A data processing method applied to an intelligent bathroom mirror

Abstract

The application discloses a data processing method applied to an intelligent bathroom mirror, relates to the technical field of intelligent bathroom equipment and man-machine interaction, and comprises the following steps: in response to detecting a mirror fogging state, determining a fogging level according to overall attenuation of transmitted light intensity and switching to an optical touch mode; continuously collecting multispectral band transmitted light intensity distribution; identifying a local enhanced region meeting spatial continuity, time persistence and a transmitted light intensity transition proportion in line with a biological tissue scattering characteristic threshold as a transmission window and tracking a motion track of the transmission window; generating a touch pressure level according to an area change rate of the transmission window, an equivalent ellipse length-width ratio change rate and an edge light intensity gradient intensity change rate, and then generating a control instruction. The fog layer is taken as an optical modulation layer, and multispectral biological tissue identification and trajectory predictive partition heating are combined, so that reliable optical touch interaction can be realized under a fogging environment, and thus the misrecognition rate is reduced.

Description

Technical Field

[0001] This invention belongs to the field of intelligent bathroom equipment and human-computer interaction technology, specifically a data processing method applied to intelligent bathroom mirrors. Background Technology

[0002] In bathroom environments, mirror fogging is a thermodynamically inevitable phenomenon. A temperature difference exists between hot water steam and the mirror glass, causing steam molecules to condense and form nuclei on the mirror surface. The fog layer accumulates over time, transforming the mirror from an optically transparent object into an optical scatterer. Existing solutions primarily employ two strategies: passive and active defogging. Passive defogging reduces the adhesion of fog droplets by coating the mirror surface with a hydrophilic or hydrophobic coating. This approach is effective in mild steam environments, but in the consistently high temperature and humidity of a bathroom, the defogging ability rapidly diminishes after the coating becomes saturated. Furthermore, the coating interferes with touch signals, rendering traditional touch control systems inoperable. Active defogging raises the mirror temperature above the dew point using heating wires on the back of the mirror to evaporate the fog layer. This solution is reliable, but continuous heating of the entire mirror surface results in high energy consumption and slow response. Uneven temperature distribution during heating can lead to localized thermal stress, posing a risk of mirror cracking.

[0003] The common flaw of the above solutions is that they treat defogging and touch control as two separate problems and fail to utilize the fog layer itself as a signal carrier for changes in the optical medium properties, thus failing to establish an effective interaction channel in the foggy state. Summary of the Invention

[0004] This invention provides a data processing method for smart bathroom mirrors, aiming to solve the technical problem that traditional touch control solutions fail due to fogging of the mirror surface in high temperature and humidity environments in bathrooms.

[0005] A data processing method for smart bathroom mirrors includes the following steps: S1. In response to detecting that the smart bathroom mirror is in a fogged state, the current fogging level is determined according to the overall attenuation of the current transmitted light intensity of the mirror surface. Different fogging levels correspond to different light intensity shift detection sensitivities, and the smart bathroom mirror is controlled to switch from the conventional interaction mode to the optical touch mode. S2. In the optical touch mode, the transmitted light intensity distribution of the mirror surface is continuously collected; wherein, the transmitted light intensity distribution includes the transmitted light intensity sub-distributions collected in two different spectral bands respectively; S3. In response to the transmitted light intensity distribution, the transmitted light intensity of a local area is enhanced relative to other areas of the same mirror surface, and the enhanced area has spatial continuity and temporal persistence between adjacent acquisition frames. The transmission light intensity transition ratio of the local area in each spectral band falls within a preset biological tissue scattering characteristic threshold range. The threshold range is calibrated based on the difference in transmitted light intensity between skin and water droplets / fabric in the two spectral bands. The enhanced area is identified as a transmission window formed by the removal of the fog layer by external force, and the movement trajectory of the transmission window on the mirror surface is tracked. S4. Obtain the area change rate, equivalent ellipse aspect ratio change rate, and edge light intensity gradient intensity change rate of the transmission window in continuous acquisition frames. Generate touch pressure level based on the area change rate, the equivalent ellipse aspect ratio change rate, and the edge light intensity gradient intensity change rate. S5. Generate control commands based on the morphological characteristics of the motion trajectory and the touch pressure level.

[0006] Optionally, detecting that the smart bathroom mirror is in a fogged state includes: The temperature values ​​of multiple sampling points on the mirror surface and the dew point temperature value of the environment surrounding the mirror surface are obtained; Spatial interpolation is performed based on the temperature values ​​of the multiple sampling points to generate a mirror temperature field distribution; Based on the comparison between the mirror temperature field distribution and the dew point temperature value, a fogging risk gradient map is generated. The fogging risk gradient map is used to guide the spatial differentiation configuration of the light intensity shift detection sensitivity. When the temperature value of any of the sampling points is lower than the dew point temperature value, the mirror surface is determined to be in a fogged state.

[0007] Optionally, identifying the enhanced region as a transmission window formed due to the fog layer being wiped away by an external force includes: The transmitted light intensity distribution of the mirror surface in a fog-free state is obtained as the first reference distribution; The transmitted light intensity distribution of the area of ​​the mirror surface that is not in contact with external force in a fogged state is obtained as a second reference distribution; The transmitted light intensity distribution of the mirror surface under two different spectral bands is obtained, and a first reference distribution and a second reference distribution corresponding to each band are generated respectively. When the transmitted light intensity of the local area transitions from close to the second reference distribution to close to the first reference distribution within a dynamic time window, and the light intensity value after the transition remains stable within consecutive acquisition frames, the local area is identified as the transmission window. The identification also needs to meet the following requirements: the ratio of transmitted light intensity transition in the local area under each spectral band conforms to the preset threshold range of biological tissue scattering characteristics, and the threshold range is experimentally calibrated based on the difference in transmitted light intensity between skin and water droplets / fabric under the two spectral bands. The duration of the dynamic time window is determined by the difference between the second reference distribution and the first reference distribution; the larger the difference, the longer the duration of the dynamic time window.

[0008] Optionally, tracking the motion trajectory of the transmission window on the mirror surface includes: Obtain the geometric center position of the transmission window in the transmitted light intensity distribution of continuously acquired multiple frames; The geometric center positions between adjacent frames are correlated and matched to generate a sequence of the position of the transmission window changing over time; When the displacement between adjacent positions in the sequence falls within a preset speed range, the sequence is retained as the motion trajectory; the preset speed range corresponds to the natural sliding speed range of a human hand on a mirror surface. When the displacement of the geometric center position of the transmission window within a continuous acquisition frame is less than a preset displacement threshold, the area change rate of the transmission window is obtained. When the area change rate presents a pulse pattern of first increasing and then decreasing within a preset time window, it is identified as a tap operation, and short taps and long taps are distinguished according to the pulse width of the pulse pattern. When a tap operation is detected, a tap control command containing pressure information is generated based on the area of ​​the transmission window at the peak of the pulse pattern, the aspect ratio of the equivalent ellipse, and the intensity of the edge light gradient. The larger the area of ​​the transmission window, the closer the aspect ratio of the equivalent ellipse is to 1, and the greater the intensity of the edge light gradient, the higher the corresponding touch pressure level.

[0009] Optionally, when at least two of the transmission windows exist simultaneously, the method further includes: Obtain the erasure width, erasure area change rate, trajectory curvature, erasure speed, and transmitted light intensity recovery degree of each transmission window; The operation dynamics characteristics of the motion trajectory corresponding to each transmission window are obtained. The operation dynamics characteristics include the mean trajectory velocity, velocity variance, acceleration distribution, click interval duration, and statistical moments of trajectory curvature distribution. Based on the erase width, erase area change rate, trajectory curvature, erase speed, and degree of transmission light intensity recovery, as well as the operation dynamics characteristics, each transmission window is clustered to generate at least two operation source groups. The operation dynamics characteristics of each operation source group are compared with a preset user touch behavior feature library, which includes statistical features such as the historical trajectory speed average, tap duration distribution, and trajectory curvature distribution, in order to determine the user identity confidence level corresponding to each operation source group. The motion trajectory of the transmission window corresponding to each operation source group is tracked, and corresponding control commands are generated respectively.

[0010] Optionally, tracking the motion trajectory of the transmission window on the mirror surface further includes: The light intensity attenuation rate of the transmission window in consecutive acquisition frames is obtained, and the light intensity attenuation rate characterizes the rate at which the transmitted light intensity of the transmission window decreases with time. The temperature difference between the local mirror temperature in the area where the transmission window is located and the temperature difference between the surrounding fogging area is obtained. When the temperature difference value gradually converges from above the preset temperature difference threshold to below the preset temperature difference threshold, and the light intensity attenuation rate is lower than the dynamic threshold, it is determined that the transmission window has entered the fading state, the position of the transmission window in the current frame is marked as the trajectory endpoint, and the tracking of the transmission window is stopped. The dynamic threshold is determined by the local temperature of the mirror surface at the location of the transmission window; the lower the local temperature, the smaller the dynamic threshold. In subsequent acquisition frames, in response to the reappearance of enhanced transmitted light intensity at the same spatial location, and the enhanced light intensity value being higher than the light intensity value at the end point of the previous trajectory, the reappearing enhanced area is identified as a new transmission window, and a new motion trajectory is generated.

