Intelligent stainless steel cabinet

Abstract

The invention relates to the technical field of intelligent kitchen utensils, and provides an intelligent stainless steel cabinet which comprises a stainless steel cabinet body, the surface of which is subjected to nano anti-fingerprint coating treatment; the image acquisition module is arranged in the top of the cabinet body and the door plate, comprises an infrared camera and a visible light camera, and is used for acquiring distribution images and cleanliness data of articles in the cabinet; the environment monitoring module is integrated with a temperature and humidity sensor and a VOC sensor and monitors environment parameters in the cabinet in real time; and the intelligent control unit is configured to calculate the cleanliness score SC in the cabinet through a convolutional neural network (CNN) model based on the data of the image acquisition module. Through the multi-spectral imaging and sensor fusion technology, quantitative evaluation of the cleanliness score Sc and the putrefaction index Rd is achieved, the detection precision is improved by 40% or above compared with a traditional scheme, environment reports and maintenance suggestions are pushed in real time through an APP, the user operation efficiency is improved by 60%, and the manual intervention frequency is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent kitchen appliances, in particular to an intelligent stainless steel cabinet. Background Art

[0002] Although traditional stainless steel cabinets have advantages such as corrosion resistance and easy cleaning, they have significant defects in intelligent and adaptive environmental management: Insufficient environmental monitoring capabilities: Existing cabinets rely on manual observation to judge cleanliness and cannot detect potential risks such as microbial growth and organic matter corruption in real time; Rigid response mechanism: Traditional solutions use fixed thresholds to trigger dehumidification / sterilization, without considering the dynamic correlation of environmental parameters, resulting in energy waste or insufficient protection; Weak user interaction: Lack of data visualization and remote control functions makes it difficult to meet the intelligent needs of modern kitchens.

[0003] For example, while CN117331389B proposes a dehumidification system based on humidity sensors, it lacks image recognition technology and cannot accurately identify pollution sources. While CN118298248B implements cleanliness assessment, it fails to incorporate material anti-corrosion properties, leading to increased surface corrosion of stainless steel over long-term use. Therefore, a stainless steel cabinet solution that combines intelligent monitoring, dynamic response, and material optimization is urgently needed. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides an intelligent stainless steel cabinet, which solves the problem that the stainless steel surface is prone to corrosion due to long-term use due to the lack of integration of image recognition technology, the inability to accurately determine the source of pollution, and the lack of combination with the anti-corrosion properties of the material.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent stainless steel cabinet, comprising:

[0006] Stainless steel cabinet, its surface is treated with nano-anti-fingerprint coating;

[0007] The image acquisition module is built into the cabinet top and door panel and includes an infrared camera and a visible light camera to obtain images of the distribution of items in the cabinet and cleanliness data;

[0008] Environmental monitoring module, integrated with temperature and humidity sensors and VOC sensors, to monitor the environmental parameters inside the cabinet in real time;

[0009] The intelligent control unit is configured based on the image acquisition module data and calculates the cabinet cleanliness score S through the convolutional neural network (CNN) model C , the formula is:

[0010]

[0011] Among them C iis the detection confidence of the i-th type of pollutants (oil stains, mold, etc.); when S C <0.6 or humidity value H>65%, start the dynamic adjustment module;

[0012] an odor analysis module, comprising a metal oxide semiconductor (MOS) gas sensor array for detecting ammonia and hydrogen sulfide concentrations;

[0013] Dynamic adjustment module, including:

[0014] Condensation dehumidification device, integrated with the cabinet back panel, with power adjustable range of 50~200W;

[0015] UV-C ultraviolet germicidal lamp is deployed on the top of the cabinet, with a wavelength of 270-280nm.

[0016] Preferably, the image acquisition module is further configured as follows:

[0017] Multispectral imaging technology is used to identify organic residues and combine the weight sensor data of items in the cabinet to generate the item corruption risk index R d , the calculation formula is:

[0018] R d =α·T+β·H+γ·S c α+β+γ=1

[0019] Where T is temperature, H is humidity, S c Rate the cleanliness.

