consumption and includes:

[0033] Image data acquisition: The device's rear or front camera is used, but instead of high-resolution recording or photography, a specially optimized low-power image acquisition module (which can be a background app or system service) is employed for acquisition. This module controls the camera to operate at low resolution (e.g., 480p or lower) or directly acquires thumbnails from the camera preview stream. This approach avoids the enormous computational and energy burden of processing full-resolution video streams and is the core invention for resolving the conflict between continuous monitoring and device battery life.

[0034] Ambient humidity data acquisition: Simultaneously, the device's built-in humidity sensor is used to acquire the ambient humidity value around the device. This data will be used as supplementary evidence.

[0035] Location data acquisition (optional): In some implementations, a GPS or Wi-Fi positioning module can be invoked to obtain the device's approximate geographical location information. This information is mainly used to help determine whether the device is located in the vicinity of a large body of water (such as a lake, ocean, river, or pond).

[0036] Step S3: AI Risk Analysis.

[0037] The multimodal data, especially image data, collected in step S2 is input into the AI ​​agent (i.e., a lightweight neural network model) built into the terminal device for real-time analysis. This model is deployed on the device side and is pre-trained specifically for identifying water-related risk features.

[0038] Computer vision analysis is used to analyze image data and identify the presence of water surface features in the image. This model is specifically trained to identify macroscopic water bodies, including lakes, rivers, oceans, and swimming pools, as well as microscopic water risk areas frequently encountered in daily life, such as sinks, basins, toilets, and water cups. This highlights the fundamental difference between this invention and applications focused on macroscopic environmental recognition.

[0039] Simultaneously, environmental humidity data is analyzed. If the humidity sensor reading suddenly increases or remains consistently high, this is used as supplementary evidence that the device may be near a body of water, corroborating the visual analysis conclusions and improving the accuracy of the judgment.

[0040] Step S4: Risk decision-making and early warning triggering.

[0041] The AI ​​agent (i.e., the risk decision-making module) conducts a comprehensive risk assessment and makes a decision based on the analysis results of step S3.

[0042] If the device is detected to be in or very close to the aforementioned aquatic environment (for example, image analysis identifies that the device's camera is facing a toilet or basin, and the water occupies a prominent position in the image), and combined with data analysis (such as the device's posture potentially indicating a slippage) it is assessed that there is a high risk of falling into the water, then a risk is determined to exist, and an early warning mechanism is immediately triggered.

[0043] If no risky water area is identified or the risk level is low, the system returns to step S2 and continues to cycle through low-power monitoring.

[0044] Step S5: User warning notification.

[0045] Once the warning mechanism is triggered, the system immediately issues a strong alert to the user through the device's human-machine interface. This step is implemented by the warning execution module (e.g., by calling the system's built-in alarm clock app or underlying warning interface). The warning method aims to instantly alert the user in a high-priority, highly disruptive manner, specifically including at least one of the following methods:

[0046] Start the alarm: The control device emits a rapid, loud alarm sound.

[0047] Visual warning: Display a full-screen or pop-up warning at the top of the screen, accompanied by eye-catching icons and text, such as "Caution! The device is near water, please hold it carefully!"

[0048] Tactile feedback: Initiates strong, continuous equipment vibration.

[0049] These immediate and strong warning methods aim to buy users valuable reaction time in the final moments before an accident occurs, prompting them to hold their equipment firmly and thus enabling effective intervention.

[0050] Corresponding to the above method, this embodiment provides a terminal device and system for implementing anti-fall-in-water early warning, the module structure of which can be referred to the appendix. Figure 2 As shown.

[0051] The terminal device includes the anti-drowning warning system described in this invention, which mainly includes:

[0052] Environmental perception module: This includes hardware such as cameras, positioning modules (such as GPS), and attitude sensors, as well as their drivers. It is responsible for collecting environmental and equipment status data and specifically performs the data acquisition function in step S2 above.

[0053] Low-power image acquisition module: As a software control unit, it is used to specifically control the camera to acquire images with low power parameters (low resolution, thumbnail) in monitoring mode.