[0011] Optionally, the method further includes: Obtain the transmitted light intensity distribution in the non-erased area on the mirror surface where the transmission window is not formed; Based on the difference between the transmitted light intensity distribution of the non-erased area and the reference transmitted light intensity distribution of the mirror in a fog-free state, a residual fog layer accumulation distribution map of the mirror is generated. Based on the cumulative distribution map of the residual fog layer, the fog layer thickness level of each region is determined; By conducting calibration experiments with standard touch points at each fog layer thickness level, the measured coordinate offset between the geometric center position of the transmission window and the standard touch coordinates is obtained, and a correspondence table between fog layer thickness level and the measured coordinate offset is established. Based on the cumulative distribution map of the residual fog layer, the identifiable and non-identifiable areas on the mirror surface are determined. The identifiable areas are those where the light intensity difference in the cumulative distribution map of the residual fog layer is lower than a preset threshold. When generating the motion trajectory, only the transmission window located within the identifiable area is tracked; Based on the corresponding relationship table, the geometric center position of the transmission window located in each fog layer thickness level region in the motion trajectory is compensated by superimposing the corresponding measured coordinate offset to generate a corrected touch coordinate sequence.

[0012] Optionally, after clustering each transmission window to generate at least two operational source groups, the method further includes: Obtain the erase width change sequence and light intensity recovery degree change sequence of each transmission window within each operation source group during continuous tracking; Based on the erase width change sequence and the light intensity recovery degree change sequence, calculate the feature consistency coefficient within each operation source group; When the feature consistency coefficient continues to decrease within a preset continuous sampling window and falls below a preset threshold, the operation source group is split into at least two sub-operation source groups; The operational dynamics characteristics of each sub-operation source group are compared with the user touch behavior feature library; When the similarity exceeds a preset matching threshold, the sub-operation source group is bound to the corresponding user identity; When the similarity comparison is lower than the preset visitor threshold, it is marked as a visitor operation source and permission downgrade control is triggered. The motion trajectory of the transmission window corresponding to each sub-operation source group is tracked, and corresponding control commands are generated respectively.

[0013] Optionally, before determining that the transmission window has entered the fading state, the method further includes: The current direction and speed of the transmission window are obtained, and the predicted position of the transmission window after a preset time is predicted based on the current direction and speed of the transmission window. Based on the predicted position, determine the corresponding prediction area on the mirror surface; A heating command is sent to the mirror back heating unit corresponding to the prediction area. The heating command is used to raise the mirror surface temperature of the prediction area to above the dew point temperature before the transmission window reaches the prediction area, so as to prevent the prediction area from being in a fogging state when the transmission window arrives. Obtain the historical heating power sequence of the mirror back heating unit and the fog layer dissipation rate sequence of the corresponding region under each historical heating power, and determine the mapping relationship table between each heating power and the fog layer dissipation rate; Based on the mapping table and the current movement speed of the transmission window, the target heating power required for the predicted area is calculated so that the predicted area is in a fog-free or foggy state when the transmission window arrives, and naturally recovers to a foggy state after the transmission window leaves using thermal inertia.

[0014] Optionally, after sending a heating command to the mirror back heating unit corresponding to the predicted region, the method further includes: Obtain the actual fog layer thickness value of the prediction area when the transmission window reaches the prediction area; Based on the actual fog thickness value, the heating power of the mirror back heating unit is adjusted so that the predicted area returns to the fogged state at a preset cooling rate after the transmission window leaves. The touch pressure level and dwell time of the transmission window within the prediction area are obtained; Based on the touch pressure level and the dwell time, the thermal inertia compensation coefficient in the mapping table is adjusted. The higher the touch pressure level or the longer the dwell time, the larger the thermal inertia compensation coefficient. In subsequent predictions, the target heating power is reduced accordingly based on the lookup result in the mapping table.

[0015] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: This invention uses the fog layer as an optical modulation layer rather than a simple barrier, and utilizes the multispectral transmitted light intensity distribution to detect the transmission window formed after the fog layer is wiped away, thereby realizing optical touch interaction in foggy conditions and solving the problem of complete failure of traditional capacitive / resistive touch in foggy environments. By identifying biological tissue contact using a dual-band transition ratio threshold based on Mie scattering theory, the actual touch is physically separated from environmental interference such as water droplets falling and fabric rubbing, thus significantly reducing the touch misidentification rate in foggy environments. Furthermore, by establishing a joint mapping between the area change rate, aspect ratio change rate, edge gradient change rate, and touch pressure level through Hertzian contact theory and a fluid dynamics model, touch pressure sensing is achieved under conditions without a dedicated pressure sensor, thereby enriching the interaction dimensions in foggy conditions. By using multi-transmission window clustering based on operational dynamics characteristics and identity confidence comparison, operation source separation and hierarchical permission are achieved under multi-user parallel operation, thus supporting multi-user shared interaction in bathroom scenarios. By using trajectory prediction and temporal complementarity of mirror back zone heating, active defogging is transformed from continuous heating of the entire mirror surface to pulse preheating following the touch path. Combined with thermal inertia compensation coefficient, the heating power is dynamically adjusted, thereby reducing system energy consumption and mirror thermal stress while ensuring touch continuity. Through residual fog layer accumulation distribution map and coordinate offset compensation mechanism, touch positioning drift caused by fog layer scattering is eliminated, thereby improving the positioning accuracy of touch coordinates in foggy environments. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and 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.

[0018] In a bathroom environment, mirror fogging is a thermodynamic inevitability. A temperature difference exists between hot water steam and the mirror glass; steam molecules condense and nucleate on the mirror surface, and the fog layer accumulates over time, causing the mirror to degenerate from an optically transparent object into an optical scatterer. Traditional solutions treat this as a purely physical obstacle, resorting to either passive hydrophobic coatings or active full-mirror heating for defogging. The former fails rapidly under saturated steam, while the latter consumes continuous high energy to achieve a fog-free environment, neglecting the potential of the fogged mirror itself to be reconstructed into an optical touch-sensitive medium.

[0019] This application provides a data processing method for smart bathroom mirrors. The method does not treat the fog layer as an enemy to be completely eliminated, but rather transforms it into an optical modulation layer for touch interaction. When the fog layer is uniformly distributed, the transmitted light intensity received by the photoelectric sensor on the back of the mirror forms a stable low-value dark field; when a finger wipes away a localized fog layer, the transmitted light intensity in that area jumps, forming a bright spot. This bright spot is the transmission window, which is the only usable touch signal source in the fogged state.

[0020] The transmitted light intensity distribution refers to the spatial distribution mapping of light power intensity received by the photoelectric sensor array on the back side of the mirror after the light emitted from the light source passes through the mirror glass substrate and the surface fog layer. The transmitted light window refers to the optical phenomenon where, after the fog layer is significantly thinned or even disappeared due to contact and removal by external force such as a finger or palm, the transmitted light intensity is enhanced relative to the surrounding fogged area. The optical touch mode refers to the operating mode in which the system disables the conventional capacitive touch sampling circuit and instead uses the transmitted light intensity distribution acquisition and analysis as the sole input channel. The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. Please refer to... Figure 1 This application provides a data processing method for smart bathroom mirrors, including: In response to the detection that the smart bathroom mirror is fogged up, the method initiates a fogging monitoring process. An array of thermocouple temperature sensors is embedded around the frame of the smart bathroom mirror, distributed in a grid pattern along the edge and center of the mirror surface, with a sampling period set to once per second.

[0021] Specifically, the method reads the voltage signals of each thermocouple through an analog-to-digital converter interface and converts them into temperature values. At the same time, it obtains the relative humidity value of the environment through a humidity sensor and calculates the dew point temperature value of the environment around the mirror based on the Magnus formula. The dew point temperature value refers to the temperature at which air is cooled to saturation under the condition that the water vapor content and air pressure remain constant. Its calculation formula is that the dew point temperature is equal to the ambient dry-bulb temperature minus one-fifth of (100 minus the relative humidity percentage) and then multiplied by a correction factor.

[0022] Furthermore, after obtaining the temperature values ​​of multiple sampling points on the mirror and the dew point temperature of the surrounding environment, the method performs spatial interpolation: using each sampling point as a control point, a bilinear interpolation algorithm is used to fill the two-dimensional plane of the mirror with a grid, generating a mirror temperature field distribution covering the entire mirror area. The data structure of the mirror temperature field distribution is a two-dimensional floating-point array, where the array index corresponds to the mirror pixel coordinates, and the array element value is the interpolated temperature at that coordinate.

[0023] Subsequently, the method generates a fogging risk gradient map based on the comparison between the mirror temperature field distribution and the dew point temperature value. The data structure of the fogging risk gradient map is a two-dimensional byte array with the same resolution as the mirror. Its element value is equal to the difference between the dew point temperature and the mirror temperature at that point, multiplied by a preset scaling factor and then rounded. The larger the value, the higher the fogging risk.

