Monitoring camera with demisting function for coal mine field and demisting system

Abstract

The invention discloses a monitoring camera with a demisting function for a coal mine field and a demisting system, and relates to the technical field of monitoring camera equipment. According to the monitoring camera with the demisting function for the coal mine field and the demisting system, environmental data of the coal mine field are collected through the sensor, the data are analyzed through a principal component analysis algorithm, the demisting influence index is obtained, whether the environment meets the demisting condition or not is judged, and if yes, the demisting function is started; the method comprises the following steps: carrying out feature learning on an image through a deep convolutional neural network, identifying a haze region and evaluating a haze intensity index, and dynamically adjusting a defogging parameter through a particle swarm optimization algorithm according to the change of the haze intensity index. Finally, according to the demisting influence index, the temperature of the heating film is adjusted through a fuzzy control algorithm, and mist on the surface of the camera lens is effectively removed.

Description

Technical Field

[0001] The invention relates to the technical field of monitoring camera equipment, in particular to a monitoring camera for coal mines with a defogging function and a defogging system. Background Art

[0002] With the rapid development of the coal mining industry, the complexity and harshness of the coal mining operating environment have brought challenges to the normal operation of monitoring equipment, especially in the adverse environmental conditions such as low temperature, high humidity and haze in the mining area. The lens of the monitoring camera is easily blurred due to fogging, affecting the effect of real-time monitoring. With the increasing demand for coal mine safety management, it is particularly important to rely on efficient and clear monitoring images for safety monitoring, accident prevention and environmental monitoring. Therefore, how to ensure the clear image output of monitoring equipment in the coal mine operating environment has become one of the important topics of mine safety assurance technology.

[0003] At present, the solution to the haze problem of coal mine monitoring camera lenses mainly relies on traditional physical defogging technology and simple heating film heating adjustment. However, these methods have several disadvantages. First, the traditional heating film only prevents haze by increasing the temperature, but it cannot be flexibly adjusted according to environmental conditions for dynamic changes in harsh environments, and excessive or insufficient heating may affect the normal operation of the equipment. Secondly, most of the existing image processing methods rely on simple image defogging algorithms, which are difficult to effectively handle the special haze types in the coal mine environment, resulting in limited improvement in image clarity. At the same time, traditional methods often ignore the combined effects of multiple complex factors in the coal mine environment during the defogging process, resulting in unstable defogging effects and difficulty in achieving efficient and automated monitoring image clarity assurance. Therefore, the existing technology urgently needs further innovation, especially through comprehensive environmental perception, deep learning and intelligent control algorithms to achieve intelligent dynamic defogging technology. Summary of the invention

[0004] In view of the deficiencies of the prior art, the present invention provides a monitoring camera and a defogging system for coal mines with a defogging function. The coal mine environment data is collected by sensors, and the data is analyzed by a principal component analysis algorithm to obtain a defogging influence index; the defogging influence index is compared with a preset defogging threshold to determine whether the environment meets the defogging conditions, and if so, the defogging function is activated; a deep convolutional neural network is used to learn features of the image, identify the haze area and evaluate the haze intensity index; according to the change of the haze intensity index, the defogging parameters are dynamically adjusted by a particle swarm optimization algorithm; finally, according to the defogging influence index, the heating film temperature is adjusted by a fuzzy control algorithm to effectively remove the fog on the surface of the camera lens. The above-mentioned problems of the background technology are solved.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a monitoring camera and a defogging system for a coal mine with a defogging function, comprising the following modules: an environmental perception module, a judgment module, a defogging module, and a heating module; the environmental perception module is used to collect coal mine environmental data through sensors, and perform a multi-dimensional comprehensive analysis of the coal mine environmental data through a principal component analysis algorithm to obtain a defogging influence index; the judgment module is used to compare the defogging influence index with the defogging threshold to determine whether the current environment meets the defogging conditions, and if so, start the defogging module; the defogging module is used to perform feature learning on images under the coal mine environment through a deep convolutional neural network, identify haze areas in the image, evaluate the haze intensity index, and dynamically adjust the defogging parameters of the monitoring camera through a particle swarm optimization algorithm according to changes in the haze intensity index; the heating module is used to control the heating film to adjust the heating temperature of the camera lens surface through a fuzzy control algorithm according to the defogging influence index to remove fog on the surface of the monitoring camera.

[0006] Furthermore, the specific process of conducting a multi-dimensional comprehensive analysis of coal mine environmental data through the principal component analysis algorithm is as follows: obtaining coal mine environmental data, including temperature, humidity, air pressure and haze concentration, and performing normalization processing to eliminate the influence of different data units and magnitudes; performing dimensionality reduction processing on the normalized data through the principal component analysis algorithm, constructing the covariance matrix of the data, calculating the eigenvalues ​​and eigenvectors of each principal component, and selecting the first few principal components by eigenvalue sorting; projecting the normalized data onto the first few principal components to form a low-dimensional comprehensive environmental feature, in which each dimension represents the main characteristic information of the coal mine environmental status.

