Intelligent fireproof and high-temperature steam oil smoke cleaning integrated smoke hood

Abstract

The invention provides an intelligent fireproof and high-temperature steam cleaning oil fume integrated hood, and relates to the technical field of range hoods, and the intelligent fireproof and high-temperature steam cleaning oil fume integrated hood comprises an oil fume diffusion monitoring module used for obtaining an oil fume image; through a difference image method, according to the difference between the oil smoke image and the reference image, obtaining oil smoke diffusion data; vOC data are obtained; calculating a lampblack diffusion quantized value; the intelligent cleaning module is used for determining the optimal steam release temperature and the optimal steam release amount through an improved artificial neural network; according to the optimal steam release temperature and the optimal steam release amount, the steam release amount and temperature of the range hood are dynamically adjusted based on a PID control system of a radial basis function neural network, and the smoke hood is cleaned by releasing high-temperature steam; the flame monitoring module is used for acquiring a cooking bench image through a camera; through a convolutional neural network, identifying a flame combustion degree, and determining a kitchen fire occurrence probability value; and the fire extinguishing module is used for spraying water to extinguish the kitchen.

Description

Technical Field

[0001] This invention relates to the field of range hood technology, and in particular to an integrated intelligent fireproof and high-temperature steam cleaning range hood. Background Technology

[0002] High-temperature steam technology generates high-temperature steam by heating water. During the cleaning process, the steam rapidly dissolves and emulsifies grease and grime upon contact with the surface of the range hood hood. Unlike traditional chemical cleaners, high-temperature steam cleaning is completely chemical-free, making it a more environmentally friendly and healthier cleaning method. Chemical cleaners can not only pollute the environment, but prolonged exposure can also have adverse effects on human health, while high-temperature steam cleaning completely avoids these problems. Due to the strong penetrating and dissolving power of high-temperature steam, it can effectively penetrate the grease layer on the surface of the range hood, quickly decomposing and removing grease. This method not only cleans the range hood hood but also reduces the growth of harmful microorganisms such as bacteria and viruses, further improving the cleanliness and safety of the kitchen.

[0003] Many current intelligent cleaning technologies rely on preset cleaning parameters, which are usually fixed, such as steam temperature and steam output. This fixed parameter design cannot be flexibly adjusted according to the actual diffusion of cooking fumes, resulting in suboptimal cleaning efficiency. Especially when there is little cooking fume, excessive steam release may occur, wasting energy and potentially causing over-wetting of the fume hood surface, thus negatively impacting cleaning effectiveness. Conversely, when cooking fumes are heavily accumulated, a fixed cleaning intensity and steam output may not completely remove stubborn grease, leading to incomplete cleaning. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide an integrated intelligent fireproof and high-temperature steam cleaning fume hood that can solve the technical problem that existing intelligent cleaning technologies rely on preset cleaning parameters and cannot be flexibly adjusted according to the actual diffusion of fumes, resulting in the inability to optimize cleaning efficiency.

[0005] This invention provides an integrated intelligent fireproof and high-temperature steam cleaning fume hood, comprising:

[0006] The oil fume diffusion monitoring module is used to acquire oil fume images through a camera; obtain oil fume diffusion data based on the difference between the oil fume images and a reference image using the difference image method; arrange VOC sensors on the left, right, and top edges of the range hood, and use the VOC sensors to acquire VOC data; calculate the oil fume diffusion quantification value based on the oil fume diffusion data and the VOC data.

[0007] The cleaning judgment module is used to determine whether the oil fume diffusion quantification value is greater than the preset diffusion quantification value; if so, the intelligent cleaning module is invoked to perform intelligent cleaning; otherwise, the oil fume diffusion monitoring module is invoked to continue monitoring.

[0008] The intelligent cleaning module is used to determine the optimal steam release temperature and the optimal steam release amount by taking the oil fume diffusion data and the VOC data as input and through an improved artificial neural network. Based on the optimal steam release temperature and the optimal steam release amount, the PID control system based on the radial basis function neural network dynamically adjusts the steam release amount and temperature of the range hood, and cleans the range hood by releasing high-temperature steam.

[0009] The flame monitoring module is used to acquire images of the stove through a camera; based on the images of the stove, a convolutional neural network is used to identify the degree of flame combustion and determine the probability value of a kitchen fire.