[0024] The fogging risk gradient map is used to guide the spatial differentiation configuration of the light intensity shift detection sensitivity. Specifically, it is implemented through the following mapping rules: the element values ​​at each coordinate position of the fogging risk gradient map are used as the fogging risk index for that position, and the local sensitivity correction coefficient for that coordinate position is calculated according to a piecewise linear mapping relationship; when the fogging risk index is less than 50, the local sensitivity correction coefficient is set to 1.2; when the fogging risk index is between 50 and 150, the local sensitivity correction coefficient is set to 1.0; when the fogging risk index is greater than 150, the local sensitivity correction coefficient is set to 0.6. The fogging risk index is obtained by multiplying the difference between the mirror temperature and the dew point temperature by a preset proportional coefficient and then rounding it down. This preset proportional coefficient is determined based on the thermal conductivity of the mirror material and historical fogging data, and is set to 10 quantization units per degree Celsius. This maps the temperature difference range in a typical bathroom environment to a risk index on the order of 0 to 200. The local sensitivity correction coefficient is multiplied by the global sensitivity coefficient corresponding to the current fogging level to obtain the final light intensity shift detection sensitivity at that coordinate location. This sensitivity is then used as a multiplicative factor on the local light intensity enhancement threshold at that coordinate location. This reduces sensitivity in high-fogging-risk areas to suppress noise and increases sensitivity in low-fogging-risk areas to capture weak touch signals. When the temperature value of any sampling point is lower than the dew point temperature, the mirror is determined to be fogged, the fogging flag in the system status register is set from 0 to 1, and the subsequent optical touch mode switching process is triggered. If the temperature of all sampling points is higher than the dew point temperature, the fogging flag remains at 0, and the system maintains the normal interaction mode. Correspondingly, if the temperature sensor reading interface returns a timeout error or exceeds the range value, the sampling point is marked as an invalid point, and interpolation compensation is performed using the average value of adjacent valid sampling points, thereby ensuring the integrity of the temperature field distribution.

[0025] After determining that the mirror is in a fogged state, the method determines the current fogging level based on the overall attenuation of the current transmitted light intensity of the mirror.

[0026] Specifically, the method calls the global sampling interface of the back-side photoelectric sensor array to obtain the average transmitted light intensity of the mirror surface under conditions without external light source interference, and records it as the current light intensity value; simultaneously, it reads the reference average transmitted light intensity of the mirror surface under a fog-free state from non-volatile memory, and records it as the reference light intensity value. The overall attenuation degree is defined as the normalized attenuation ratio obtained by dividing the difference between the reference light intensity value and the current light intensity value by the reference light intensity value.

[0027] Specifically, the fogging level is determined by the normalized attenuation ratio: when the ratio is less than 0.2, the fogging level is Level 1, with an extremely thin fog layer; when the ratio is between 0.2 and 0.5, the fogging level is Level 2, with a medium fog layer; when the ratio is between 0.5 and 0.8, the fogging level is Level 3, with a relatively thick fog layer; and when the ratio is greater than 0.8, the fogging level is Level 4, with an extremely thick fog layer. The boundary values ​​for each level are assigned to the right-hand interval, i.e., a ratio of 0.2 is assigned to Level 2, 0.5 to Level 3, and 0.8 to Level 4. Different fogging levels correspond to different light intensity shift detection sensitivities: Level 1 fogging corresponds to a sensitivity coefficient of 1.0, Level 2 to 0.7, Level 3 to 0.4, and Level 4 to 0.2. These sensitivity coefficients act as multiplicative factors on the subsequent local light intensity enhancement judgment threshold. Subsequently, the method controls the smart bathroom mirror to switch from the conventional interactive mode to the optical touch mode: a mode switching command is sent to the display control unit to turn off the power supply circuit of the capacitive touch panel to reduce electromagnetic interference; a command is sent to the light source driving unit to turn on the near-infrared light-emitting diode array and the visible light light-emitting diode array on the back of the mirror, wherein the center wavelength of the near-infrared band is 940 nanometers and the center wavelength of the visible light band is 550 nanometers, and the two light sources are alternately lit in a time-division multiplexing manner to avoid spectral crosstalk.

[0028] In the optical touch mode, the transmitted light intensity distribution of the mirror surface is continuously acquired. An array of photodiodes, matching the mirror resolution, is arranged on the back side of the mirror. Each photodiode corresponds to a physical pixel area on the mirror surface, with a spatial resolution of one sampling point per millimeter. The method continuously calls the array read interface at a rate of 60 frames per second to acquire the raw light intensity data matrix for each frame. The transmitted light intensity distribution includes transmitted light intensity sub-distributions acquired in two different spectral bands: during the near-infrared LED illumination period, the array reads the near-infrared band transmitted light intensity sub-distribution; during the visible light LED illumination period, the array reads the visible light band transmitted light intensity sub-distribution. The two sub-distributions are stored in a circular buffer in a dual-channel image frame format, with a buffer depth of 10 frames, to support time-duration analysis. Accordingly, if the photoelectric sensor array read interface returns a frame synchronization error or data verification failure, the frame is discarded and the valid data from the previous frame is reused, thereby ensuring tracking continuity.

[0029] On a fogged mirror, water droplets scatter visible and near-infrared light differently. When biological tissue comes into contact with and removes the fog layer, the local optical boundary conditions are reconstructed, resulting in a surge in light intensity. If only single-band light intensity changes are relied upon, the falling of water droplets or the rubbing of fabric will produce amplified signals similar to those of a finger, leading to an extremely high false positive rate. If image recognition algorithms are used directly to distinguish materials, the interference from water and steam on the mirror surface makes feature extraction highly unstable.

[0030] Meanwhile, human skin exhibits a significantly larger equivalent scattering cross-section in the 940 nm near-infrared band compared to the 550 nm visible light band due to Mie resonance scattering by hemoglobin and collagen fibers. Water droplets, as a homogeneous medium, have a complex refractive index close to 1.33 with a very small imaginary part in both bands, resulting in only a slight difference in the Mie scattering cross-section between the two bands. Although fabric fibers have a higher refractive index, their fiber diameter is much larger than the wavelength of light, leading to predominantly geometric forward scattering. This results in low enhancement amplitudes in both bands, and the ratio deviates from the characteristics of skin. While this difference in scattering cross-section cannot be directly converted into touch coordinates, it can serve as a hard threshold for material identification, helping the system physically separate real touch from environmental interference.

[0031] This application employs a dual-band transition ratio and spatiotemporal continuity to provide dual verification. In response to the transmitted light intensity distribution, when the transmitted light intensity in a local area is enhanced relative to other areas of the same mirror surface, the application simultaneously verifies the spatial continuity and temporal persistence of the enhanced area between adjacent acquisition frames. Furthermore, it utilizes the band differences in Mie scattering to compensate for the material ambiguity inherent in single-band identification. Through this complementary dual-band design, the physical constraints of scattering theory can be leveraged to offset the risk of misjudgment from a single spectrum, avoiding erroneous operations triggered by non-biological contact such as water droplets or fabrics, while simultaneously improving the reliability of biological tissue identification.

[0032] Specifically, the method performs spatial filtering on the transmitted light intensity distribution of the current frame, removes photoelectric noise using a Gaussian smoothing kernel, and calculates the deviation between the local light intensity of each pixel and the average light intensity of the entire frame. If the average deviation of a connected region exceeds the sensitivity threshold corresponding to the current fogging level, the region is marked as a candidate enhancement region. Spatial continuity requires that the geometric center distance between the candidate enhancement region and the previous frame is less than a preset pixel threshold, which is determined to be 15 pixels based on the upper limit of human hand sliding speed and frame rate. Temporal continuity requires that the candidate region be detected in at least 3 consecutive frames, thereby eliminating false enhancements caused by instantaneous noise flashes or random water droplets falling.

[0033] The proportion of transmitted light intensity transitions in the local region across various spectral bands falls within a preset threshold range for biological tissue scattering characteristics. This threshold range is calibrated experimentally based on the difference in transmitted light intensity between skin and water droplets / fabric in the two spectral bands. The specific calibration method is as follows: Calibration environmental conditions: Ambient temperature controlled at 25±2℃, relative humidity controlled at 85±5%, stable fogging state of the mirror (the overall transmitted light intensity of the mirror is reduced to 20%~30% of the fog-free state), constant light source drive current at 80% of the rated value, and fixed photoelectric sensor gain at the calibration range.

[0034] Calibrated samples: Collect four types of samples respectively: (1) the fingertip of an adult index finger (dry state, wiped with anhydrous ethanol and air-dried before contact), (2) a clean water droplet with a diameter of 2mm to 5mm (dropped onto the fogged area of ​​the mirror using a micropipette), (3) dry cotton fabric (pure cotton towel, cut into small pieces of 20mm×20mm, dried in a drying oven at 60℃ for 2 hours before use), and (4) moist cotton fabric (the above dry fabric is soaked in deionized water at 25℃ and squeezed until it does not drip). Collect no less than 100 sets of valid data for each type of sample.

[0035] Calibration steps: Press the sample to be tested vertically onto the fogged mirror surface with natural force (approximately 1N to 2N, monitored and controlled in real time by the calibration personnel through a fingertip pressure sensor) and wipe away the fog layer to form a light-transmitting area with a diameter of not less than 10mm. Record the transition ratio of transmitted light intensity in the 940nm near-infrared band and the 550nm visible light band of the contact area (the transition ratio is defined as the stable light intensity value after contact divided by the average light intensity of the fogged area before contact). Calculate the dual-band ratio (near-infrared transition ratio divided by visible light transition ratio) and take the statistical average.

[0036] Data processing: For each class of samples, outliers were removed from the 100 valid data sets, and the mean of the remaining data ± 2 times the standard deviation was used as the range of the dual-band ratio for that class of samples.

[0037] Calibration results: Based on the above calibration experiments, the dual-band transition ratio for human skin is 1.4 to 1.6, for water droplets it is 0.9 to 1.1, for dry fabric it is 0.7 to 0.9, and for wet fabric it is 0.8 to 1.0. Therefore, the threshold range for biological tissue scattering characteristics is set as follows: the ratio of dual-band transitions is between 1.3 and 1.7, and the near-infrared band transition ratio is greater than 2.2.