[0007] Furthermore, the specific process of obtaining the defog influence index is as follows: the low-dimensional comprehensive environmental characteristics obtained by principal component analysis are used as input, and weighted calculation is performed to comprehensively consider the influence of various environmental factors. By setting the weight coefficient, the principal component of each environmental characteristic is multiplied by the corresponding weight value to obtain a weighted feature vector; the weighted feature vector is input into a multi-layer perceptron neural network for nonlinear modeling. The network obtains the mathematical relationship between the environmental characteristics and the defog influence index through training to obtain the defog influence index; the defog influence index is the output result of the multi-layer perceptron network, which reflects the comprehensive influence of temperature, humidity, air pressure, and haze concentration factors in the coal mine environment on the haze intensity, and indicates the influence of the current environmental conditions on the defogging function of the camera lens.

[0008] Furthermore, the specific process of judging whether the current environment meets the defog conditions is as follows: according to the defog influence index, it is compared with the set defog threshold. When the defog influence index is greater than or equal to the defog threshold, the defog module is started. If the defog influence index is less than the defog threshold, the standby state is maintained and the mine environment data is continued to be monitored in real time.

[0009] Furthermore, the features of images in the coal mine environment are learned through deep convolutional neural networks, and the specific process of identifying the haze area in the image is as follows: the image data taken by the coal mine monitoring camera is obtained and preprocessed, including image denoising and contrast enhancement, to ensure that the image quality is suitable for network input; the preprocessed image data is input into the deep convolutional neural network, and the spatial features of the image are automatically extracted through multi-layer convolution operations, and the image features of different haze concentration areas are gradually identified; the probability value of each pixel belonging to the haze area is generated through the pooling layer and the fully connected layer; based on the probability value, the haze area in the image is determined and marked.

[0010] Furthermore, the specific process of evaluating the haze intensity index is as follows: according to the identified haze area, the area and pixel density of the haze area are calculated to evaluate the coverage of the haze area; according to the brightness and contrast of the haze area, combined with the area and density of the haze area, a weighted evaluation is performed to obtain a preliminary haze intensity index; the haze intensity index is associated with the coal mine environmental data, and the preliminary haze intensity index is corrected through a regression analysis model to obtain the haze intensity index, which quantifies the impact on the clarity of the monitoring image.

[0011] Furthermore, according to the change of the haze intensity index, the specific process of dynamically adjusting the defogging parameters of the surveillance camera through the particle swarm optimization algorithm is as follows: according to the haze intensity index, the particle swarm is initialized, the number of particle swarms is set, and the defogging parameters of each particle are defined, including contrast, brightness, and sharpness, and a random defogging parameter combination is assigned to each particle, and the parameter value is between the preset minimum and maximum values; a fitness function is defined to evaluate the effect of the defogging parameter combination corresponding to the particle; for each particle, the defogging parameters are updated according to the current speed and position, the defogging parameters of each particle are iteratively updated, the defogging effect is evaluated according to the fitness function, and the optimal parameter combination is selected; after reaching the preset maximum number of iterations or when the fitness value of the particle swarm converges, the particle swarm optimization algorithm is stopped and the defogging parameters are output.

[0012] Furthermore, the specific process of controlling the heating film to adjust the heating temperature of the camera lens surface through the fuzzy control algorithm is as follows: according to the defogger influence index, a fuzzy control rule is established, the defogger influence index is used as the input variable, and the heating temperature of the camera lens surface is set as the output variable; the defogger influence index is input into the fuzzy control rule library, the input variable is fuzzified, and a fuzzy value of the heating temperature is generated, the fuzzy reasoning result is defuzzified, and the fuzzy value is converted into an actual heating temperature value; according to the defuzzified heating temperature value, the heating intensity of the heating film is adjusted, the heating temperature of the camera lens surface is controlled, and the heating film temperature is dynamically adjusted.

[0013] The present invention has the following beneficial effects:

[0014] (1) A monitoring camera and defogging system for coal mines with defogging function. The environmental perception module collects coal mine environmental data through sensors, and uses the principal component analysis algorithm to conduct a multi-dimensional comprehensive analysis of the data to accurately obtain the defogging impact index, thereby providing a scientific basis for defogging control. The judgment module compares the defogging impact index with the preset defogging threshold, and intelligently determines whether to start the defogging function, effectively reducing human intervention and improving response speed and adaptability. The defogging module uses a deep convolutional neural network to learn the features of images in the coal mine environment, identify and evaluate the haze area and its intensity index, and dynamically adjusts the defogging parameters through the particle swarm optimization algorithm to ensure the clarity and stability of the monitoring image in a special haze environment.