[0010] The fire prevention judgment module is used to determine whether the probability value of a kitchen fire is greater than a preset probability value; if so, the fire extinguishing module is invoked to extinguish the fire; otherwise, the flame monitoring module is invoked to continue monitoring.

[0011] The fire extinguishing module is used to spray water to extinguish fires in the kitchen.

[0012] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0013] (1) In this embodiment of the invention, the range hood can automatically determine whether cleaning needs to be started based on the actual diffusion of oil fumes, avoiding the arbitrariness of the frequency and intensity of manual cleaning, ensuring that the range hood is automatically cleaned when oil fumes accumulate to a certain level, and reducing the frequency and error of manual operation.

[0014] (2) In this embodiment of the invention, the steam release and temperature of the range hood can be flexibly adjusted according to the actual diffusion of oil fumes during the cleaning process to achieve precise cleaning and improve cleaning efficiency.

[0015] (3) In this embodiment of the invention, an intelligent fire prevention mechanism is integrated to ensure the safety of the kitchen in a high-temperature and high-fume environment and reduce the risk of fire caused by the accumulation of fumes or during cooking. Attached Figure Description

[0016] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0017] Figure 1 This is a schematic diagram of the structure of an integrated intelligent fireproof and high-temperature steam cleaning fume hood provided in an embodiment of the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] The intelligent fireproof and high-temperature steam cleaning integrated fume hood provided by the present invention will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0020] Reference manual attached Figure 1 The diagram shows a structural schematic of an integrated intelligent fireproof and high-temperature steam cleaning fume hood provided by an embodiment of the present invention.

[0021] This invention provides an integrated intelligent fireproof and high-temperature steam cleaning fume hood 10, which may include:

[0022] The oil fume diffusion monitoring module 101 is used to acquire oil fume images via a camera. Oil fume diffusion data is obtained by using the difference image method to determine the difference between the oil fume image and a reference image. VOC sensors are placed on the left, right, and top edges of the range hood to acquire VOC data. Based on the oil fume diffusion data and the VOC data, a quantitative value for oil fume diffusion is calculated.

[0023] In one possible implementation, the oil fume diffusion monitoring module 101 is specifically used for:

[0024] The difference image method is used to calculate the difference image between the oil fume image and the reference image:

[0025]

[0026] Among them, D N Represents the difference image, F N F0 represents the current image of cooking fumes, F0 represents the reference image, and (x,y) represents the coordinates of the pixel position in the image.

[0027] It should be noted that by calculating the difference between the current oil fume image and the reference image, changes in the oil fume area can be clearly identified, the diffusion of oil fumes can be accurately detected, and background and environmental interference can be avoided.

[0028] The difference image is binarized to generate a mask image of the oil fume area:

[0029]

[0030] Among them, R N This represents the masked image of the oil fume area, and T represents the binarization threshold.

[0031] It should be noted that binarizing the difference image to generate a mask image of the oil fume region can effectively separate the oil fume region from the non-oil fume region, simplifying subsequent processing and improving computational efficiency.

[0032] Erosion is performed on discrete noise points in the mask image of the oil fume area to obtain an eroded mask image:

[0033]

[0034] Among them, E N This represents the erosion mask image, where m represents a specific location, and B represents the erosion convolution template. m This indicates the state of the erosion convolution template after it slides to position m on the mask image of the oil fume region. This represents the erosion operation. The specific rules for the erosion operation are: when all pixels within the area covered by the convolution template are 1, the output pixel value is 1. Otherwise, the output pixel value is 0.

[0035] Erosion is a morphological operation in image processing, primarily used to remove small noise points in binary images or to reduce the boundaries of objects. During erosion, a convolutional template (or structuring element) slides across the image; the output value is 1 only when the structuring element perfectly matches a region in the image, otherwise it is 0. In this way, erosion can reduce fine noise and artifacts, extracting clearer target regions.

[0036] It should be noted that by performing an erosion operation on the mask image of the oil fume area, small noise and discrete points in the image are removed, reducing misjudgments, ensuring the accuracy of the oil fume area, and avoiding the impact of noise interference on subsequent analysis.

[0037] The erosion mask image is processed by median filtering to obtain the filtered mask image.