[0038] According to Mie scattering theory, the scattering cross section is related to the particle's complex refractive index, particle size distribution, and light wavelength. The complex refractive index of human skin tissue is approximately 1.4 minus 0.05 imaginary part in the near-infrared band (940 nm) and approximately 1.38 minus 0.03 imaginary part in the visible light band (550 nm). Due to the presence of hemoglobin and collagen fibers in the skin, its equivalent particle size distribution is similar to the near-infrared wavelength, resulting in a larger Mie scattering cross section. After removing the fog layer, the near-infrared transmitted light intensity is significantly enhanced. The complex refractive index of water droplets is approximately 1.33 minus 0.0001 imaginary part in the 940 nm band and approximately 1.33 minus 0 imaginary part in the 550 nm band. Its scattering cross section is much smaller than that of skin tissue, and after removal, the enhancement amplitudes in both bands are similar. The complex refractive index of fabric fibers is approximately 1.55 minus 0.01 imaginary part in both bands, but its fiber diameter is much larger than the light wavelength. Scattering is mainly geometrical optics, and the enhancement amplitudes in both bands are low, with the ratio deviating from the characteristics of skin.

[0039] If the dual-band characteristics of the candidate enhancement region fall within this range, the enhancement region is identified as a transmission window formed by the removal of the fog layer by external force. The system assigns a unique identifier to the transmission window and initializes the tracking data structure. If it does not fall within this range, it is determined to be due to environmental interference or water droplet falling, and the candidate region is discarded.

[0040] During the transmission window recognition process, the method obtains the transmitted light intensity distribution of the mirror surface in a fog-free state as a first reference distribution. This first reference distribution is a reference light intensity matrix acquired during system initialization under fog-free conditions and stored in non-volatile memory. The method also obtains the transmitted light intensity distribution of the unaffected area of ​​the mirror surface in a fogged state as a second reference distribution. This second reference distribution is obtained by spatially averaging the light intensity values ​​of the remaining fogged areas after removing all candidate enhancement areas in the current frame. Furthermore, the method obtains the transmitted light intensity distribution of the mirror surface in two different spectral bands, generating a first reference distribution and a second reference distribution for each band. Specifically, it stores the first and second reference distributions for the near-infrared band and the visible light band, respectively, with each distribution being a two-dimensional floating-point matrix with the same resolution as the mirror surface.

[0041] When the transmitted light intensity of the local region transitions from close to the second reference distribution to close to the first reference distribution within a dynamic time window, and the intensity value after the transition remains stable within consecutive acquisition frames, the local region is identified as the transmission window. The criterion for proximity is that the absolute difference between the average light intensity of the local region and the light intensity at the corresponding position of the reference distribution is less than 0.5 times the standard deviation of the light intensity of the reference distribution. The identification also needs to satisfy that the proportion of transmitted light intensity transitions in each spectral band of the local region conforms to the preset biological tissue scattering characteristic threshold range, thereby ensuring that the transmission window is caused by biological tissue contact rather than other physical disturbances. The duration of the dynamic time window is determined by the difference between the second reference distribution and the first reference distribution. The larger the difference, the longer the duration of the dynamic time window. The difference is defined as the average of the absolute values ​​of the differences between the mean of the first reference distribution and the mean of the second reference distribution for each of the two bands, that is, the difference is equal to the absolute value of the difference between the first reference mean minus the second reference mean in the near-infrared band plus the absolute value of the difference between the first reference mean minus the second reference mean in the visible light band, divided by 2. When the difference is greater than the preset high threshold of 200 gray levels, the dynamic time window is set to 5 frames because light intensity recovery takes longer to stabilize when the fog layer is thick. When the difference is between the preset low threshold of 80 gray levels and the preset high threshold of 200 gray levels, the dynamic time window is determined by linear interpolation to be an integer between 3 and 5 frames. Specifically, the duration of the dynamic time window is equal to 3 frames plus the difference minus 80, divided by 120, multiplied by 2, and rounded to the nearest integer to ensure recognition stability when the fog layer thickness is moderate. When the difference is less than the preset low threshold of 80 gray levels, the dynamic time window is set to 3 frames because light intensity recovery is faster after the thin fog layer is removed.

[0042] The method tracks the motion trajectory of the transmission window on the mirror surface. It acquires the geometric center position of the transmission window in continuously acquired multi-frame transmission light intensity distributions. The geometric center position is calculated by weighting the coordinates of each pixel within the masked area of ​​the transmission window according to its light intensity value. The calculation formula is: the abscissa of the geometric center equals the sum of the products of the abscissa of each pixel within the masked area and its light intensity value, divided by the total light intensity value of the area; the ordinate is calculated similarly. The geometric center positions of adjacent frames are correlated and matched to generate a sequence of the transmission window's position changing over time. The correlation matching uses a nearest neighbor algorithm, using the geometric center position of the transmission window in the previous frame as a reference, searching for the window with the smallest Euclidean distance among all transmission windows in the current frame as the matching result for the same target. If the minimum distance exceeds a preset correlation threshold of 30 pixels, the target is determined to be lost, and the current trajectory tracking is terminated.

[0043] When the displacement between adjacent positions in the sequence falls within a preset speed range, the sequence is retained as the motion trajectory. The preset speed range corresponds to the natural sliding speed range of a human hand on a mirror surface, with a lower limit of 5 millimeters per second to avoid false displacements caused by breathing or micro-vibrations of the mirror surface, and an upper limit of 500 millimeters per second to eliminate interference from high-speed objects such as splashing water droplets. If the displacement exceeds this range, the current sequence is interrupted and the subsequent position is used as the starting point of a new trajectory.

[0044] When the displacement of the geometric center of the transmission window within consecutive acquisition frames is less than a preset displacement threshold, the area change rate of the transmission window is obtained. When the area change rate exhibits a pulse pattern of first increasing and then decreasing within a preset time window, it is identified as a tap operation, and a short tap or long tap is distinguished based on the pulse width of the pulse pattern. The preset displacement threshold is set to 3 pixels. When the geometric center displacement is less than this threshold for 3 consecutive frames, the tap detection logic is triggered. The area change rate is calculated by subtracting the area of ​​the previous frame from the pixel area of ​​the transmission window in the current frame and then dividing by the area of ​​the previous frame. The pulse pattern requires that the area change rate exceeds a positive threshold of 15% on the rising edge and is lower than a negative threshold of -10% on the falling edge, and the pulse width is between 2 and 15 frames: a pulse width less than 5 frames is determined as a short tap, and a pulse width greater than or equal to 5 frames is determined as a long tap.

[0045] When a tap operation is detected, a tap control command containing pressure information is generated based on the area of ​​the transmission window at the peak of the pulse pattern, the aspect ratio of the equivalent ellipse, and the intensity of the edge light gradient. The larger the area of ​​the transmission window, the closer the aspect ratio of the equivalent ellipse is to 1, and the greater the intensity of the edge light gradient, the higher the corresponding touch pressure level.

[0046] When a finger presses on the fog layer, an elastic contact is formed between the fingertip and the mirror surface. As the pressure increases, the soft tissue of the fingertip deforms more severely, and the contact area expands non-linearly. At the same time, the fog layer is squeezed out at the contact edge. If the pressure is determined solely by the area size, the difference in fingertip size among different users will cause the pressure level to drift; if it is determined solely by the edge gradient, uneven fog layer thickness will cause the gradient reference to be inaccurate; if it is determined solely by the aspect ratio, the trailing shape during the sliding process will introduce shape noise.

[0047] Meanwhile, according to Hertzian contact theory, when an elastic sphere contacts a rigid plane, the contact area is proportional to the square of the third of the normal pressure. Therefore, the variance of the area change rate directly reflects pressure fluctuations. The aspect ratio change rate reflects the degree to which the contact surface shape approaches a circle; as pressure increases, soft tissue deformation makes the contact surface more circular. The edge light intensity gradient change rate reflects the rate at which the fog layer is squeezed out at the contact edge. According to the Hele-Shaw flow model in fluid mechanics, the expulsion rate is positively correlated with the contact pressure. These three factors are coupled with Hertzian contact mechanics and fluid mechanics. Although they cannot directly output Newtonian force values, they can serve as a soft mapping of pressure gradation, helping the system establish and operate accordingly. Figure 1 The resulting sense of pressure.

[0048] Based on the aforementioned physical laws, this application employs the following control logic: a weighted fusion of the area change rate, aspect ratio change rate, and edge gradient change rate is used. During the sliding process, the temporal evolution of these three indicators is integrated, and the area-pressure nonlinear relationship of Hertzian contact theory is utilized to compensate for individual differences in a single indicator. Through multi-parameter coupling design, the physical laws of contact mechanics can be used to offset pressure drift caused by user physiological differences, avoiding misjudgments of light touches as hard presses or missed hard presses, while simultaneously achieving normalization and stabilization of pressure perception.