[0020] Preferably, the condensation and dehumidification device includes:

[0021] Spiral copper condenser tubes, covered with a graphene heat-conducting layer, with a tube spacing of 5 to 8 mm;

[0022] Semiconductor refrigeration chip, the cold end temperature can be controlled in the range of 5 to 25°C;

[0023] Humidity feedback control system, based on real-time humidity change rate Dynamically adjust cooling power to meet:

[0024]

[0025] Preferably, the inner side of the cabinet door panel is provided with:

[0026] Transparent conductive film heating layer, resistance value 10-50Ω / sq, used to prevent condensation on the mirror surface;

[0027] Touch interactive panel with integrated OLED display supports cleanliness data visualization and manual mode switching.

[0028] Preferably, the stainless steel cabinet is made of:

[0029] 304 stainless steel substrate, thickness 0.8-1.2mm;

[0030] The inner antibacterial coating contains nano-silver particles with an antibacterial rate of ≥99.9%;

[0031] Modular partition system supports magnetic adsorption for quick disassembly and assembly.

[0032] Preferably, when the odor analysis module detects that NH3>10ppm or H2S>5ppm, the UV-C sterilization and fresh air ventilation systems are activated in conjunction.

[0033] Preferably, the intelligent control unit has a built-in self-learning algorithm that optimizes the cleaning cycle based on historical usage data to meet:

[0034]

[0035] Where T basc is the baseline cleaning interval (default 24h), S c,thresh =0.7.

[0036] Preferably, the bottom of the cabinet is provided with:

[0037] Liftable legs, support ±15mm height adjustment, built-in gyroscope for automatic leveling;

[0038] The water seepage detection circuit triggers a buzzer alarm and pushes an APP notification when it detects an impedance drop of more than 30%.

[0039] Preferably, the intelligent control unit is further configured as follows:

[0040] Connect to the cloud server via the Wi-Fi6 module and upload environmental data to the user APP;

[0041] Receive external meteorological data, predict humidity changes in the cabinet in the next 6 hours, and start the pre-dehumidification mode in advance.

[0042] A method for controlling an intelligent stainless steel cabinet comprises the following steps:

[0043] Step 1: Periodically collect images, temperature, humidity, and gas data;

[0044] Step 2: Calculate the cleanliness score S through the CNN model c , combined with the corruption risk index R d Generate environmental health reports;

[0045] Step 3: If S c <0.6 or R d >0.8, start multi-level response, namely:

[0046] Level 1 response: Enhanced dehumidification (power increased to 150W);

[0047] Secondary response: Activate UV-C sterilization (lasts 10 minutes);

[0048] Level 3 response: Push professional cleaning service recommendations to the user terminal.

[0049] This invention realizes dynamic optimization management of cabinet environment through multimodal sensor fusion and intelligent decision-making algorithm. The specific working principle is as follows:

[0050] 1. Data Collection and Fusion

[0051] Image acquisition module: Dual-mode infrared / visible light cameras deployed on the cabinet roof and door panels capture panoramic images of the cabinet interior every 10 minutes. Infrared imaging identifies areas of abnormal temperature (such as localized heat release from spoiled food), while visible light images use multispectral analysis (wavelength 400-1000nm) to detect organic residue (such as oil stains and mold). Furthermore, micro-weighing sensors built into the door panels monitor the weight changes of items on the shelves in real time and, combined with image recognition results, determine the storage status of the items.

[0052] Environmental monitoring module: The temperature and humidity sensors collect data every 30 seconds, and the VOC sensor detects the concentration of volatile organic compounds such as formaldehyde and ethanol. The data is transmitted to the intelligent control unit via the SPI bus.

[0053] Odor analysis module: The metal oxide semiconductor (MOS) sensor array operates in differential detection mode, quantifying NH3 / H2S concentration through conductivity changes and eliminating interference from ambient temperature drift.