[0038] Median filtering is a common image denoising technique that applies a sliding window around each pixel in an image and replaces the value of the center pixel with the median value of the pixels within the window. This method effectively removes salt-and-pepper noise and other abrupt noise while preserving the image's edge features, thus improving image quality. Median filtering has strong noise suppression capabilities and does not blur image details.

[0039] It should be noted that median filtering was applied to the etched mask image to further remove irregular noise, smooth the image, and ensure greater accuracy and stability in the oil fume area.

[0040] Perform a dilation operation on the filtered mask image to obtain a dilated mask image:

[0041]

[0042] Among them, S N Bs represents the dilation mask image, and Bs represents the dilation convolution template. m This represents the state of the dilated convolution template after it slides to position m on the mask image of the oil fume region, where M is the template position. N Represents the filtered mask image. This represents the dilation operation. The specific rule for the dilation operation is: if there is a pixel with a value of 1 within the region covered by the convolution template, the output pixel value is 1. Otherwise, the output pixel value is 0.

[0043] Dilation is the dual operation of erosion. It is another fundamental operation in morphological image processing. Dilation enlarges objects in an image, filling in small holes or gaps within them. Appropriate use of dilation can improve the connectivity of objects and enhance subsequent image analysis, especially when processing areas with cooking fumes, allowing for more complete connection of the target region.

[0044] It should be noted that the expansion operation can enhance the connectivity of the fume area, fill small gaps, and make the fume area more complete, thereby improving the detection accuracy and reliability of the fume area.

[0045] The difference image is multiplied point-by-point with the dilation mask image to determine the image of the oil fume diffusion area:

[0046]

[0047] Among them, O N This image represents the area where cooking fumes diffuse.

[0048] It should be noted that by multiplying the dilatation mask image and the difference image point by point and combining the features of both, a more accurate image of the oil fume diffusion area is obtained, further optimizing the detection and analysis effect of the oil fume area.

[0049] The grayscale values ​​of each pixel in the image of the oil fume diffusion area are accumulated to obtain the oil fume diffusion data.

[0050] It should be noted that by accumulating the grayscale values ​​of each pixel in the image of the fume diffusion area, fume diffusion data was obtained, providing a precise basis for subsequent cleaning judgment and intelligent control.

[0051] In one possible implementation, the oil fume diffusion monitoring module 101 is specifically used for:

[0052] VOC sensors are placed on the left, right, and top edges of the range hood.

[0053] Among them, VOC (volatile organic compound) sensors are used to detect the concentration of volatile organic compounds in the air. VOC sensors can sense and analyze harmful gases in the air, such as formaldehyde, benzene, and toluene, which may be released in the kitchen from sources such as cooking fumes, combustion, or cleaning agents. VOC sensors are widely used in air quality monitoring systems to help achieve intelligent control, such as adjusting the operating status of range hoods or air purification systems based on real-time VOC data to ensure fresh and safe air.

[0054] It's important to note that VOC sensors are used because image processing methods can capture most of the visible area of ​​oil fume diffusion. However, oil fumes contain many colorless substances such as alkanes, alcohols, and ketones, which cannot be detected by machine vision. VOC sensors can overcome this limitation by monitoring and identifying volatile organic compounds in the air in real time, providing more comprehensive oil fume monitoring.

[0055] VOC data is acquired through each VOC sensor.

[0056] Based on the VOC data measured by each VOC sensor, a weighted sum is performed to calculate the comprehensive VOC data:

[0057]

[0058] Where V represents the comprehensive VOC data value, V l This represents the VOC data value measured by the VOC sensor on the left edge, V r This represents the VOC data value measured by the VOC sensor on the right edge, V u λ1 represents the VOC data value measured by the VOC sensor on the upper edge, λ2 represents the weighting coefficient of the VOC sensors on the left and right edges, and λ3 represents the weighting coefficient of the VOC sensor on the upper edge.

[0059] Those skilled in the art can set the weight coefficients of the VOC sensors on the left and right sides and the VOC sensor on the top side according to the actual situation; this invention does not impose any limitations.

[0060] In this embodiment of the invention, the measurement results of VOC sensors at different locations (left, right and top) may be affected by environmental factors and uneven distribution of oil fumes. By using a weighted summation method, different weights can be assigned to each sensor according to its importance or accuracy. By weighted summation of the VOC data measured by each VOC sensor, the comprehensive VOC data can be calculated to more comprehensively reflect the overall diffusion of oil fumes.