[0049] The aspect ratio of the equivalent ellipse is determined by fitting the second-order moments of the region: First, the zero-order moment and the second-order central moment of the transmission window mask region are calculated. The zero-order moment is the sum of the light intensity values ​​of each pixel in the mask region. The second-order central moments include the second-order central moments in the horizontal direction, the second-order central moments in the vertical direction, and the mixed second-order central moments. Second, an equivalent ellipse with the same second-order moment distribution characteristics as the mask region is constructed based on the second-order central moments. The lengths of the major and minor axes are obtained by solving for the principal inertial axis length of the equivalent ellipse. Finally, the ratio of the major axis length to the minor axis length is taken as the aspect ratio of the equivalent ellipse. The edge light intensity gradient intensity is determined by the following steps: First, extract the set of pixels of the outer contour of the transmission window mask region; second, extend the outer contour outward by 2 pixels along the normal direction to form a ring-shaped boundary sampling band with a width of 2 pixels; third, for each pixel in the ring-shaped boundary sampling band, use the Sobel gradient operator to calculate the horizontal and vertical gradient components respectively, and obtain the magnitude of the gradient vector of each pixel; finally, take the arithmetic mean of the gradient magnitudes of all pixels in the ring-shaped boundary sampling band as the edge light intensity gradient intensity.

[0050] The area change rate, equivalent ellipse aspect ratio change rate, and edge light intensity gradient intensity change rate of the transmission window in consecutive acquisition frames are obtained. Touch pressure levels are generated based on these rates. The area change rate is defined as the current frame area minus the previous frame area, divided by the previous frame area. The equivalent ellipse aspect ratio change rate is defined as the current frame aspect ratio minus the previous frame aspect ratio, divided by the previous frame aspect ratio. The edge light intensity gradient intensity change rate is defined as the current frame boundary average gradient minus the previous frame boundary average gradient, divided by the previous frame boundary average gradient.

[0051] For swiping operations, the method calculates the sliding window variance of the area change rate between consecutive frames, the absolute value of the first difference of the aspect ratio change rate of the equivalent ellipse, and the moving average of the edge light intensity gradient intensity change rate. These three indicators are weighted and fused into a touch pressure level, with weighting coefficients determined to be 0.4, 0.3, and 0.3 based on calibration experiments. The pressure levels are divided into four levels: Level 1 corresponds to a light touch, with an area change rate variance less than 0.1 and a gradient change rate less than 20% of the baseline value; Level 2 corresponds to a medium press, with an area change rate variance between 0.1 and 0.3 and a gradient change rate between 20% and 50% of the baseline value; Level 3 corresponds to a heavy press, with an area change rate variance between 0.3 and 0.6 and a gradient change rate between 50% and 80% of the baseline value; and Level 4 corresponds to a strong swipe, with an area change rate variance greater than 0.6 and a gradient change rate greater than 80% of the baseline value. Control commands are generated based on the morphological characteristics of the motion trajectory and the touch pressure level: for sliding trajectories, scrolling or page turning commands are mapped according to the trajectory direction and length; for tap operations, single tap, double tap, or long press menu commands are mapped according to the pressure level. For complex trajectories, template matching is performed based on a preset gesture library to generate corresponding function control commands. The preset gesture library stores standard gesture trajectory coordinate sequence templates, including at least straight lines, polylines, arcs, and closed shapes. Template matching uses a dynamic time warping algorithm to calculate the minimum cumulative distance between the current motion trajectory and the standard gesture trajectory template. When the minimum cumulative distance is less than a preset matching threshold, the gesture matching is considered successful and the corresponding function control command is generated.

[0052] When at least two of the aforementioned transmissive windows exist simultaneously, the method performs multi-source touch separation. The method acquires the erase width, erase area change rate, trajectory curvature, erase speed, and transmissive light intensity recovery degree for each transmissive window. The erase width is defined as the maximum span of the transmissive window mask area perpendicular to the direction of motion. The trajectory curvature is calculated using the reciprocal radius of the circle fitted to the geometric center positions of three consecutive frames. The transmissive light intensity recovery degree is defined as the ratio of the average light intensity of the transmissive window in the current frame to the light intensity of the corresponding region of the first reference distribution. The method acquires the operational dynamics characteristics of the motion trajectory corresponding to each transmissive window. These operational dynamics characteristics include the mean trajectory velocity, velocity variance, acceleration distribution, tap interval duration, and statistical moments of the trajectory curvature distribution. The statistical moments include the first-order origin moment, the second-order central moment, and the third-order skewness. Based on the erase width, erase area change rate, trajectory curvature, erase speed, and degree of transmitted light intensity recovery, as well as the operation dynamics characteristics, each transmission window is clustered to generate at least two operation source groups. The clustering employs a density-based spatial clustering algorithm, using the erase width and mean trajectory speed as core feature dimensions. The density threshold is set based on typical human hand size and sliding speed, with an erase width between 10 mm and 30 mm and a mean speed between 20 mm / s and 300 mm / s. The neighborhood radius is set to 5 mm, and the minimum number of samples is set to 2. If any transmission window remains unassigned after clustering, it is marked as isolated noise and discarded.

[0053] The operational dynamics characteristics of each operation source group are compared with a preset user touch behavior feature library. This feature library includes statistical features such as historical trajectory speed average, tap duration distribution, and trajectory curvature distribution. The user identity confidence level corresponding to each operation source group is determined by calculating the Euclidean distance between the current group's feature vector and each user template in the feature library. The confidence level is equal to 1 divided by 1 plus the natural exponent of the Euclidean distance. The motion trajectory of the transmission window corresponding to each operation source group is tracked, and corresponding control commands are generated. The control commands for each group are sent to the application layer through an independent message queue, thereby supporting multi-user parallel operation.

[0054] The light intensity attenuation rate of the transmission window in consecutive acquisition frames is obtained. The light intensity attenuation rate characterizes the rate at which the transmitted light intensity of the transmission window decreases over time. The light intensity attenuation rate is calculated by subtracting the average light intensity of the previous frame from the average light intensity of the transmission window in the current frame, and then dividing by the average light intensity of the previous frame. A negative value indicates light intensity attenuation. The temperature difference between the local mirror temperature of the region where the transmission window is located and the temperature difference between the surrounding foggy region and the local mirror temperature is obtained. The local mirror temperature is obtained by querying the interpolation result of the aforementioned mirror temperature field distribution at the geometric center of the transmission window. The temperature of the surrounding foggy region is taken as the average temperature of the annular region after extending the outer rectangle of the transmission window by 20 pixels. When the temperature difference gradually converges from above a preset temperature difference threshold to below the preset temperature difference threshold, and the light intensity attenuation rate is below the dynamic threshold, the transmission window is determined to have entered a fading state. The position of the transmission window in the current frame is marked as the trajectory endpoint, and the tracking of the transmission window is stopped. The preset temperature difference threshold is set to 2 degrees Celsius. When the temperature difference between the local area and the surrounding fogged area continuously decreases from greater than 2 degrees Celsius to less than 2 degrees Celsius, it indicates that the area is being recooled by ambient vapor after the finger is removed. The dynamic threshold is determined by the local mirror temperature at the location of the transmission window. The lower the local temperature, the smaller the dynamic threshold: -5% per frame when the local temperature is above 30 degrees Celsius, -8% per frame when the local temperature is between 20 and 30 degrees Celsius, and -12% per frame when the local temperature is below 20 degrees Celsius. In subsequent capture frames, in response to the reappearance of enhanced transmitted light intensity at the same spatial location, where the enhanced light intensity is higher than the light intensity at the previous trajectory endpoint, the reappearing enhanced area is identified as a new transmission window, and a new motion trajectory is generated, thereby supporting repeated touch operations at the same location.

[0055] The transmitted light intensity distribution of the non-erased area on the mirror surface where the transmission window has not been formed is obtained. The non-erased area is obtained by inverting the transmission window mask of the current frame's transmitted light intensity distribution; that is, the set of pixels in the non-erased area equals the set of pixels on the entire mirror surface minus the set of pixels identified by the transmission window mask. Based on the difference between the transmitted light intensity distribution of the non-erased area and the reference transmitted light intensity distribution of the mirror surface in a fog-free state, a residual fog layer accumulation distribution map of the mirror surface is generated. The difference is calculated using pixel-by-pixel absolute difference; the larger the difference, the thicker the residual fog layer at that location.

[0056] Furthermore, based on the cumulative distribution map of the residual fog layer, the fog layer thickness levels for each region are determined: regions with a difference less than 20% of the baseline difference are considered light fog regions, those with a difference between 20% and 50% are considered medium fog regions, and those with a difference greater than 50% are considered thick fog regions. The baseline difference is defined as the absolute value of the difference between the average light intensity of the entire mirror surface in the fog-free state and the average light intensity of the untouched area after fogging. Calibration experiments are conducted using standard touch points at each fog layer thickness level to obtain the measured coordinate offset between the geometric center position of the transmission window and the standard touch coordinates. A correspondence table between the fog layer thickness level and the measured coordinate offset is established. The calibration experiment uses a robotic arm to press the mirror surface with known coordinates, recording the deviation between the geometric center of the transmission window and the actual pressing point. The offset is approximately 2 pixels in the light fog region, approximately 5 pixels in the medium fog region, and approximately 12 pixels in the thick fog region. The calibration experiment was conducted after the fogging level of the mirror surface stabilized. A three-axis robotic arm was used to press the mirror surface vertically with a constant pressing force. At least 25 calibration points were evenly selected on the mirror surface for each fog layer thickness level. The deviation between the geometric center of the transmission window and the coordinates of the end of the robotic arm was recorded and averaged to establish the corresponding relationship table.