[0015] (2) A monitoring camera and a defogging system for use in coal mines with a defogging function. The heating module uses a fuzzy control algorithm to adjust the temperature of the heating film according to the defogging influence index to remove the fog on the surface of the camera lens in real time. The heating control method intelligently adjusts the heating temperature according to the temperature and humidity changes in the coal mine environment, avoiding overheating or underheating, ensuring that the lens surface always remains clear, and improving the stability of the camera and its reliability during long-term operation. This enables coal mine monitoring cameras to continuously and effectively monitor in harsh environments and ensure safe management of mining areas.

[0016] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a monitoring camera with a defogging function for use in a coal mine and a defogging system flow chart. DETAILED DESCRIPTION

[0018] The embodiment of the present application solves the problem of blurred monitoring images in coal mines due to environmental humidity, haze and other factors through a monitoring camera and a defogging system with a defogging function. The existing monitoring camera system has limited image clarity in bad weather in mining areas, which affects real-time monitoring and safety management. The defogging technology of the present invention can effectively remove haze and moisture on the camera lens and imaging system, thereby improving image quality, ensuring the clarity and visibility of the monitoring screen, and ensuring the efficiency and accuracy of coal mine safety production monitoring.

[0019] The overall idea of ​​the solution in the embodiments of this application is as follows:

[0020] The coal mine environmental data is collected through sensors, and the principal component analysis algorithm is used to conduct a multi-dimensional comprehensive analysis of the coal mine environmental data to obtain the defogging impact index.

[0021] The defogging impact index is compared with the defogging threshold to determine whether the current environment meets the defogging conditions. If so, the defogging module is started.

[0022] The deep convolutional neural network is used to learn the features of images in the coal mine environment, identify the haze area in the image, and evaluate the haze intensity index. According to the changes of the haze intensity index, the particle swarm optimization algorithm is used to dynamically adjust the defogging parameters of the surveillance camera.

[0023] According to the defogging influence index, the heating film is controlled by the fuzzy control algorithm to adjust the heating temperature of the camera lens surface and remove the fog on the surveillance camera surface.

[0024] See also Figure 1 The embodiment of the present invention provides a technical solution: a monitoring camera and a defogging system for a coal mine with a defogging function, comprising the following modules: an environment perception module, a judgment module, a defogging module, and a heating module; the environment perception module is used to collect coal mine environment data through sensors, and perform multi-dimensional comprehensive analysis on the coal mine environment data through a principal component analysis algorithm to obtain a defogging influence index; the judgment module is used to compare the defogging influence index with the defogging threshold to determine whether the current environment meets the defogging conditions, and if so, start the defogging module; the defogging module is used to perform feature learning on images under the coal mine environment through a deep convolutional neural network, identify haze areas in the image, evaluate the haze intensity index, and dynamically adjust the defogging parameters of the monitoring camera through a particle swarm optimization algorithm according to changes in the haze intensity index; the heating module is used to control the heating film to adjust the heating temperature of the camera lens surface through a fuzzy control algorithm according to the defogging influence index to remove fog on the surface of the monitoring camera.

[0025] In this implementation scheme, the environment perception module: This module collects environmental data of the coal mine in real time through sensors, such as temperature, humidity, air pressure and haze concentration. These data are the main factors affecting the image quality of the surveillance camera in the coal mine environment. Working principle: The environmental data is input into the principal component analysis (PCA) algorithm. Through the PCA algorithm, the different data are first normalized to eliminate the unit differences and magnitude differences between different parameters. Then, the PCA algorithm performs dimensionality reduction processing on the data to extract the main components that best reflect the changes in the coal mine environment. These components will help the system evaluate the impact of haze on the camera image quality and generate a defogging impact index. Judgment module: This module is responsible for judging whether the coal mine environment meets the defogging conditions. It uses the defogging impact index obtained from the environment perception module to compare with the pre-set defogging threshold. Working principle: The judgment module determines whether the current environmental conditions are suitable for starting the defogging process by comparing the defogging impact index with the defogging threshold. If the defogging impact index exceeds the threshold, it means that the haze in the environment is more serious and the fog needs to be removed. At this time, the module will send a signal to the downstream defogging module to start the defogging function. Defogging module: The main function of this module is to process the haze area in the camera image and improve the clarity of the image. It uses deep convolutional neural network (CNN) to perform image analysis and feature learning to identify and process the haze area. Working principle: The defogging module uses CNN to learn the features of the coal mine environment images collected by the camera and identify the haze area. The system evaluates the intensity of the haze and calculates the haze intensity index. Subsequently, the defogging parameters of the camera (such as image enhancement coefficient, defogging filter setting, etc.) are dynamically adjusted through the particle swarm optimization algorithm (PSO) to achieve the best image clarity. The particle swarm optimization algorithm can automatically adjust the defogging parameters according to the changes in the haze intensity, thereby ensuring the continuous optimization of image quality. Heating module: This module removes fog from the surface of the camera lens through a heating film to further ensure image clarity. It automatically adjusts the heating temperature of the camera lens according to changes in the coal mine environment. Working principle: The heating module uses a fuzzy control algorithm to control the heating temperature of the heating film according to the defogging influence index. The fuzzy control algorithm is an adaptive control method that can adjust the temperature according to the current environmental changes, so that the heating film can effectively remove water vapor or frost on the lens surface without causing damage to the equipment. This process avoids the single control method of overheating or overcooling of the traditional heating film, thereby improving the flexibility and accuracy of the defogging effect.

