Industrial building indoor oil mist image processing method, system, equipment and medium
Through the mixed Gaussian background model and wavelet analysis method, the dynamic changes of oil mist particles are monitored in real time, which solves the problem that traditional exhaust systems cannot accurately regulate, and achieves accurate identification of oil mist particles and improves environmental quality.
Patent Information
- Application Number
- CN202411093375.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-09
- Publication Date
- 2024-11-22
AI Technical Summary
Traditional exhaust systems are difficult to achieve accurate regulation when facing dynamic changes in oil mist concentration, resulting in waste of energy and poor environmental quality. The existing technology cannot monitor the emanation status of oil mist particles in real time, affecting the operating efficiency of local exhaust systems.
The hybrid Gaussian background model is used to combine morphological methods and wavelet analysis to monitor the dynamic changes of oil mist particles in real time, and the dynamic change areas and static areas are separated by the pixel background of the video frame, energy characteristics are extracted and thresholds are set to determine the oil mist area. Provide data to support local exhaust system regulation.
Accurate monitoring and identification of oil mist particles is achieved, the regulation accuracy of local exhaust systems is improved, energy consumption is reduced, and the workshop environmental quality is improved.
Smart Images

Figure CN119007108A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of oil mist image recognition, and in particular, relates to a method, system, equipment and medium for processing indoor oil mist images of industrial buildings. Background Art
[0002] In many industrial processing fields, the generation of oil mist particles is common, especially in the processing of non-ferrous metals such as copper and aluminum rolling. The emission of these particles is affected by a variety of processing parameters, such as the rolling mill's reduction, strip width, rolling force, and oil injection volume, resulting in significant fluctuations in the emission intensity. Traditional exhaust systems often use a fixed air volume design to cope with the most unfavorable working conditions and ensure effective pollution control of pollutants. However, this design method cannot adapt to the significant fluctuations in source intensity in actual working conditions, resulting in energy waste. Therefore, regulating the local exhaust system to adapt to the dynamic changes in oil mist concentration during the production process is of great significance to improving the workshop environment quality level and energy conservation and carbon reduction.
[0003] The operation control system dynamically adjusts the ventilation system by real-time monitoring of process parameters to ensure that it is in the best operating state. However, in the environment of small-particle, high-viscosity oil mist, the traditional sensor-based operation control system faces two major challenges: due to the mutual interference caused by the high viscosity and diverse sources of oil mist, it is difficult for feedforward control to accurately predict the generation and distribution of oil mist; secondly, the feedback control of oil mist is limited by the long-term stability of detection data, and it is difficult to provide continuous and accurate control feedback.
[0004] With the development of computer vision technology, cameras combined with advanced image processing algorithms are gradually becoming a new monitoring method. Compared with traditional sensors, camera systems have better adaptability and flexibility. By capturing real-time images in the workshop with high-resolution cameras and analyzing these images with image recognition algorithms, it is not only possible to realize the functions of traditional sensors, such as monitoring temperature distribution and identifying specific particle concentrations, but also to perform more complex analysis, such as evaluating the operating status of mechanical equipment and predicting potential failures. In addition, the durability and maintenance cost of camera systems in extreme industrial environments are relatively low, so the use of image recognition to achieve the perception of environmental parameters has great application prospects. Summary of the invention
[0005] In order to solve the problems existing in the prior art, the present invention provides an indoor oil mist image processing method, system, equipment and medium for industrial buildings, which are used to monitor the dispersion state of oil mist particles in the target area in real time, and provide a data basis for the control of the local exhaust system, so as to solve the problems that the prior art cannot monitor in real time and the unreasonable operation of the exhaust system.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] A method for processing indoor oil mist images of industrial buildings comprises the following steps:
[0008] Collect visual information of oil mist particles in the target area and determine video frames in the visual information;
[0009] Describing the background of each pixel in the video frame based on a mixed Gaussian background model, and separating a first dynamically changing area and a first static area based on the background of each pixel;
[0010] Using morphological method to enhance the oil mist image features in the first dynamic change area;
[0011] A morphological method is used to enhance the image features of the first dynamic area, and a wavelet analysis method is used to extract the energy features of the enhanced first dynamic area. A threshold is set based on the energy features, and the second dynamic area is determined by threshold judgment, and the second dynamic area is used as the oil mist area.