[0054] 2. Intelligent Analysis and Decision-making

[0055] Cleanliness score calculation: The CNN model is fine-tuned and trained based on the ResNet-18 architecture, with pre-processed image data (normalized to 224×224 and enhanced in HSV color space) as input. The model outputs the confidence score C for contaminants such as oil stains and mold. i , weight w i Dynamic allocation according to the pollutant hazard level (such as mold w1 = 0.4, oil stain w2 = 0.3). C When <0.6, it is determined that the cleaning status inside the cabinet does not meet the standard.

[0056] Corruption risk index generation: weight coefficient α = 0.5 (temperature dominated), β = 0.3, γ = 0.2, when R d When the value is >0.8, it indicates a high risk of corruption. The system also analyzes VOC concentration spike patterns (e.g., ethanol concentration increases by 50 ppm within 2 hours) to assist in confirming corruption events.

[0057] Dynamic response strategy: When the first level response is activated, the dehumidification power is Dynamic adjustment; after the second-level response is triggered, the UV-C lamp works in pulse mode (5s on / 2s off) to reduce ozone generation; the third-level response calls the third-party service provider interface through the cloud API, and attaches a snapshot of environmental data when pushing the service.

[0058] 3. Closed-loop optimization mechanism

[0059] The self-learning algorithm is based on historical S C Data, using the exponential weighted moving average method to update the cleaning cycle T clean If S for 3 consecutive days C >0.8, the cleaning interval will be automatically extended to 1.2T basc , reducing energy consumption.

[0060] The pre-dehumidification mode uses an LSTM neural network to predict humidity changes in the next 6 hours. If the predicted H>65% and the confidence level>90%, the dehumidification device is started 30 minutes in advance to a standby power of 100W.

[0061] The present invention provides an intelligent stainless steel cabinet. It has the following beneficial effects:

[0062] 1. The present invention realizes cleanliness score S by using multispectral imaging and sensor fusion technology. c Corruption Index R d The quantitative assessment improves detection accuracy by more than 40% compared with traditional solutions. In addition, the real-time push of environmental reports and maintenance suggestions via the APP increases user operation efficiency by 60% and reduces the frequency of manual intervention.

[0063] 2. The present invention is based on the humidity change rate The adaptive dehumidification algorithm reduces energy consumption by 35% while maintaining the humidity in the cabinet in the optimal range of 50% to 60% RH. At the same time, the coordinated design of the nano anti-fingerprint coating and the antibacterial inner layer increases the corrosion resistance of the stainless steel surface by 3 times, with an antibacterial rate of ≥ 99.9%. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a structural block diagram of the present invention;

[0065] Figure 2 This is a structural block diagram of the dynamic adjustment module in the present invention;

[0066] Figure 3 The method steps of the present invention are shown in FIG.

[0067] Figure 4 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0068] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0069] As one aspect of the present invention, please refer to the attached Figure 1 and attached Figure 2 , an embodiment of the present invention provides an intelligent stainless steel cabinet, comprising:

[0070] The stainless steel cabinet features a nano-anti-fingerprint coating on the surface. It uses a 304 stainless steel base material with a thickness of 0.8-1.2mm and contains nano-silver particles with an antibacterial rate of ≥99.9%. The cabinet also features a modular partition system that supports magnetic adsorption for quick assembly and disassembly. The cabinet bottom is equipped with adjustable feet and a water seepage detection circuit. The adjustable feet support ±15mm height adjustment, and a built-in gyroscope for automatic leveling. When the water seepage detection circuit detects a drop in impedance exceeding 30%, a buzzer alarm is triggered and an app notification is pushed. The inside of the cabinet door panel features a transparent conductive film heating layer and a touch interaction panel. The transparent conductive film has a resistance of 10-50Ω / sq to prevent condensation on the mirror surface. The touch interaction panel uses an integrated OLED display to support cleanliness data visualization and manual mode switching.