[0057] Furthermore, based on the cumulative distribution map of the residual fog layer, identifiable and non-identifiable areas on the mirror surface are determined. Identifiable areas are those where the light intensity difference in the cumulative distribution map of the residual fog layer is below a preset threshold. The preset threshold is set to 60% of the baseline difference. Areas exceeding this threshold are unable to reliably identify touch because the light intensity signal is submerged in noise due to excessive fog thickness. When generating the motion trajectory, only the transmission windows located within the identifiable areas are tracked. For trajectories crossing the boundaries of the identifiable areas, truncation is performed at the boundaries, retaining the trajectory segments within the identifiable areas and discarding the trajectory segments within the non-identifiable areas. According to the correspondence table, the geometric center positions of the transmission windows located within each fog layer thickness level area in the motion trajectory are compensated by superimposing the corresponding measured coordinate offset, generating a corrected touch coordinate sequence, thereby eliminating positioning drift caused by fog scattering.

[0058] The erase width change sequence and light intensity recovery degree change sequence of each transmission window within each operation source group are obtained during continuous tracking. The erase width change sequence is an array of erase width values ​​arranged over time, and the light intensity recovery degree change sequence is an array of light intensity recovery degree values ​​arranged over time. Based on the erase width change sequence and light intensity recovery degree change sequence, a feature consistency coefficient is calculated within each operation source group. The feature consistency coefficient is equal to the weighted average of the coefficient of variation of the erase width sequence and the coefficient of variation of the light intensity recovery degree sequence. The coefficient of variation is the standard deviation divided by the mean, and each has a weight of 0.5. When the feature consistency coefficient continuously decreases within a preset continuous sampling window and falls below a preset threshold, the operation source group is split into at least two sub-operation source groups. The preset continuous sampling window is set to 30 frames, and the preset threshold is set to 0.3. A continuous decrease in the consistency coefficient indicates that two different operation sources may have been mixed in within the original group, such as the simultaneous operation of an adult's and a child's hands, resulting in systematic differences in their erase width and light intensity recovery dynamics. The operation dynamics characteristics of each sub-operation source group are compared with the user touch behavior feature database. When the similarity exceeds a preset matching threshold, the sub-operation source group is bound to the corresponding user identity. The matching threshold is set to 0.85. After binding, the control commands generated by this group have the full permissions of the user, including accessing personal health data and adjusting personal preference settings. When the similarity is lower than a preset visitor threshold, it is marked as a visitor operation source and permission downgrade control is triggered. The visitor threshold is set to 0.5. Permission downgrade includes prohibiting the use of personalized settings, restricting access to sensitive functions, and shortening the timeout time for a single operation to 10 seconds. The motion trajectory of the transmission window corresponding to each sub-operation source group is tracked, and corresponding control commands are generated respectively.

[0059] On fogged mirrors, the fog layer is an inherent optical medium. When a user touches the mirror, the fog layer is wiped away, forming a transmission window. However, the fog layer at the edge of the window regenerates rapidly under thermodynamic drive. If the entire area is heated for defogging at this time, the continuously rising mirror temperature not only consumes a huge amount of energy but also leads to the accumulation of thermal stress and the risk of cracking. If no heating is applied at all, the transmission window disappears rapidly after the finger is removed, and subsequent touches require re-establishing the window, resulting in a broken interaction. At the same time, the area where the finger touches conducts heat to the mirror during pressing, making the local substrate temperature higher than the surrounding fogged area, and the heat conduction flux from high-pressure, long-term contact is even greater. Although this thermal inertia is insufficient for continuous defogging, it can serve as a soft benchmark for predictive heating, helping the system reduce preheating power requirements.

[0060] Based on the above, this application achieves a complementary timing between trajectory prediction and zoned heating, preheating in advance before the transmission window moves, and reducing preheating power by utilizing the thermal inertia compensation of residual finger heat. Through the complementary timing design, the heating lag caused by prediction error can be offset by the thermal inertia soft reference, avoiding thermal stress and energy consumption caused by continuous full-power heating, while achieving a seamless balance between touch continuity and energy efficiency.

[0061] The current direction and speed of the transmission window are obtained. Based on the current direction and speed, the predicted position of the transmission window after a preset time is predicted. The current direction of movement is determined by the slope of the least-squares fitted line of the geometric center positions of the most recent three frames, and the current speed is calculated by dividing the displacement of the most recent three frames by the frame interval. The preset time is set to 800 milliseconds based on the response delay of the mirror back heating unit. The predicted position is equal to the current position plus the velocity vector multiplied by the preset time. Based on the predicted position, the corresponding prediction area on the mirror surface is determined. The prediction area is a circular area centered on the prediction position with a radius equal to the maximum historical erasure width of the transmission window. A heating command is sent to the mirror back heating unit corresponding to the prediction area. The heating command is used to raise the mirror surface temperature of the prediction area to above the dew point temperature before the transmission window reaches the prediction area, so as to prevent the prediction area from being in a fogged state when the transmission window arrives. The mirror back heating unit consists of a partitioned controllable thin-film heating resistor array. Each partition corresponds to a rectangular physical area on the mirror surface, with a size of 50 mm by 50 mm. The controller adjusts the heating power of each partition through a pulse width modulation signal with a pulse width modulation frequency of 1 kHz to avoid low-frequency flicker.

[0062] The historical heating power sequence of the mirror back heating unit and the fog layer dissipation rate sequence of the corresponding region under each historical heating power are obtained. A mapping relationship table between each heating power and the fog layer dissipation rate is determined. The fog layer dissipation rate is defined as the rate of change of fog layer thickness over time, and is indirectly calculated by monitoring the transmitted light intensity recovery speed through a photoelectric sensor. The mapping relationship table is established through a one-dimensional heat conduction equation and an evaporation rate model. According to Fourier's law of heat conduction, the rate of temperature rise of the mirror is directly proportional to the heating power density, and the proportionality coefficient is determined by the thermal conductivity, thickness, and backplate thermal resistance of the mirror glass. According to the Clausius-Clapeyron equation, the evaporation rate of the fog layer is exponentially related to the degree to which the mirror temperature exceeds the dew point. Combining these two factors, a nonlinear mapping between heating power and fog layer dissipation rate can be obtained. This mapping table is stored in discrete lookup table form, with a power step of 0.5 watts per square centimeter and a dissipation rate step of 0.01 millimeters per second. The mapping table is established before the system leaves the factory through calibration in a standard temperature and humidity chamber: the mirror is placed in a constant dew point temperature environment, and heating power is applied step-by-step according to the power step. The time required for the transmitted light intensity to recover to 90% of the first reference distribution at each power level is monitored by a photoelectric sensor, and the fog layer dissipation rate is calculated by inversion, thereby establishing the nonlinear mapping.

[0063] Based on the mapping table and the current movement speed of the transmission window, the target heating power required for the predicted area is calculated so that the predicted area is in a fog-free or foggy state when the transmission window arrives, and naturally recovers to a foggy state after the transmission window leaves using thermal inertia.

[0064] Specifically, the calculation of the target heating power follows these steps: First, the reference heating power required to reduce the current fog layer thickness at a reference reduction rate is retrieved according to the mapping table, where the reference reduction rate is set to 0.05 mm per second. Second, a speed compensation coefficient is calculated based on the current movement speed. This coefficient is equal to the quotient obtained by multiplying the reference reduction rate by the predicted distance and then dividing by the current movement speed. The predicted distance is equal to the current movement speed multiplied by a preset time of 800 milliseconds. When the quotient is greater than 1, the quotient itself is used; when the quotient is less than or equal to 1, 1 is used, thus ensuring that the faster the movement speed, the higher the preheating power. Finally, the reference heating power is multiplied by the speed compensation coefficient to obtain the target heating power. Correspondingly, if the back-of-mirror heating unit feedback resistance is abnormal or the temperature exceeds the limit, the heating of that zone is skipped and a fault code is reported to the system to avoid local overheating.

[0065] The actual fog layer thickness value of the predicted area is obtained when the transmission window reaches the predicted area. The actual fog layer thickness value is calculated by inversion of the difference between the transmitted light intensity in the area and the first reference distribution. The inversion formula is that the fog layer thickness is equal to the inverse of the extinction coefficient of light in the fog layer multiplied by the negative natural logarithm of the quotient of the difference between the reference transmitted light intensity and the current transmitted light intensity divided by the reference transmitted light intensity. The extinction coefficient is determined by the Mie scattering theory of fog droplet size distribution and light wavelength. For the 940 nm near-infrared band, the value is about 2.3 per millimeter in a typical bathroom steam environment.

[0066] Based on the actual fog thickness value, the heating power of the mirror back heating unit is adjusted so that the predicted area recovers to the fogged state at a preset cooling rate after the transmission window leaves. Specifically, the method compares the actual fog thickness value with the predicted fog thickness value: if the actual value is greater than the predicted value, it indicates insufficient preheating and a thick residual fog layer, so the heating power is increased to 1.2 times the original target power to accelerate evaporation; if the actual value is less than the predicted value and the transmitted light intensity is close to the first reference distribution, it indicates overheating, so the heating power is reduced to 0.6 times the original target power. Furthermore, to achieve the preset cooling rate of 0.5 degrees Celsius per second, the method establishes a first-order thermal resistance network model based on the heat capacity of the mirror glass substrate, the thermal resistance of the backplate, and the environmental convective heat transfer coefficient. It calculates the target heat flux density required to reduce the local temperature of the current mirror surface to below the ambient dew point temperature, and adjusts the output power of the mirror back heating unit to 0.6 times the power value corresponding to the target heat flux density. This ensures that the mirror surface cools uniformly at the preset cooling rate after the transmission window leaves, thereby promptly restoring the fogged state during the touch gap and avoiding the accumulation of thermal stress and energy waste caused by continuous high temperatures on the mirror surface. Correspondingly, if the adjusted heating power exceeds the unit's rated maximum power, it is limited to the rated power, and an anomaly log is recorded.