[0054] Risk Decision Module: This module has a built-in lightweight AI recognition model, which is responsible for receiving environmental data and performing the analysis and decision-making functions of steps S3 and S4 above, namely analyzing images to identify risk areas in the water body and making a decision on whether to issue a warning.

[0055] Warning Execution Module: This module triggers corresponding warning prompts based on instructions from the risk decision module. In actual implementation, the strong warning in step S5 can be achieved by calling the interface of the system alarm clock APP or by directly controlling the system's audio, vibration, and display components.

[0056] In one embodiment, a bathroom scenario is provided: When a user uses their mobile phone in the bathroom, the phone's front or rear camera inadvertently captures an image of a toilet or a basin of water. The AI ​​model identifies the water feature and, combined with the slight shaking that the phone may experience due to the user's movements, determines that there is a risk of the phone falling. It then immediately triggers a strong vibration and a warning sound to alert the user.

[0057] In one embodiment, a kitchen scenario is provided: a user is looking at a recipe by the kitchen sink, and the phone's camera captures the sink area. The AI ​​model identifies the sink as a risk area. If the phone tilts rapidly due to being wet and slippery, the system combines visual and posture data to determine a high risk and pops up a full-screen visual warning.

[0058] In one embodiment, an outdoor waterside scenario is provided: a user takes photos or plays with their mobile phone by a pond, swimming pool, or river. The positioning module provides location information of nearby water bodies, and the camera simultaneously captures images of the water surface. Once the AI ​​model confirms the presence of water, the system enters a high-alert state. If any throwing or slipping motion of the device is detected (via the attitude sensor), an alarm immediately sounds at maximum volume.

[0059] In summary, this invention shifts the focus of protection from "waterproofing" and "contact alarm" after a body falls into water to "environmental early warning" before it does, achieving a fundamental shift from post-event remediation to pre-event prevention. By utilizing multimodal data (images, humidity, location) and a lightweight end-side AI model for comprehensive analysis, it can accurately identify everyday micro and macro water bodies with a low false alarm rate.

[0060] This invention significantly reduces the power consumption of the system during continuous operation by using technologies such as status monitoring (monitoring only during use) and acquiring low-resolution images, without affecting the normal battery life of the device.

[0061] The warning method of this invention is direct and effective (alarm clock, strong vibration, eye-catching pop-up), which can instantly alert the user, give the user reaction time, and does not affect the use of the device's core functions.

[0062] The solution of this invention is based entirely on the existing sensors (camera, humidity sensor, IMU) and computing power of the terminal device, without adding any new hardware costs, and is easy to integrate and promote.

[0063] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for early warning of falling water based on environmental visual recognition, characterized in that, Running on a terminal device includes the following steps: S1: Service Activation and Status Monitoring: When the terminal device starts up, the background data collection service automatically starts to continuously monitor the device's usage status. When the device screen is on and there is user interaction, it enters S2; when the device screen is locked or there is no operation, data collection is paused. S2: Environmental Data Acquisition: During device use, it silently collects multimodal data of the surrounding environment with low power consumption as the principle, including calling the camera at low resolution or capturing thumbnails to obtain image data, and calling the humidity sensor to obtain humidity data; S3: AI Risk Analysis: The multimodal data collected by S2 is input into the terminal's built-in lightweight AI model for analysis. This model is used to identify water-related risk characteristics, perform visual analysis on images to identify water body characteristics, and combine humidity data to assist in judgment. S4: Risk Decision and Early Warning Trigger: The AI ​​model assesses the risk based on the analysis results. If it identifies that the device is in or near water and there is a risk of falling into the water, it triggers an early warning; if there is no risk, it returns to S2 to continue monitoring. S5: User warning prompt: After the warning is triggered, the user is alerted through the device's human-machine interface.

2. The anti-fall-into-water early warning method based on environmental visual recognition according to claim 1, characterized in that, In step S2, the environmental data collection also includes: calling the positioning module to obtain the device's rough geographical location information to determine whether the device is located in the vicinity of a lake, ocean, river or pond; in the analysis of step S3, the AI ​​model combines the geographical location information to make a comprehensive risk assessment.