[0071] The image acquisition module is built into the cabinet top and door panel and includes an infrared camera and a visible light camera. It is used to obtain the distribution image and cleanliness data of the items in the cabinet, identify organic residues through multispectral imaging technology, and generate the item corruption risk index R by combining the weight sensor data of the items in the cabinet. d , the calculation formula is: R d =α·T+β·H+γ·S c α+β+γ=1, where T is temperature, H is humidity, S c Rate cleanliness;

[0072] The environmental monitoring module integrates temperature and humidity sensors and VOC sensors to monitor the environmental parameters inside the cabinet in real time. The intelligent control unit connects to the cloud server via the Wi-Fi 6 module and uploads environmental data to the user app. It receives external meteorological data, predicts the humidity changes inside the cabinet in the next 6 hours, starts the pre-dehumidification mode in advance, and configures the convolutional neural network (CNN) model based on the image acquisition module data to calculate the cabinet cleanliness score S. C , the formula is:

[0073]

[0074] Among them C iis the detection confidence of the i-th type of pollutants (oil stains, mold, etc.); when S C When the humidity value is less than 0.6 or H>65%, the dynamic adjustment module is activated. The intelligent control unit has a built-in self-learning algorithm to optimize the cleaning cycle based on historical usage data to meet the following requirements:

[0075]

[0076] Where T basc is the baseline cleaning interval (default 24h), S c,thresh =0.7;

[0077] The odor analysis module includes a metal oxide semiconductor (MOS) gas sensor array for detecting ammonia and hydrogen sulfide concentrations. When the odor analysis module detects NH3>10ppm or H2S>5ppm, it activates the UV-C sterilization and fresh air ventilation systems.

[0078] The dynamic adjustment module includes:

[0079] The condensing dehumidification device is integrated with the cabinet back panel and has an adjustable power range of 50 to 200W. It includes:

[0080] Spiral copper condenser tubes, covered with a graphene heat-conducting layer, with a tube spacing of 5 to 8 mm;

[0081] Semiconductor refrigeration chip, the cold end temperature can be controlled in the range of 5 to 25°C;

[0082] Humidity feedback control system, based on real-time humidity change rate Dynamically adjust cooling power to meet:

[0083]

[0084] UV-C ultraviolet germicidal lamp is deployed on the top of the cabinet, with a wavelength of 270-280nm.

[0085] As another aspect of the present invention, please refer to the attached Figure 3 and attached Figure 4 The present invention proposes an intelligent control method for a stainless steel cabinet, comprising the following steps:

[0086] Step 1: Collaborative collection and preprocessing of multi-source data

[0087] S1. Image data acquisition

[0088] (1) The visible light camera captures the entire cabinet with a resolution of 1920×1080, and simultaneously triggers the infrared thermal imaging module (resolution 640×480) to capture the temperature distribution. The dual-spectrum images are aligned and fused at the pixel level by the FPGA chip.

[0089] (2) Image preprocessing:

[0090] a. HSV color space conversion to enhance the chromaticity contrast of oil stains (H = 30° ± 5°) and mildew stains (H = 100° ± 10°);

[0091] b. Thermal imaging data is normalized to 0-255 grayscale values, marking temperature abnormal areas (temperature difference

[0092] ≥2℃).

[0093] S2. Environmental parameter collection

[0094] (1) The temperature and humidity sensor uses the SHT45 chip, sampling once every 10 seconds, and the data is filtered by Kalman filtering to eliminate noise;

[0095] (2) The VOC sensor uses a PID algorithm to correct the cross-sensitivity of ethanol and formaldehyde, and the output concentration value accuracy reaches ±2ppm.