[0026] Specifically, the specific process of multi-dimensional comprehensive analysis of coal mine environmental data through principal component analysis algorithm is as follows: obtain coal mine environmental data, including temperature, humidity, air pressure and haze concentration, and perform normalization processing to eliminate the influence of different data units and magnitudes; perform dimensionality reduction processing on the normalized data through principal component analysis algorithm, construct the covariance matrix of the data, calculate the eigenvalues ​​and eigenvectors of each principal component, and select the first principal components by eigenvalue sorting; project the normalized data onto the first principal components to form low-dimensional comprehensive environmental characteristics, in which each dimension represents the main characteristic information of the environmental status of the coal mine.

[0027] In this embodiment, the environmental data of the coal mine is obtained: first, the environmental data of the coal mine is collected by sensors, which usually include indicators such as temperature, humidity, air pressure and haze concentration. Since these data come from different environmental parameters, their units and magnitudes may be different, for example, the unit of temperature is Celsius, and the humidity is usually a percentage. In order to ensure that different types of data can be compared and processed in the same model, normalization is required. The purpose of normalization is to convert all data to the same dimension range (for example, between 0 and 1), thereby eliminating the unit difference and magnitude effect between different data, and ensuring that the data is analyzed on the same scale. Dimensionality reduction and construction of covariance matrix: The principal component analysis (PCA) algorithm is applied to the normalized data for dimensionality reduction. The purpose of dimensionality reduction is to compress high-dimensional environmental data into lower-dimensional features, so as to extract the most representative environmental feature information. First, a covariance matrix is ​​constructed, which reflects the correlation between different data dimensions. The covariance matrix is ​​used to measure the correlation between various variables (such as temperature, humidity, air pressure, etc.). The larger the covariance value, the stronger the correlation between these variables. Calculate eigenvalues ​​and eigenvectors: Next, the main components of the environmental data are obtained by calculating the eigenvalues ​​and eigenvectors of the covariance matrix. The eigenvalue reflects the importance of each principal component, while the eigenvector represents the direction or characteristic pattern of these principal components. The larger the eigenvalue, the greater the proportion of the principal component in the data variation. Select the first few principal components: Sort by the size of the eigenvalue and select the first few principal components with the largest eigenvalue. These principal components represent the main information in the original data and can effectively capture the variability of environmental data. For example, the first two or three principal components may be selected to ensure the integrity of the information while reducing the dimension of the data and simplifying subsequent processing. Data projection and low-dimensional feature extraction: Finally, the normalized data is projected onto the first few selected principal components. In this way, the original high-dimensional data is mapped to a low-dimensional space, and each dimension represents an important feature of the environmental status of the coal mine. This low-dimensional comprehensive environmental feature contains the main information of the coal mine environment, such as the combined influence of factors such as temperature change, humidity change, and haze concentration.

[0028] Specifically, the specific process of obtaining the defog influence index is as follows: the low-dimensional comprehensive environmental characteristics obtained by principal component analysis are used as input, and weighted calculation is performed to comprehensively consider the influence of various environmental factors. By setting the weight coefficient, the principal component of each environmental characteristic is multiplied by the corresponding weight value to obtain a weighted feature vector; the weighted feature vector is input into a multi-layer perceptron neural network for nonlinear modeling. The network obtains the mathematical relationship between the environmental characteristics and the defog influence index through training to obtain the defog influence index; the defog influence index is the output result of the multi-layer perceptron network, which reflects the comprehensive influence of temperature, humidity, air pressure, and haze concentration factors in the coal mine environment on the haze intensity, and indicates the influence of the current environmental conditions on the defogging function of the camera lens.

[0029] In this embodiment, the specific process of the calculation formula of the defogging influence index is as follows: The low-dimensional comprehensive environment feature vector obtained by principal component analysis (PCA) is X = [X 1 ,X 2 ,..,X j .,X n ], where X j represents the environmental characteristics represented by the jth principal component (such as temperature, humidity, air pressure, haze concentration). Through weighted calculation, the principal components of each environmental characteristic and the weight coefficient w j Perform point multiplication to obtain the weighted feature vector W = [w 1 ,w 2 ,..w j .,w n ], where w j is the jth feature X j The corresponding weight. The weighted feature vector calculation formula is: Where: X j is the environmental characteristic value of the jth principal component; w j is the weight coefficient of the jth principal component. Next, the weighted eigenvector X w The input is sent to a multi-layer perceptron (MLP) neural network for nonlinear modeling. The multi-layer perceptron network learns the mapping relationship between environmental features and the defogging influence index through training, and the output defogging influence index D is what we are looking for. Defogging influence index calculation formula: D = f(X w ,θ) where: D represents the defogging influence index, which reflects the influence of the coal mine environment on the defogging effect; f(·) is the nonlinear mapping function of the multi-layer perceptron neural network; X w is the weighted environmental feature vector; θ is the trainable parameter in the neural network.