[0067] The touch pressure level and dwell time of the transmission window within the prediction area are obtained. The dwell time is calculated by multiplying the frame interval by the number of frames from the first entry of the geometric center of the transmission window into the boundary of the prediction area to the last exit from the boundary. Based on the touch pressure level and dwell time, the thermal inertia compensation coefficient in the mapping table is adjusted. This thermal inertia compensation coefficient acts as a correction factor on the base target heating power output by the mapping table. Specifically, the higher the touch pressure level or the longer the dwell time, the larger the thermal inertia compensation coefficient: Level 1 pressure corresponds to a compensation coefficient of 1.0, Level 2 pressure to 1.2, Level 3 pressure to 1.4, and Level 4 pressure to 1.6; an additional 0.1 is added for every full second of dwell time, with a maximum of 2.0. The physical basis of the thermal inertia compensation coefficient is that high pressure or long-term contact increases the heat conduction flux between the finger and the mirror surface. The finger, as a isothermal body, continuously absorbs heat from the mirror surface, while the forceful wiping action more thoroughly dissipates the fog layer, resulting in a higher heat loss rate in this area compared to low-pressure, short-term contact. Therefore, a larger compensation coefficient is needed to characterize the thermal inertia deficit. The reduction of the target heating power in subsequent predictions, based on the lookup result in the mapping table, means dividing the base target heating power obtained from the lookup by the thermal inertia compensation coefficient, thereby reducing the preheating power requirement in the next trajectory prediction. Because the previous high-pressure, long-duration contact has already caused additional frictional and conductive heat to accumulate in this area, the thermal inertia of the mirror substrate is sufficient to maintain a temperature higher than the surrounding fogged area for several seconds, thus utilizing residual heat to maintain a brief fog-free state. Therefore, reducing the heating power avoids overheating, thereby reducing system energy consumption and mirror thermal stress while ensuring touch continuity.

[0068] As a specific implementation method, assuming the bathroom ambient temperature is 25 degrees Celsius and the relative humidity is 85%, the dew point temperature calculated based on the Magnus formula is 22.3 degrees Celsius. After showering begins, the mirror surface temperature drops to 20 degrees Celsius due to contact with cold air. The temperature reading at the center sampling point of the thermocouple array is 19.8 degrees Celsius, lower than the dew point temperature, and the system determines a fogging state. The global average transmitted light intensity decreases from 4095 in the fog-free state to 820, with a normalized attenuation ratio of 0.8, corresponding to four levels of fogging. The light intensity shift detection sensitivity coefficient is set to 0.2. The system switches to optical touch mode and activates the 940nm and 550nm dual-band light source.

[0069] The user touches the center area of ​​the mirror with their index finger, wiping away a circular fog layer approximately 15 mm in diameter. In this area, the transmitted light intensity at the 940 nm band jumps from an average of 120 to 380, and at the 550 nm band, it jumps from 80 to 180. The ratio of the two band transitions is 3.17 to 2.25, or 1.41, falling within the biological tissue scattering characteristic threshold range of 1.3 to 1.7. The system identifies this area as a transmission window and assigns it the identifier TW001. In subsequent frames, the geometric center of TW001 moves horizontally at a speed of 80 mm per second. The displacement falls within the natural sliding speed range of 5 to 500 mm per second, and the system retains this sequence as a valid motion trajectory. The trajectory length is 80 mm, the direction is to the right, and it is mapped to a page-turning command.

[0070] When a user taps the lower right corner of the mirror with their thumb, the geometric center of the transmissive window shifts by less than 3 pixels within three consecutive frames. The area change rate exhibits a pulse pattern: 45% at the rising edge of the first frame, peak at the second frame, and -30% at the falling edge of the third frame, with a pulse width of three frames, indicating a short press. At the peak, the transmissive window area is 180 square millimeters, the equivalent ellipse aspect ratio is 1.1, and the edge light intensity gradient is 85 gray levels per millimeter, corresponding to a level 2 touch pressure, generating a short press confirmation command.

[0071] Assuming two users simultaneously use the mirror, the left transmission window has an erasing width of 22 mm and an average speed of 60 mm / s, while the right window has an erasing width of 18 mm and an average speed of 150 mm / s. The system uses DBSCAN clustering to divide the two transmission windows into operation source groups A and B. Group A's operation dynamics feature has a Euclidean distance of 0.3 with the template of user A in the feature library, with a confidence level of 0.74, which is below the matching threshold of 0.85 but above the visitor threshold of 0.5, and is therefore marked as a low-confidence user. Group B has an Euclidean distance of 0.1 and a confidence level of 0.90, exceeding the matching threshold, and is therefore bound to user B. The two groups generate independent control commands respectively; user B enjoys full permissions, while user A is limited to basic functions.

[0072] When the transmission window TW001 moves towards the upper right corner of the mirror, the system predicts its position 800 milliseconds later based on its current direction and speed. It then sends a heating command to the corresponding back heating zone of the mirror, with a target heating power of 8 watts per square centimeter (based on a lookup table). When the transmission window reaches the prediction area, the actual fog layer thickness is calculated to be 0.05 mm, lower than the predicted 0.08 mm, indicating overheating. The system then reduces the heating power to 4.8 watts per square centimeter. The transmission window remains in this area for 0.5 seconds, with the touch pressure level set to level three. The thermal inertia compensation coefficient is adjusted to 1.4 plus 0, i.e., 1.4. In the next prediction, the target heating power is divided by 1.4 based on the lookup table, thus utilizing thermal inertia to reduce energy consumption.

[0073] The core of this method lies in the fact that the condensation and removal of the fog layer on the mirror surface is essentially a dynamic reconstruction process of the boundary conditions of the optical medium. Traditional touch technology treats the fog layer as an interference that needs to be completely eliminated, while this application utilizes the fog layer as an optical modulation layer: when the fog layer is present, the transmitted light intensity is uniformly attenuated, forming a stable dark field background; when biological tissue contacts and removes the fog layer, a sudden change in local optical transmittance forms a bright spot, which is the transmission window. Since the Mie scattering characteristics of biological tissue in a specific spectral band are unique, the actual touch and environmental interference can be physically separated through dual-band transition ratio analysis. Furthermore, there is a mapping relationship between the geometric changes of the transmission window and contact mechanics: according to Hertzian contact theory, increased pressure leads to an increase in the deformation area of ​​the fingertip and the contact surface becomes closer to a circle; according to the Hele-Shaw flow model, increased pressure causes the edge fog layer to be squeezed out, forming a steep gradient. Therefore, the three parameters of area, aspect ratio, and edge gradient are sufficient to inversely determine the touch pressure. At the trajectory prediction and heating control level, the system uses the thermal inertia of the mirror substrate to establish an energy balance model, transforming active defogging from continuous heating of the entire mirror surface to precise preheating in front of the trajectory, thereby reducing energy consumption from full-power continuous mode to pulse heating mode that follows the touch path, while avoiding the thermal stress problem caused by uneven temperature distribution of the mirror surface in traditional solutions.

[0074] In summary, this application achieves reliable optical touch interaction for smart bathroom mirrors in foggy environments through a progressive data processing flow, including optical detection of fogging conditions, multispectral transmission window recognition, motion trajectory tracking, pressure level generation, multi-user separation, fog residue compensation, and predictive zone heating. The method utilizes the fog layer itself as an optical modulation medium, transforming traditional touch failure scenarios into usable optical signal sources. It distinguishes between real touch and environmental interference through biological tissue scattering characteristics, clears obstacles before touch is detected through predictive heating, and dynamically adapts to user operating habits through thermal inertia compensation. Thus, it balances touch accuracy, response speed, and system energy efficiency in the high-temperature and high-humidity environment of a bathroom.

[0075] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data processing method for smart bathroom mirrors, characterized in that, The method includes: In response to the detection that the smart bathroom mirror is in a fogged state, the current fogging level is determined according to the overall attenuation of the current transmitted light intensity of the mirror surface. Different fogging levels correspond to different light intensity shift detection sensitivities, and the smart bathroom mirror is controlled to switch from the conventional interaction mode to the optical touch mode. In the optical touch mode, the transmitted light intensity distribution of the mirror surface is continuously acquired; wherein, the transmitted light intensity distribution includes the transmitted light intensity sub-distributions acquired in two different spectral bands respectively; In response to the transmitted light intensity distribution, the transmitted light intensity in a local area is enhanced relative to other areas of the same mirror surface, and the enhanced area has spatial continuity and temporal persistence between adjacent acquisition frames. Furthermore, the transmission light intensity transition ratio of the local area in each spectral band falls within a preset biological tissue scattering characteristic threshold range. The threshold range is experimentally calibrated based on the difference in transmitted light intensity between skin and water droplets / fabric in the two spectral bands. The enhanced area is identified as a transmission window formed by the removal of the fog layer by external force, and the movement trajectory of the transmission window on the mirror surface is tracked. The area change rate, the equivalent ellipse aspect ratio change rate, and the edge light intensity gradient intensity change rate of the transmission window in continuous acquisition frames are obtained. Based on the area change rate, the equivalent ellipse aspect ratio change rate, and the edge light intensity gradient intensity change rate, the touch pressure level is generated. Control commands are generated based on the morphological characteristics of the motion trajectory and the touch pressure level.