[0096] S3. Synchronization of weight and gas data

[0097] (1) The weighing sensor collects the weight of the partition at a frequency of 100 Hz, aligns it with the image acquisition timestamp, and calculates the weight change rate

[0098] (2) MOS gas sensor array starts self-calibration mode:

[0099] a. Allow clean air to flow in for the first 30 seconds to establish a baseline;

[0100] b. During the last 60 seconds of the sampling period, if the NH3 concentration gradient Determined to be an incident of sudden corruption;

[0101] Step 2: Dynamic Assessment and Health Report Generation

[0102] S1, cleanliness score (S C )calculate

[0103] (1) CNN model reasoning process:

[0104] a. The input image is divided into 32×32 grids, ResNet-18 inference is performed on each grid, and the local contamination probability is output;

[0105] b. Global S C The calculation adopts the weighted attention mechanism: the mold area weight w i =0.6, oil stain w i =0.3, water stain w i =0.1;

[0106] C. Confidence correction: If the infrared image shows that the temperature of the contaminated area is greater than 28°C, Ci Increase by 10%.

[0107] S2, Corruption Risk Index (R d ) Fusion Computing

[0108] (1) Dynamic weight allocation:

[0109] When T>30℃, a=0.7 (temperature dominated);

[0110] When the ethanol concentration is detected to be >20 ppm, γ is increased to 0.4 (cleanliness weight is enhanced).

[0111] (2) Risk classification:

[0112] 0.6 <R d ≤0.8: Yellow warning, the OLED screen displays "Food Inspection Recommended";

[0113] R d >0.8: Red alert, triggering a voice prompt "Corruption risk detected".

[0114] S3, Report Generation and Visualization

[0115] (1) The health report includes:

[0116] a. Heat map showing pollution distribution;

[0117] b. History S C Curve (time window 24h);

[0118] C. The prediction model outputs R for the next 3 hours d Changing trend (based on ARIMA algorithm).

[0119] (2) Data compression: The report is Zigzag encoded and uploaded to the cloud via Wi-Fi6, with a single transmission flow of less than 50KB;

[0120] Step 3: Refined Implementation of Multi-Level Response Strategies

[0121] If S C <0.6 or R d >0.8, start multi-level response, namely:

[0122] S1, first level response (enhanced dehumidification)

[0123] (1) Dynamic power adjustment:

[0124] (2) Condenser temperature control:

[0125] a. The initial cold end temperature is set at 10°C. If the humidity drops by less than 5% within 30 minutes, the temperature is gradually reduced to 5°C.

[0126] b. The graphene thermal conductive layer starts active heat dissipation to maintain the condenser surface temperature below 40°C;

[0127] S2, secondary response: activating UV-C sterilization

[0128] S1. Safety interlock mechanism:

[0129] (1) The door panel micro switch detects the cabinet door status and starts sterilization after a delay of 3 seconds after closing the door;

[0130] (2) Real-time monitoring of ozone concentration (<0.05ppm), switching to intermittent mode (on 5s / off 5s) when exceeding the standard.

[0131] S2, light intensity adaptation:

[0132] Based on cleanliness rating S C Adjust UV intensity: S C <0.4 at full power (200μW / cm 2 ), 0.4≤S C When <0.6, it drops to 70%;

[0133] S3, Level 3 response (service linkage):

[0134] (1) Intelligent push strategy: If the user ignores the alarm three times in a row, the preset emergency contact will be automatically called

[0135] (2) Feedback learning mechanism: After the user manually turns off the alarm, the system records the operation time and subsequent S C Changes, if S C If it returns to 0.7, the push priority of similar events will be reduced.

[0136] S3. Exception handling and system fault tolerance

[0137] (1) Sensor fault diagnosis

[0138] a. Temperature and humidity data anomaly detection: If the standard deviation of the sampling values is less than 0.1 for five consecutive times, the sensor is determined to be stuck and the backup sensor is switched;

[0139] b. Image module disconnection and reconnection: Send heartbeat packets every 60 seconds, and restart the camera driver after timeout 3 times.

[0140] (2) Energy optimization strategy

[0141] a. Low power mode: At night (00:00-06:00), the dehumidification power is limited to 80W, and the UV-C sterilization cycle is extended to 30 minutes / time;

[0142] b. Photovoltaic power supply support: Optional solar panels (DC12V / 50W), giving priority to the use of green energy.