[0030] Specifically, the specific process of judging whether the current environment meets the defog conditions is as follows: according to the defog influence index, it is compared with the set defog threshold. When the defog influence index is greater than or equal to the defog threshold, the defog module is started. If the defog influence index is less than the defog threshold, the standby state is maintained and the mine environment data continues to be monitored in real time.

[0031] In this implementation, the process of determining whether the current environment meets the defogging conditions is as follows: First, the calculated defogging influence index is compared with the preset defogging threshold. When the defogging influence index is greater than or equal to the defogging threshold, it means that the haze intensity in the current environment has reached the level that requires defogging, and the system will start the defogging module to perform image defogging; if the defogging influence index is less than the defogging threshold, it means that the haze in the environment is relatively light, and the system remains in standby mode, continues to monitor the environmental data of the coal mine in real time, and starts the defogging function when the conditions are met.

[0032] Specifically, the feature learning of images in the coal mine environment is carried out through deep convolutional neural networks, and the specific process of identifying the haze area in the image is as follows: the image data taken by the coal mine monitoring camera is obtained and preprocessed, including image denoising and contrast enhancement, to ensure that the image quality is suitable for network input; the preprocessed image data is input into the deep convolutional neural network, and the spatial features of the image are automatically extracted through multi-layer convolution operations, and the image features of areas with different haze concentrations are gradually identified; through the pooling layer and the fully connected layer, the probability value of each pixel belonging to the haze area is generated; based on the probability value, the haze area in the image is determined and marked.

[0033] In this embodiment, image preprocessing: image denoising and contrast enhancement processing are performed. The denoising process can use a filtering algorithm, such as median filtering or Gaussian filtering, to remove noise in the image. Contrast enhancement can use histogram equalization technology. Suppose the preprocessed image is I pre The image is input into a deep convolutional neural network (CNN) and the preprocessed image I pre Input into the deep convolutional neural network, the network consists of multiple convolutional layers, pooling layers and fully connected layers. Set the input image to I input , extract the spatial features of the image through convolution operation. The convolution layer performs convolution operation on the image through the filter (or convolution kernel) K. Each convolution operation generates a feature map (FeatureMap) F. Assume that F i is the output feature map of the i-th layer. The formula for the convolution operation is as follows: i =I input *K i +b i ; where * represents the convolution operation, K i is the convolution kernel of the i-th layer, b iis the bias term. Pooling operation: The pooling layer is used to reduce the size of the image, reduce the amount of calculation, and retain important spatial information. Let the pooling operation be P, and the pooling formula is: P i =max(F i [m,M]); where P i represents the feature map after pooling, F i [m,M] is the value in the region (m,M) in the i-th layer convolution feature map, and the maximum value is taken. Fully connected layer: The fully connected layer takes the feature map P output by the pooling layer i It is converted into a vector and multiplied with the weight matrix W to generate the output of image classification. The formula of the fully connected layer is as follows: O = W·P i +b; where W is the weight matrix of the fully connected layer, b is the bias term, and P i is the output of the pooling layer, and O is the output. The output of the fully connected layer is processed nonlinearly through the activation function (Sigmoid). Let the activation function be σ, then: O activation =σ(O); Through the Sigmoid activation function, the output of each pixel is mapped to the probability value of belonging to the haze area. For each pixel point (x, y), the probability P of belonging to the haze area mist It can be calculated by the following formula: Among them, e is a natural constant, O activation (x,y) is the output value after the activation function. The haze area is determined based on the probability value P mist (x,y), set a threshold T threshold To determine whether the pixel belongs to the haze area. mist (x,y)≥T threshold , then the pixel is considered to be a haze area, otherwise it is considered not to be a haze area. Finally, the haze area R mist It can be determined by the following formula: R mist ={(x,y)|P mist (x,y)≥T threshold}.

[0034] Specifically, the specific process of evaluating the haze intensity index is as follows: according to the identified haze area, the area and pixel density of the haze area are calculated to evaluate the coverage of the haze area; according to the brightness and contrast of the haze area, combined with the area and density of the haze area, a weighted evaluation is performed to obtain a preliminary haze intensity index; the haze intensity index is associated with the coal mine environmental data, and the preliminary haze intensity index is corrected through a regression analysis model to obtain the haze intensity index, which quantifies the impact on the clarity of the monitoring image.