2. The method according to claim 1, characterized in that, The detection that the smart bathroom mirror is in a fogged state includes: The temperature values ​​of multiple sampling points on the mirror surface and the dew point temperature value of the environment surrounding the mirror surface are obtained; Spatial interpolation is performed based on the temperature values ​​of the multiple sampling points to generate a mirror temperature field distribution; based on the comparison result between the mirror temperature field distribution and the dew point temperature value, a fogging risk gradient map is generated, which is used to guide the spatial differentiation configuration of the light intensity shift detection sensitivity. When the temperature value of any of the sampling points is lower than the dew point temperature value, the mirror surface is determined to be in a fogged state.

3. The method according to claim 1, characterized in that, The step of identifying the enhanced region as a transmission window formed due to the removal of the fog layer by external force includes: The transmitted light intensity distribution of the mirror surface in a fog-free state is obtained as the first reference distribution; The transmitted light intensity distribution of the area of ​​the mirror surface that is not in contact with external force in a fogged state is obtained as a second reference distribution; The transmitted light intensity distribution of the mirror surface under two different spectral bands is obtained, and a first reference distribution and a second reference distribution corresponding to each band are generated respectively. When the transmitted light intensity of the local area transitions from close to the second reference distribution to close to the first reference distribution within a dynamic time window, and the light intensity value after the transition remains stable within consecutive acquisition frames, the local area is identified as the transmission window. The identification also needs to meet the following requirements: the ratio of transmitted light intensity transition in the local area under each spectral band conforms to the preset threshold range of biological tissue scattering characteristics, and the threshold range is experimentally calibrated based on the difference in transmitted light intensity between skin and water droplets / fabric under the two spectral bands. The duration of the dynamic time window is determined by the difference between the second reference distribution and the first reference distribution; the larger the difference, the longer the duration of the dynamic time window.

4. The method according to claim 1, characterized in that, The tracking of the movement trajectory of the transmission window on the mirror surface includes: Obtain the geometric center position of the transmission window in the transmitted light intensity distribution of multiple continuously acquired frames; The geometric center positions between adjacent frames are correlated and matched to generate a sequence of the position of the transmission window changing over time; When the displacement between adjacent positions in the sequence falls within a preset speed range, the sequence is retained as the motion trajectory; the preset speed range corresponds to the natural sliding speed range of a human hand on a mirror surface. When the displacement of the geometric center position of the transmission window within a continuous acquisition frame is less than a preset displacement threshold, the area change rate of the transmission window is obtained. When the area change rate presents a pulse pattern of first increasing and then decreasing within a preset time window, it is identified as a tap operation, and short taps and long taps are distinguished according to the pulse width of the pulse pattern. When a tap operation is detected, a tap control command containing pressure information is generated based on the area of ​​the transmission window at the peak of the pulse pattern, the aspect ratio of the equivalent ellipse, and the intensity of the edge light gradient. The larger the area of ​​the transmission window, the closer the aspect ratio of the equivalent ellipse is to 1, and the greater the intensity of the edge light gradient, the higher the corresponding touch pressure level.

5. The method according to claim 1, characterized in that, When at least two of the transmission windows exist simultaneously, the method further includes: Obtain the erasure width, erasure area change rate, trajectory curvature, erasure speed, and transmitted light intensity recovery degree of each transmission window; The operation dynamics characteristics of the motion trajectory corresponding to each transmission window are obtained. The operation dynamics characteristics include the mean trajectory velocity, velocity variance, acceleration distribution, click interval duration, and statistical moments of trajectory curvature distribution. Based on the erase width, erase area change rate, trajectory curvature, erase speed, and degree of transmission light intensity recovery, as well as the operation dynamics characteristics, each transmission window is clustered to generate at least two operation source groups. The operation dynamics characteristics of each operation source group are compared with a preset user touch behavior feature library, which includes statistical features such as the historical trajectory speed average, tap duration distribution, and trajectory curvature distribution, in order to determine the user identity confidence level corresponding to each operation source group. The motion trajectory of the transmission window corresponding to each operation source group is tracked, and corresponding control commands are generated respectively.

6. The method according to claim 1, characterized in that, The method of tracking the movement trajectory of the transmission window on the mirror surface further includes: The light intensity attenuation rate of the transmission window in consecutive acquisition frames is obtained, and the light intensity attenuation rate characterizes the rate at which the transmitted light intensity of the transmission window decreases with time. The temperature difference between the local mirror temperature of the area where the transmission window is located and the temperature difference between the surrounding fogged area is obtained. When the temperature difference gradually converges from above the preset temperature difference threshold to below the preset temperature difference threshold, and the light intensity attenuation rate is below the dynamic threshold, the transmission window is determined to enter the fading state, the position of the transmission window in the current frame is marked as the trajectory endpoint, and the tracking of the transmission window is stopped. The dynamic threshold is determined by the local temperature of the mirror surface at the location of the transmission window; the lower the local temperature, the smaller the dynamic threshold. In subsequent acquisition frames, in response to the reappearance of enhanced transmitted light intensity at the same spatial location, and the enhanced light intensity value being higher than the light intensity value at the end point of the previous trajectory, the reappearing enhanced area is identified as a new transmission window, and a new motion trajectory is generated.

7. The method according to claim 1, characterized in that, Also includes: Obtain the transmitted light intensity distribution in the non-erased area on the mirror surface where the transmission window is not formed; Based on the difference between the transmitted light intensity distribution of the non-erased area and the reference transmitted light intensity distribution of the mirror in a fog-free state, a residual fog layer accumulation distribution map of the mirror is generated. Based on the cumulative distribution map of the residual fog layer, the fog layer thickness level of each region is determined; by conducting calibration experiments with standard touch points at each fog layer thickness level, the measured coordinate offset between the geometric center position of the transmission window and the standard touch coordinates is obtained, and a correspondence table between the fog layer thickness level and the measured coordinate offset is established. Based on the cumulative distribution map of the residual fog layer, the identifiable and non-identifiable areas on the mirror surface are determined. The identifiable areas are those where the light intensity difference in the cumulative distribution map of the residual fog layer is lower than a preset threshold. When generating the motion trajectory, only the transmission window located within the identifiable area is tracked; Based on the corresponding relationship table, the geometric center position of the transmission window located in each fog layer thickness level region in the motion trajectory is compensated by superimposing the corresponding measured coordinate offset to generate a corrected touch coordinate sequence.

8. The method according to claim 5, characterized in that, After clustering each transmission window to generate at least two operational source groups, the process further includes: Obtain the erase width change sequence and light intensity recovery degree change sequence of each transmission window within each operation source group during continuous tracking; Based on the erase width change sequence and the light intensity recovery degree change sequence, calculate the feature consistency coefficient within each operation source group; When the feature consistency coefficient continues to decrease within a preset continuous sampling window and falls below a preset threshold, the operation source group is split into at least two sub-operation source groups; The operation dynamics features of each sub-operation source group are compared with the user touch behavior feature library; when the similarity exceeds a preset matching threshold, the sub-operation source group is bound to the corresponding user identity; when the similarity is lower than a preset visitor threshold, it is marked as a visitor operation source and permission downgrade control is triggered. The motion trajectory of the transmission window corresponding to each sub-operation source group is tracked, and corresponding control commands are generated respectively.

9. The method according to claim 6, characterized in that, Before determining that the transmission window has entered the fading state, the method further includes: The current direction and speed of the transmission window are obtained, and the predicted position of the transmission window after a preset time is predicted based on the current direction and speed of the transmission window. Based on the predicted position, determine the corresponding predicted area on the mirror surface; A heating command is sent to the mirror back heating unit corresponding to the prediction area. The heating command is used to raise the mirror surface temperature of the prediction area to above the dew point temperature before the transmission window reaches the prediction area, so as to prevent the prediction area from being in a fogging state when the transmission window arrives. Obtain the historical heating power sequence of the mirror back heating unit and the fog layer dissipation rate sequence of the corresponding region under each historical heating power, and determine the mapping relationship table between each heating power and the fog layer dissipation rate; according to the mapping relationship table and the current movement speed of the transmission window, calculate the target heating power required for the predicted region, so that the predicted region is in a fog-free or light fog state when the transmission window arrives, and naturally recovers to a fogged state by thermal inertia after the transmission window leaves.

10. The method according to claim 9, characterized in that, After sending a heating command to the mirror back heating unit corresponding to the predicted region, the method further includes: Obtain the actual fog layer thickness value of the prediction area when the transmission window reaches the prediction area; Based on the actual fog thickness value, the heating power of the mirror back heating unit is adjusted so that the predicted area returns to the fogged state at a preset cooling rate after the transmission window leaves. Obtain the touch pressure level and dwell time of the transmission window within the prediction area; adjust the thermal inertia compensation coefficient in the mapping table according to the touch pressure level and the dwell time, wherein the higher the touch pressure level or the longer the dwell time, the larger the thermal inertia compensation coefficient, and the target heating power in subsequent predictions is reduced accordingly based on the lookup result in the mapping table.

Images