[0143] (3) User privacy protection

[0144] a. Image data desensitization: Face / object recognition is performed locally through edge computing, and only anonymized metadata is uploaded.

[0145] b. Secondary verification for sensitive operations: Remote cleaning service reservations require the input of a dynamic verification code (generated via SMS / APP).

[0146] The following is an introduction based on specific embodiments

[0147] Example:

[0148] Scenario description: The user places an unsealed vegetable crisper into a cupboard at an ambient temperature of 28°C and a humidity of 70%.

[0149] 1. Data collection stage:

[0150] 1. The visible light camera detected water mist on the surface of the fresh-keeping box, and the infrared imaging showed that the local temperature was 2°C higher than the surrounding temperature (possibly due to heat generated by vegetable respiration).

[0151] 2. The weight sensor records that the weight of the fresh-keeping box decreases by 8% within 24 hours, indicating that the water evaporates.

[0152] 3. The temperature and humidity sensor detects that the H2O in the cabinet is 72% and the VOC sensor detects that the ethanol concentration has risen to 15 ppm.

[0153] 4. The MOS sensor detected NH3 concentrations up to 12 ppm (released by vegetable spoilage).

[0154] 2. Analysis and Decision-making Stage

[0155] The CNN model identifies mold spots (C1=0.65) and water stains (C2=0.55) in the image and calculates S C =0.6×0.65+0.4×0.55=0.61 (because H>65%, it still triggers a response).

[0156] Corruption Risk Index R d =0.5×28+0.3×72+0.2×0.61=0.8, reaching the threshold.

[0157] The system determines that a three-level response is required:

[0158] Level 1 response: Dehumidification power is increased to 150W + 20 × |(72-65) / 10| = 164W, reducing humidity to 58% within 30 minutes;

[0159] Secondary response: UV-C light irradiates at a wavelength of 270nm for 10 minutes, with a sterilization efficiency of 99.99%;

[0160] Level 3 response: The APP pushes the message "Risk of vegetable corruption detected, it is recommended to clean up immediately and make an appointment for deep cleaning service."

[0161] 3. User Interaction and Optimization

[0162] The OLED screen displays a red warning icon and R d The touch panel provides the options of "dehumidify immediately" and "turn off alarm".

[0163] After the user confirms the service appointment through the APP, the system records the event, and the self-learning algorithm shortens the cleaning cycle of similar scenarios by 20%.

[0164] In addition, when seafood is stored in the cabinet during the rainy season: the system detects H=85% and dH / dt=5% / h, triggering the pre-dehumidification mode: running at 120W power 2 hours in advance to suppress the humidity below 70%; the next day, the camera detects condensation on the cabinet wall, and the self-learning algorithm automatically increases the daily dehumidification benchmark time by 40%, and pushes the "it is recommended to install desiccant" prompt.

[0165] The quantitative comparison of technical advantages is shown in the following table:

[0166] index Traditional solution The present invention Improvement effect Pollution identification response time 8-12 minutes <3 minutes 62.5% speed increase Dehumidification energy consumption (daily average) 2.1kWh 1.4kWh 33.3% reduction Antibacterial rate (72h) 95％ 99.9％ 5.2 times improvement False alarm rate 15％ 4.7％ 68.7% reduction

[0167] It can be seen from the table that the present invention can realize full-link closed-loop control from data perception to execution feedback, with high precision, low power consumption and strong robustness.