[0035] In this embodiment, the area and pixel density of the haze region are calculated based on the haze region R identified above. mist , calculate the area A of the haze area mist and pixel density D mist . Area calculation formula: A mist =∑ (x,y)∈Rmist 1; where (x, y) is each pixel in the image, R mist It is a haze area. Area A mist is the number of all pixels belonging to the haze area. Pixel density calculation formula Among them, A total is the total area of ​​the image (i.e. the number of pixels in the image). Pixel density D mist It reflects the relative density of the haze area in the image. According to the weighted evaluation of brightness and contrast, the brightness and contrast of the haze area are weighted. Set the brightness and contrast to L mist and C mist ,They can be calculated by the average brightness and contrast of all pixels in the haze area. Brightness calculation formula: Where I(x,y) is the brightness value of pixel (x,y) in the image. Contrast calculation formula: in, Is the average brightness of the haze area. By using brightness and contrast, calculate the comprehensive evaluation value of the haze area: E mist =αL mist +βC mist ; Among them, α and β are weighting coefficients, reflecting the influence of brightness and contrast on haze intensity. According to the area, pixel density and comprehensive evaluation value E of the haze area mist , the preliminary haze intensity index I can be calculated mist,init , I mist,init =γA mist +δD mist +εE mist ; Among them, γ, δ and ε are weighted coefficients corresponding to area, pixel density and comprehensive evaluation value. The preliminary haze intensity index I is corrected by regression analysis model. mist,init The low-dimensional comprehensive environmental characteristic vector X of the coal mine is associated with it, and the preliminary haze intensity index is corrected through the regression analysis model. The regression model is: I mist =f(I mist,init ,X) where X is the vector of coal mine environmental data, and the regression function f is based on the environmental data and the preliminary haze intensity index I mist,init Generate the modified haze intensity index I mist . Haze Intensity Index I mistThe impact of haze on the clarity of surveillance images can be quantified. mist It means that the haze is more serious and has a greater impact on the image; while the lower I mist It means the haze is lighter and the image clarity is better.

[0036] Specifically, according to the change of the haze intensity index, the specific process of dynamically adjusting the defogging parameters of the surveillance camera through the particle swarm optimization algorithm is as follows: according to the haze intensity index, initialize the particle swarm, set the number of particle swarms, and define the defogging parameters of each particle including contrast, brightness, and sharpness, and assign a random defogging parameter combination to each particle, and the parameter value is between the preset minimum and maximum values; define a fitness function to evaluate the effect of the defogging parameter combination corresponding to the particle; for each particle, update the defogging parameters according to the current speed and position, iteratively update the defogging parameters of each particle, evaluate the defogging effect according to the fitness function, and select the optimal parameter combination; after reaching the preset maximum number of iterations or when the fitness value of the particle swarm converges, stop the particle swarm optimization algorithm and output the defogging parameters.

[0037] In this implementation, the particle swarm is initialized according to the haze intensity index I mist , initialize the particle swarm P = {p 1 ,p 2 ,…,p N}, where N is the number of particles in the swarm. Each particle p a The dehazing parameters include contrast C dehaze , Brightness L dehaze and sharpness S dehaze These parameters are set at the preset minimum values ​​C min ,L min ,S min and the maximum value C max ,L max ,S max Randomly generated between. Formula: Where a represents the particle index. The fitness function F(p i ) is used to evaluate the effect of the defogging parameter combination corresponding to the particle. This fitness function can be evaluated based on the defogging effect of the image, such as the image quality evaluation after defogging (for example, by calculating contrast, clarity, distortion, etc.). The fitness function F(p a ) is in the form of: F(p a )=α·Q dehazed +β·PSNR(I dehazed ,I original ), where Q dehazed is the quality evaluation value of the dehazed image, PSNR(I dehazed ,I original ) is the dehazed image Idehazed With the original image I original α and β are weighting coefficients that control the impact of quality assessment and signal-to-noise ratio on fitness. Each particle updates the defogging parameters according to its current speed and position. The update formula is: in: For particle p a The velocity at iteration t. For particle p a Position (dehaze parameter combination) at iteration t. For particle p a The best historical position. best is the historical optimal position of all particles. w is the inertia weight, which controls the search range of the particle swarm. 1 ,c 2 is the learning factor, which indicates the degree of dependence of the particle on individual experience and group experience. 1 ,r 2 is a random number in the range [0,1]. After each iteration, the fitness value F(p a ), and update the optimal solution according to the fitness function: If particle p a The fitness value F(p a ) is better than the current particle's historical optimal value Update If the fitness value of all particles F(p a ) is better than the current global optimal value F(g best ), then update g best The particle swarm optimization algorithm stops when any of the following termination conditions is reached: when the preset maximum number of iterations T is reached max The fitness value of the particle swarm F(g best ) converges, that is, the change in fitness is less than the preset threshold. After the algorithm terminates, the global optimal particle defogging parameter combination is output That is the final defogging parameters of the surveillance camera, including: These defogging parameters can be directly used to dynamically adjust the image defogging effect of surveillance cameras.