[0061] In one possible implementation, the oil fume diffusion monitoring module 101 is specifically used for:

[0062] We performed a weighted fitting of the oil fume diffusion data and VOC data to determine the quantitative value of oil fume diffusion:

[0063]

[0064] Where Y represents the quantitative value of oil fume diffusion, Y a V represents the oil fume diffusion data value determined by the difference image method, V represents the comprehensive VOC data value, and μ represents the VOC data fitting coefficient.

[0065] Those skilled in the art can set the size of the VOC data fitting coefficient according to the actual situation, and the present invention does not limit it.

[0066] In this embodiment of the invention, by performing a weighted fitting of oil fume diffusion data and VOC data to calculate a quantitative value for oil fume diffusion, the overall situation of oil fume diffusion can be reflected more accurately. The difference image method provides quantitative data on visible oil fume diffusion, while VOC data can supplement information on colorless oil fume components (such as alkanes, alcohols, and ketones). Combining the two through weighted fitting fully utilizes the advantages of both data sources. This helps to more comprehensively assess the oil fume diffusion situation and provides a more accurate basis for decision-making regarding subsequent cleaning and fire prevention measures.

[0067] The cleaning judgment module 102 is used to determine whether the oil fume diffusion quantification value is greater than the preset diffusion quantification value. If so, the intelligent cleaning module 103 is invoked to perform intelligent cleaning. Otherwise, the oil fume diffusion monitoring module 101 is invoked to continue monitoring.

[0068] Those skilled in the art can set the preset diffusion quantization value according to the actual situation, and the present invention does not limit it.

[0069] In this embodiment of the invention, the range hood can automatically determine whether cleaning needs to be started based on the actual diffusion of oil fumes, avoiding the arbitrariness of the frequency and intensity of manual cleaning, ensuring that the range hood automatically cleans itself when oil fumes accumulate to a certain level, and reducing the frequency and error of manual operation.

[0070] The intelligent cleaning module 103 uses oil fume diffusion data and VOC data as input to determine the optimal steam release temperature and optimal steam release rate through an improved artificial neural network. Based on the optimal steam release temperature and optimal steam release rate, a PID control system based on a radial basis function neural network dynamically adjusts the steam release rate and temperature of the range hood, cleaning the hood by releasing high-temperature steam.

[0071] Artificial Neural Networks (ANNs) are computational models that mimic the structure of biological neural networks, used for pattern recognition and machine learning tasks. ANNs process input data through multiple layers of neurons (hierarchical structure), progressively extracting features and making decisions. ANNs are particularly well-suited for solving nonlinear problems and can be used in fields such as image recognition, speech recognition, and predictive analytics. Through training, neural networks can autonomously learn patterns and optimize themselves, making them one of the foundations of deep learning.

[0072] Optionally, the improvement of the artificial neural network can be achieved by replacing the original hidden layer with a gated convolutional layer, a flattened layer, and two fully connected layers between the input and output layers.

[0073] In one possible implementation, the intelligent cleaning module 103 is specifically used for:

[0074] In the input layer, input data includes oil fume diffusion data and VOC data.

[0075] In the gated convolutional layer, extract two-dimensional gated convolutional features:

[0076]

[0077]

[0078]

[0079] Where Conv1 represents the result of the first convolution, W g Let I represent the first convolution kernel, I represent the input data, Conv2 represent the second convolution result, and W represent the second convolution result. f This represents the second convolution kernel, O represents the gated convolution feature, and Sigmoid represents the Sigmoid activation function. This indicates element-wise multiplication. This indicates two summations: the first summation is performed in the row direction, and the second summation is performed in the column direction.

[0080] It should be noted that gated convolutional layers, by combining convolution operations and gating mechanisms, can effectively extract important features from input data, especially for modeling the nonlinear features of oil fume diffusion and VOC data. By weighting the convolution operations (element-wise multiplication), the gating mechanism can enhance useful information and suppress noise, making feature extraction more accurate and improving the performance of subsequent models.

[0081] In the flattening layer, two-dimensional gated convolutional features are mapped to one-dimensional convolutional features.