3. The anti-drowning early warning method based on environmental visual recognition according to claim 1, characterized in that, In step S2, the low resolution refers to a resolution no higher than 480p; and / or, the camera is a front-facing camera or a rear-facing camera.

4. The anti-fall-into-water early warning method based on environmental visual recognition according to claim 1, characterized in that, In step S3, the water features include water surface, lake, river, ocean, swimming pool, water tank, water basin, toilet or water cup.

5. The anti-drowning early warning method based on environmental visual recognition according to claim 1, characterized in that, In step S3, the use of ambient humidity data as an auxiliary judgment specifically includes: if the ambient humidity value suddenly increases or remains at a high level, it is used as auxiliary evidence of the presence of water.

6. The anti-drowning early warning method based on environmental visual recognition according to claim 1, characterized in that, The triggering of the early warning operation includes at least one of the following methods: The terminal device is controlled to emit an alarm sound; The terminal device is controlled to perform strong vibrations; A visual warning message pops up on the display interface of the terminal device.

7. A method for early warning of falling water based on environmental visual recognition, characterized in that, Running on a terminal device includes the following steps: S1: Service Activation and Status Monitoring: When the terminal device starts up, the background data collection service automatically starts to continuously monitor the device's usage status. When the device screen is on and there is user interaction, it enters S2; when the device screen is locked or there is no operation, data collection is paused. S2: Environmental Data Acquisition: During device use, multimodal data of the surrounding environment is silently collected with low power consumption as the principle. This includes calling the camera to obtain image data at a low resolution of no more than 480p or by capturing thumbnails, calling the humidity sensor to obtain humidity data, and calling the positioning module to obtain the device's rough geographical location information to determine whether the device is in the vicinity of a lake, ocean, river or pond. The camera is a front-facing camera or a rear-facing camera. S3: AI Risk Analysis: The multimodal data collected in S2 is input into the terminal's built-in lightweight AI model for analysis. This model is used to identify water-related risk characteristics, perform visual analysis on images to identify water body features including water surfaces, lakes, rivers, oceans, swimming pools, water tanks, basins, toilets, or water cups, and combine environmental humidity data for auxiliary judgment. The combination of environmental humidity data as auxiliary judgment specifically includes: if the environmental humidity value suddenly increases or remains at a high level, it is used as auxiliary evidence of the presence of water. At the same time, a comprehensive risk judgment is made in conjunction with the geographical location information. S4: Risk Decision and Early Warning Trigger: The AI ​​model assesses the risk based on the analysis results. If it identifies that the device is in or near water and there is a risk of falling into the water, it triggers an early warning; if there is no risk, it returns to S2 to continue monitoring. S5: User Warning Prompt: After the warning is triggered, the user is alerted through the device's human-machine interface. The warning method includes at least one of the following: controlling the terminal device to emit an alarm sound; controlling the terminal device to perform strong vibration; or displaying a visual warning message on the terminal device's display interface.

8. A water-prevention early warning system based on environmental visual recognition, used to implement the water-prevention early warning method based on environmental visual recognition as described in any one of claims 1-7, characterized in that, Integrated into terminal devices, including: The environmental sensing module is used to silently collect multimodal environmental data, including low-resolution environmental images and environmental humidity data, when the device is in use. The risk decision-making module has a built-in lightweight AI model, which is used to receive and analyze the data collected by the environmental perception module, identify water risk characteristics and conduct risk assessment. When it is determined that there is a risk of falling into the water, it outputs an early warning command. The early warning execution module is used to execute early warning operations through the human-machine interface of the terminal device according to the early warning instructions output by the risk decision module; the environmental perception module includes a camera, a humidity sensor and a positioning module; the positioning module is used to provide the device's approximate geographical location information.

9. A terminal device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the anti-drowning early warning method based on environmental visual recognition as described in any one of claims 1 to 7.

10. The terminal device according to claim 9, characterized in that, The terminal device is a smartphone or tablet computer.

Images