[0038] Specifically, the specific process of controlling the heating film to adjust the heating temperature of the camera lens surface through the fuzzy control algorithm is as follows: according to the defogger influence index, a fuzzy control rule is established, the defogger influence index is used as the input variable, and the heating temperature of the camera lens surface is set as the output variable; the defogger influence index is input into the fuzzy control rule library, the input variable is fuzzified, and a fuzzy value of the heating temperature is generated; the fuzzy reasoning result is defuzzified, and the fuzzy value is converted into an actual heating temperature value; according to the defuzzified heating temperature value, the heating intensity of the heating film is adjusted, the heating temperature of the camera lens surface is controlled, and the heating film temperature is dynamically adjusted.

[0039] In this embodiment, a fuzzy control rule base is constructed according to the degree of environmental factors and the influence of haze. The rule base is mainly set as the input variable by setting the defogging influence index and the heating temperature of the camera lens surface as the output variable. The defogging influence index reflects the concentration and intensity of haze in the environment. This value can dynamically reflect the changes in haze and guide the adjustment of heating temperature. The defogging influence index is input into the fuzzy control system. First, the defogging influence index is fuzzified, that is, this continuous numerical interval is converted into a fuzzy language variable (such as: low, medium, high). This step is to convert the digital data in reality into a fuzzy set that can be processed by the system. In the fuzzy control rule base, different heating temperature requirements (such as low, medium, and high heating intensity) correspond to different fuzzy inputs (such as low, medium, and high defogging influence indexes). The fuzzy control system calculates the fuzzy value of the heating temperature based on the fuzzy value of the input variable and according to the preset rules. For example, if the defogging influence index is high, the system may infer that a higher heating temperature is required. The result of fuzzy reasoning is still a fuzzy temperature value, so it needs to be converted into a specific actual heating temperature through defuzzification. This process converts the output of fuzzy reasoning (such as medium and high temperature range) into a specific value so that the system can accurately control the heating intensity of the heating film. According to the defuzzified heating temperature value, the system adjusts the heating intensity of the heating film to control the heating temperature of the camera lens surface. If the haze is severe (the defog influence index is high), the system will increase the heating temperature to quickly remove the fog on the lens surface; conversely, when the haze is light, the system will appropriately reduce the heating temperature. Through the continuous operation of the fuzzy control algorithm, the system can monitor the changes in the environment in real time, and dynamically adjust the heating temperature of the heating film according to the changes in the haze intensity and the defog influence index to ensure that the camera lens is always in the best clear state.

[0040] In summary, this application has at least the following effects:

[0041] A monitoring camera and defogging system for coal mines with defogging function collects coal mine environmental data through the environmental perception module, and uses the principal component analysis algorithm to conduct a comprehensive analysis of multi-dimensional data, which can accurately obtain the defogging influence index and provide a scientific basis for subsequent defogging operations. The judgment module can intelligently judge whether the current environment needs defogging by comparing with the defogging threshold, ensuring that the defogging operation is only started when necessary, improving system efficiency and saving energy. The defogging module uses a deep convolutional neural network to learn image features, accurately identify the haze area in the coal mine environment, and can dynamically evaluate the image clarity and make real-time adjustments in combination with the haze intensity index. The defogging parameters of the monitoring camera are dynamically adjusted through the particle swarm optimization algorithm, which can automatically optimize the defogging effect under different haze intensities and improve the quality of the monitoring image. The heating module combines the defogging influence index and accurately adjusts the heating temperature of the heating film through the fuzzy control algorithm, effectively removing the fog on the surface of the camera lens, avoiding overheating, saving energy and ensuring the defogging effect.

[0042] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0043] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0044] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0045] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0046] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0047] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A monitoring camera and defogging system for coal mines with defogging function, characterized in that: It includes the following modules: environment perception module, judgment module, defogging module, and heating module; The environmental perception module is used to collect coal mine environmental data through sensors, perform multi-dimensional comprehensive analysis on the coal mine environmental data through a principal component analysis algorithm, and obtain a defogging impact index; The judgment module is used to compare the defogging influence index with the defogging threshold to determine whether the current environment meets the defogging conditions, and if so, start the defogging module; The defogging module is used to perform feature learning on images under coal mine environment through deep convolutional neural network, identify haze areas in images, evaluate haze intensity index, and dynamically adjust the defogging parameters of surveillance cameras through particle swarm optimization algorithm according to the change of haze intensity index; The heating module is used to control the heating film to adjust the heating temperature of the camera lens surface according to the defogging influence index through a fuzzy control algorithm to remove fog on the surveillance camera surface.