[0082] It's worth noting that the flattening layer converts two-dimensional feature maps into one-dimensional vectors. This operation allows the network to better process the spatial features extracted by the convolutional layers, facilitating further learning and integration by the fully connected layers. By compressing high-dimensional features into one-dimensional input, the model can perform calculations more efficiently and provide more structured information for subsequent decisions.

[0083] In the fully connected layer, the one-dimensional convolutional features are summarized to obtain integrated features.

[0084] It should be noted that fully connected layers, by summarizing the flattened features, can further enhance the network's learning ability and provide more parameter space for different types of features. Integrating features helps to perform deep learning on the needs of fume cleaning from multiple dimensions, enabling the model to more accurately determine and optimize the temperature and amount of steam release.

[0085] Based on the integration characteristics, the optimal steam release temperature and the optimal steam release rate are determined using an activation function.

[0086] It should be noted that by using an activation function, the network can output the optimal steam release temperature and amount based on integrated features. The role of the activation function is to map the output of the neural network to the actual control parameter range, enabling the system to adjust the cleaning intensity and temperature in real time, thereby optimizing the oil fume cleaning effect and ensuring that the system maintains a high-efficiency working state under different oil fume accumulation conditions.

[0087] Radial Basis Function Neural Network (RBFNN)-based PID control systems offer an advanced control approach that combines traditional PID control with the powerful nonlinear mapping capabilities of RBFNNs. In this system, the RBFNN is used to accurately model and predict complex system dynamics, capturing the nonlinear relationship between input and output through neural network training. The PID controller then utilizes the characteristic inputs provided by the RBFNN, combined with proportional, integral, and derivative control algorithms, to adjust the control signal in real time.

[0088] In one possible implementation, the intelligent cleaning module 103 is specifically used for:

[0089] Calculate the error function between the current output and the target output, where the target output is either the optimal steam release temperature or the optimal steam release rate.

[0090]

[0091] Where e(k) represents the error at time k, r(k) represents the target output at time k, and y(k) represents the system output at time k.

[0092] It should be noted that by calculating the error function between the current output and the target output, the system can measure in real time the deviation between the current steam release temperature or quantity of the range hood and the target value. This process helps the system understand the gap between the current state and the ideal state, providing a basis for subsequent control signal adjustments and ensuring that each cleaning is closer to the expected results.

[0093] PID control is used to determine the control signal for the steam release or temperature of the range hood:

[0094]

[0095] in, K represents the increment of the control signal at time k. p (k) represents the proportional gain coefficient at time k, K i (k) represents the integral gain coefficient at time k, K d (k) represents the differential gain coefficient at time k, e(k-1) represents the error at time k-1, and e(k-2) represents the error at time k-2.

[0096] Among them, PID control (proportional-integral-derivative control) is a classic feedback control algorithm used to adjust the system output to approach the desired value.

[0097] Based on the squared value of the error, determine the PID performance index data:

[0098]

[0099] Where E(k) represents the PID performance index value at time k.

[0100] It should be noted that by calculating the square of the error, the PID performance index can quantify the performance of the control system. This index can help the system evaluate the effectiveness of the current control strategy and serve as a basis for adjusting control parameters, thereby improving the system's response speed and accuracy.

[0101] Based on a radial basis function neural network, the Jacobian matrix is ​​calculated to describe the rate of change of the system output with respect to the increment of the control signal.

[0102]

[0103] Where ∂ represents the partial derivative operator, w j (k) represents the weight of the j-th neuron in the radial basis function neural network at time k, h j (k) represents the output of the j-th neuron in the radial basis function neural network at time k, c j (k) represents the center position of the j-th neuron in the radial basis function neural network at time k, b j (k) represents the width of the j-th neuron in the radial basis function neural network at time k.

[0104] The Jacobian matrix describes the rate of change of the system output with respect to the increment of the control signal, reflecting the system's sensitivity to input changes. By calculating the Jacobian matrix, the system can accurately understand the impact of control signal changes on the range hood's output, ensuring more precise and efficient adjustment of PID control parameters.

[0105] Radial Basis Function Neural Networks (RBFNNs) are a type of feedforward neural network that uses radial basis functions (such as Gaussian functions) as activation functions. RBFNNs consist of an input layer, hidden layers, and an output layer. The hidden layers map the input using nonlinear functions, effectively handling complex nonlinear problems. RBFNNs are commonly used for function approximation, classification, and regression tasks. Their training process is relatively simple, and they exhibit good real-time performance and adaptability.