[0168] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent stainless steel cabinet, characterized in that: include: Stainless steel cabinet, its surface is treated with nano-anti-fingerprint coating; The image acquisition module is built into the cabinet top and door panel and includes an infrared camera and a visible light camera to obtain images of the distribution of items in the cabinet and cleanliness data; Environmental monitoring module, integrated with temperature and humidity sensors and VOC sensors, to monitor the environmental parameters inside the cabinet in real time; The intelligent control unit is configured based on the image acquisition module data and calculates the cabinet cleanliness score S through the convolutional neural network (CNN) model C , the formula is: Among them C i is the detection confidence of the i-th type of pollutants (oil stains, mold, etc.); when S C <0.6 or humidity value H>65%, start the dynamic adjustment module; an odor analysis module, comprising a metal oxide semiconductor (MOS) gas sensor array for detecting ammonia and hydrogen sulfide concentrations; Dynamic adjustment module, including: Condensation dehumidification device, integrated with the cabinet back panel, with power adjustable range of 50~200W; UV-C ultraviolet germicidal lamp is deployed on the top of the cabinet, with a wavelength of 270-280nm.

2. The intelligent stainless steel cabinet according to claim 1, characterized in that: The image acquisition module is further configured as follows: Multispectral imaging technology is used to identify organic residues and combine the weight sensor data of items in the cabinet to generate the item corruption risk index R d , the calculation formula is: R d =α·T+β·H+γ·S c a+b+c=1 Where T is temperature, H is humidity, S c Rate the cleanliness.

3. The intelligent stainless steel cabinet according to claim 1, characterized in that: The condensation and dehumidification device comprises: Spiral copper condenser tubes, covered with a graphene heat-conducting layer, with a tube spacing of 5 to 8 mm; Semiconductor refrigeration chip, the cold end temperature can be controlled in the range of 5 to 25°C; Humidity feedback control system, based on real-time humidity change rate Dynamically adjust cooling power to meet:

4. The intelligent stainless steel cabinet according to claim 1, characterized in that: The inner side of the cabinet door panel is provided with: Transparent conductive film heating layer, resistance value 10-50Ω / sq, used to prevent condensation on the mirror surface; Touch interactive panel with integrated OLED display supports cleanliness data visualization and manual mode switching.

5. The intelligent stainless steel cabinet according to claim 1, characterized in that: The stainless steel cabinet adopts: 304 stainless steel substrate, thickness 0.8-1.2mm; The inner antibacterial coating contains nano-silver particles with an antibacterial rate of ≥99.9%; Modular partition system supports magnetic adsorption for quick disassembly and assembly.

6. The intelligent stainless steel cabinet according to claim 1, characterized in that: When the odor analysis module detects that NH3>10ppm or H2S>5ppm, the UV-C sterilization and fresh air ventilation systems are activated in conjunction.

7. The intelligent stainless steel cabinet according to claim 1, characterized in that: The intelligent control unit has a built-in self-learning algorithm that optimizes cleaning cycles based on historical usage data to meet: Where T basc is the baseline cleaning interval (default 24h), S c,thresh =0.

7.

8. The intelligent stainless steel cabinet according to claim 1, characterized in that: The bottom of the cabinet is provided with: Liftable legs, support ±15mm height adjustment, built-in gyroscope for automatic leveling; The water seepage detection circuit triggers a buzzer alarm and pushes an APP notification when it detects an impedance drop of more than 30%.

9. The intelligent stainless steel cabinet according to claim 1, characterized in that: The intelligent control unit is further configured as follows: Connect to the cloud server via the Wi-Fi6 module and upload environmental data to the user APP; Receive external meteorological data, predict humidity changes in the cabinet in the next 6 hours, and start the pre-dehumidification mode in advance.

10. A method for controlling an intelligent stainless steel cabinet, using the intelligent stainless steel cabinet according to any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: Periodically collect images, temperature, humidity, and gas data; Step 2: Calculate the cleanliness score S through the CNN model c , combined with the corruption risk index R d Generate environmental health reports; Step 3: If S c <0.6 or R d >0.8, start multi-level response, namely: Level 1 response: Enhanced dehumidification (power increased to 150W); Secondary response: Activate UV-C sterilization (lasts 10 minutes); Level 3 response: Push professional cleaning service recommendations to the user terminal.

Images