2. A monitoring camera and defogging system for coal mines with defogging function according to claim 1, characterized in that: The specific process of multi-dimensional comprehensive analysis of coal mine environmental data through principal component analysis algorithm is as follows: Obtain coal mine environmental data, including temperature, humidity, air pressure and haze concentration, and perform normalization to eliminate the impact of different data units and magnitudes; The normalized data is processed for dimensionality reduction by principal component analysis algorithm, the covariance matrix of the data is constructed, the eigenvalues ​​and eigenvectors of each principal component are calculated, and the first few principal components are selected by eigenvalue sorting; The normalized data are projected onto the first few principal components to form low-dimensional comprehensive environmental features, in which each dimension represents the main characteristic information of the environmental status of the coal mine.

3. The monitoring camera and defogging system for coal mines with defogging function according to claim 2, characterized in that: The specific process of obtaining the defog impact index is as follows: The low-dimensional comprehensive environmental characteristics obtained through principal component analysis are used as input for weighted calculation. The influence of various environmental factors is comprehensively considered. By setting the weight coefficient, the principal component of each environmental characteristic is multiplied by the corresponding weight value to obtain the weighted feature vector. The weighted feature vector is input into a multi-layer perceptron neural network for nonlinear modeling. The network obtains the mathematical relationship between the environmental characteristics and the defogging influence index through training, and obtains the defogging influence index. The defog impact index is the output result of a multi-layer perceptron network. It reflects the comprehensive impact of temperature, humidity, air pressure, and haze concentration factors on the haze intensity in the coal mine environment, and indicates the degree of influence of the current environmental conditions on the defog function of the camera lens.

4. The monitoring camera and defogging system for coal mines with defogging function according to claim 3, characterized in that: The specific process of judging whether the current environment meets the defogging conditions is as follows: According to the defog influence index, it is compared with the set defog threshold. When the defog influence index is greater than or equal to the defog threshold, the defog module is started. If the defog influence index is less than the defog threshold, the module remains in standby mode and continues to monitor the mine environment data in real time.

5. The monitoring camera and defogging system for coal mines with defogging function according to claim 4, characterized in that: The specific process of learning the features of images in coal mine environments through deep convolutional neural networks and identifying haze areas in images is as follows: Obtain image data captured by coal mine monitoring cameras and perform preprocessing, including image denoising and contrast enhancement, to ensure that the image quality is suitable for network input; The preprocessed image data is input into a deep convolutional neural network, which automatically extracts the spatial features of the image through multi-layer convolution operations and gradually identifies the image features of areas with different haze concentrations. Through the pooling layer and the fully connected layer, the probability value of each pixel belonging to the haze area is generated; Based on the probability value, the haze area in the image is determined and marked.

6. The monitoring camera and defogging system for coal mines with defogging function according to claim 5, characterized in that: The specific process of evaluating the haze intensity index is as follows: According to the identified haze area, the area and pixel density of the haze area are calculated to evaluate the coverage of the haze area; Based on the brightness and contrast of the haze area, combined with the area and density of the haze area, a weighted assessment is performed to obtain a preliminary haze intensity index; The haze intensity index was correlated with the coal mine environmental data, and the preliminary haze intensity index was corrected through the regression analysis model to obtain the haze intensity index, which quantified the impact on the clarity of the monitoring image.

7. A monitoring camera and defogging system for coal mines with defogging function according to claim 6, characterized in that: According to the changes in the haze intensity index, the specific process of dynamically adjusting the defogging parameters of the surveillance camera through the particle swarm optimization algorithm is as follows: According to the haze intensity index, the particle swarm is initialized, the number of particle swarms is set, and the defogging parameters of each particle are defined, including contrast, brightness, and sharpness. A random defogging parameter combination is assigned to each particle, and the parameter value is between the preset minimum and maximum values. Define a fitness function to evaluate the effect of the defogging parameter combination corresponding to the particles; For each particle, update the defogging parameters according to the current speed and position, iteratively update the defogging parameters of each particle, evaluate the defogging effect according to the fitness function, and select the optimal parameter combination; After reaching the preset maximum number of iterations or when the fitness value of the particle swarm converges, the particle swarm optimization algorithm is stopped and the dehazing parameters are output.

8. The monitoring camera and defogging system for coal mines with defogging function according to claim 7, characterized in that: The specific process of controlling the heating film to adjust the heating temperature of the camera lens surface through the fuzzy control algorithm is as follows: According to the defogging influence index, a fuzzy control rule is established, the defogging influence index is used as the input variable, and the heating temperature of the camera lens surface is set as the output variable; Input the defogging influence index into the fuzzy control rule base, perform fuzzification processing on the input variables, generate the fuzzy value of the heating temperature, perform defuzzification processing on the fuzzy reasoning result, and convert the fuzzy value into the actual heating temperature value; According to the defuzzified heating temperature value, the heating intensity of the heating film is adjusted, the heating temperature of the camera lens surface is controlled, and the heating film temperature is dynamically adjusted.

Images