[0106] Based on the PID performance data, the PID control parameters are updated using the gradient descent method and the Jacobian matrix:

[0107]

[0108]

[0109]

[0110] Among them, K p (k+1) represents the proportional gain coefficient at time k+1, K i (k+1) represents the integral gain coefficient at time k+1, K d (k+1) represents the differential gain coefficient at time k+1, and η represents the learning rate.

[0111] It should be noted that by using the gradient descent method in conjunction with the Jacobian matrix, the system can optimize the PID gain parameters based on the error, thereby making the control process more precise. The gradient descent method can automatically adjust the PID parameters, avoiding deviations from manual settings and improving the system's automation level and accuracy.

[0112] Return to recalculate the error function between the current output and the target output, and use the updated PID control parameters to dynamically adjust the steam release or temperature of the range hood at the next moment.

[0113] During the cleaning process, the steam release and temperature of the range hood can be flexibly adjusted according to the actual diffusion of oil fumes to achieve precise cleaning and improve cleaning efficiency.

[0114] In one possible implementation, the intelligent cleaning module 103 is specifically used to: effectively decompose grease into smaller oil droplets by releasing high-temperature steam, allowing the oil droplets to mix with water, and the grease will be wrapped by the water and washed away, thereby cleaning the fume hood.

[0115] In one possible implementation, the intelligent cleaning module 103 is specifically used to: perform focused cleaning on the oil fume diffusion area indicated by the oil fume diffusion area image based on the oil fume diffusion area image.

[0116] It's important to note that by analyzing images of the fume diffusion area, areas with higher levels of fume accumulation can be accurately identified and targeted for cleaning, thus improving cleaning efficiency. This method avoids indiscriminate cleaning of the entire range hood, reducing unnecessary energy waste and cleaning time. Furthermore, precisely focusing on areas with severe fume diffusion more effectively removes stubborn grease and dirt, ensuring maximum cleaning results, extending equipment lifespan, and reducing maintenance costs.

[0117] The flame monitoring module 104 is used to acquire images of the stovetop via a camera. Based on the stovetop images, a convolutional neural network is used to identify the degree of flame combustion and determine the probability of a kitchen fire.

[0118] In one possible implementation, the flame monitoring module 104 is specifically used to: output a flame region based on the stove image using a Faster R-CNN model; determine the degree of flame combustion based on the area and brightness value of the flame region; and determine the probability value of a kitchen fire based on the degree of flame combustion.

[0119] Faster R-CNN is a deep learning model for object detection that combines a convolutional neural network (CNN) and a region proposal network (RPN). Faster R-CNN generates region proposals through the RPN, and then uses the CNN for feature extraction and classification. Compared to traditional object detection methods, Faster R-CNN significantly improves detection speed and accuracy through end-to-end training. This model is widely used for tasks such as object recognition, localization, and segmentation in images, and is suitable for fields such as autonomous driving and security monitoring.

[0120] The fire prevention judgment module 105 is used to determine whether the probability value of a kitchen fire is greater than a preset probability value. If so, the fire extinguishing module 106 is invoked to extinguish the fire. Otherwise, the flame monitoring module 104 is invoked to continue monitoring.

[0121] Those skilled in the art can set the preset probability value according to the actual situation, and the present invention does not limit it.

[0122] Fire extinguishing module 106 is used to spray water to extinguish fires in the kitchen.

[0123] In this embodiment of the invention, an intelligent fire prevention mechanism is integrated to ensure the safety of the kitchen in a high-temperature, high-fume environment and reduce the risk of fire caused by the accumulation of fumes or during cooking.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A smart fireproof and high-temperature steam cleaning integrated fume hood, characterized in that, include: The oil fume diffusion monitoring module is used to acquire oil fume images through a camera; obtain oil fume diffusion data based on the difference between the oil fume images and a reference image using the difference image method; arrange VOC sensors on the left, right, and top edges of the range hood, and use the VOC sensors to acquire VOC data; calculate the oil fume diffusion quantification value based on the oil fume diffusion data and the VOC data. The cleaning judgment module is used to determine whether the oil fume diffusion quantization value is greater than the preset diffusion quantization value; If so, the intelligent cleaning module is invoked to perform intelligent cleaning; otherwise, the oil fume diffusion monitoring module is invoked to continue monitoring. The intelligent cleaning module is used to determine the optimal steam release temperature and the optimal steam release amount by taking the oil fume diffusion data and the VOC data as input and through an improved artificial neural network. Based on the optimal steam release temperature and the optimal steam release amount, the PID control system based on the radial basis function neural network dynamically adjusts the steam release amount and temperature of the range hood, and cleans the range hood by releasing high-temperature steam. The flame monitoring module is used to acquire images of the stove through a camera; based on the images of the stove, a convolutional neural network is used to identify the degree of flame combustion and determine the probability value of a kitchen fire. The fire prevention judgment module is used to determine whether the probability value of a kitchen fire is greater than a preset probability value; if so, the fire extinguishing module is invoked to extinguish the fire; otherwise, the flame monitoring module is invoked to continue monitoring. The fire extinguishing module is used to spray water to extinguish fires in the kitchen.

2. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 1, characterized in that, The oil fume diffusion monitoring module is specifically used for: The difference image between the oil fume image and the reference image is calculated using the difference image method. The difference image is binarized to generate a mask image of the oil fume area; An erosion operation is performed on the discrete noise points in the mask image of the oil fume area to obtain an eroded mask image; The etched mask image is subjected to median filtering to obtain a filtered mask image; A dilation operation is performed on the filtered mask image to obtain a dilated mask image; Among them, the expansion operation is the dual operation of the corrosion operation; The difference image is multiplied point by point with the dilation mask image to determine the image of the oil fume diffusion area; The grayscale values ​​of each pixel in the image of the fume diffusion area are accumulated to obtain the fume diffusion data.

3. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 2, characterized in that, The oil fume diffusion monitoring module is specifically used for: VOC sensors are placed on the left, right, and top edges of the range hood; VOC data are acquired through each of the aforementioned VOC sensors; The VOC data measured by each of the VOC sensors are weighted and summed to calculate the comprehensive VOC data.

4. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 3, characterized in that, The oil fume diffusion monitoring module is specifically used for: The oil fume diffusion data and the VOC data are weighted and fitted to determine the quantitative value of oil fume diffusion.

5. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 1, characterized in that, The specific improvement method of the improved artificial neural network is as follows: between the input layer and the output layer, a gated convolutional layer, a flattened layer, and two fully connected layers are used to replace the original hidden layer.

6. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 5, characterized in that, The intelligent cleaning module is specifically used for: In the input layer, the oil fume diffusion data and the VOC data are input; In the gated convolutional layer, two-dimensional gated convolutional features are extracted; In the flattening layer, the two-dimensional gated convolutional features are mapped to one-dimensional convolutional features; In the fully connected layer, the one-dimensional convolutional features are summarized to obtain integrated features; Based on the integrated characteristics, the optimal steam release temperature and the optimal steam release rate are determined using an activation function.

7. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 5, characterized in that, The intelligent cleaning module is specifically used for: Calculate the error function between the current output and the target output, where the target output is specifically the optimal steam release temperature or the optimal steam release amount; PID control is used to determine the control signal for the steam release or temperature of the range hood; PID performance index data are determined based on the squared value of the error; Based on a radial basis function neural network, the Jacobian matrix is ​​calculated to describe the rate of change of the system output with respect to the increment of the control signal. Based on the PID performance index data, the PID control parameters are updated using the gradient descent method and the Jacobian matrix. Return to recalculate the error function between the current output and the target output, and use the updated PID control parameters to dynamically adjust the steam release or temperature of the range hood at the next moment.

8. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 2, characterized in that, The intelligent cleaning module is specifically used for: By releasing high-temperature steam, the grease is effectively broken down into smaller droplets, which then mix with water. The grease is then enveloped by the water and washed away, thus cleaning the range hood.

9. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 2, characterized in that, The intelligent cleaning module is specifically used for: Based on the image of the fume diffusion area, the fume diffusion area indicated by the image is cleaned in a focused manner.

10. The integrated intelligent fireproof and high-temperature steam cleaning fume hood according to claim 1, characterized in that, The flame monitoring module is specifically used for: Based on the image of the stove, the flame area is output using the Faster R-CNN model; the degree of flame combustion is determined based on the area and brightness value of the flame area; and the probability value of a kitchen fire is determined based on the degree of flame combustion.

Images