[0012] Furthermore, the process of describing the background of each pixel in the video frame is:
[0013] The observed value of the pixel point is taken as a random variable, and the probability distribution of the random variable is composed of multiple Gaussian distributions, which represent the typical value of the pixel in different time periods;
[0014] The mixed Gaussian background model assumes that each pixel observation value is I, and the probability distribution of It for a single sample can be expressed as:
[0015]
[0016] Where K is the number of Gaussian distribution quantities. The higher the K value, the better the system describes the complex environment; ω kt is the weight of the kth Gaussian distribution at time, which is used to reflect the importance of the Gaussian distribution in the model. The sum of all weights is equal to 1; η is the probability density function of the Gaussian distribution, which is defined as:
[0017]
[0018] According to the size of the weight coefficient or the weight coefficient ω kt The ratio of the Gaussian components to the standard deviation σ is used to sort all Gaussian components, determine the importance and order of the Gaussian components in the model, and divide the Gaussian distribution into the first dynamic change area and the first static area.
[0019] Furthermore, when describing the background, several Gaussian distributions whose weighted cumulative values exceed a preset threshold value T for the first time are selected as the first static area set, and the remaining Gaussian distributions are classified as the first dynamically changing area set; the first static area set is used as the boundary between the first dynamically changing area and the first static area by accumulating and sorting the weights of the Gaussian distributions, if the accumulated value exceeds the set threshold value T;
[0020] The background formula is expressed as:
[0021]
[0022] Where B is the number of Gaussian distributions used for the background.
[0023] Furthermore, the process of separating the first dynamically changing area and the first static area based on the background of each pixel point is:
[0024] The mixed Gaussian background model includes K Gaussian distributions, each of which has a mean vector μ kt and standard deviation σ;
[0025] The mixed Gaussian background model calculates the observation value I based on the arrangement order of the K Gaussian distributions t With the mean μ of each Gaussian distribution kt The distance between the observation values I t Match it with one of the Gaussian distributions;
[0026] The matching principle is: when the distance of the Kth Gaussian distribution is less than or equal to λ times the standard deviation λσ of the distribution for the first time, the Gaussian distribution and the observation value I t Matching; the observed value I t is any pixel in the video frame;
[0027] The matching function M of the pixel point kt If the value can be assigned to 1, it means the match is successful; otherwise, the matching function remains at 0, indicating a match failure;
[0028] If the match is successful, the pixel point is used as the first static area, otherwise the pixel point is used as the first dynamically changing area.
[0029] Furthermore, the mixed Gaussian background model is used to adapt to background changes by learning and updating the parameters of each Gaussian distribution;
[0030] The parameters of the Gaussian distribution include weight, mean and variance;
[0031] The rule for the mixed Gaussian background model to learn and update the parameters of each Gaussian distribution is:
[0032] At time t, if there is no Gaussian distribution and observation value I t The mixed Gaussian background model will introduce a new Gaussian distribution, and the mean of the new Gaussian distribution is set to the current observation value I t , and set a larger initial variance and a smaller weight, and remove the existing distribution with the lowest weight;
[0033] If the observed value I t Matching with a specific Gaussian distribution, the mixed Gaussian background model adjusts the weight, mean and variance according to the update mode of formula (5);
[0034]
[0035] In the formula, ω kt+1 is the weight coefficient of the kth Gaussian distribution after update at time t+1; α is the learning rate, ranging from 0<α<1; M kt Indicator function indicating successful matching, which is 1 if the matching is successful, otherwise 0; μ kt+1 is the updated mean of the kth Gaussian distribution at time t+1; ρ kt is the adaptive learning rate, indicating the observed value I t The degree of influence on the update of the mean of the Gaussian distribution; is the probability density function of the Gaussian distribution, and calculates the observed value I t The probability of the k-th Gaussian distribution; It is the updated variance of the kth Gaussian distribution at time t+1.
[0036] Furthermore, the morphological method includes dilation, erosion, opening operation and closing operation methods.
[0037] Furthermore, the process of extracting the energy characteristics of the enhanced first dynamic region by using the wavelet analysis method is as follows:
[0038] Decomposing the enhanced first dynamic region along one direction to generate two sub-bands: a low-frequency sub-band L and a high-frequency sub-band H;
[0039] The wavelet analysis method is based on two-dimensional wavelet transform and decomposes the image by applying low-pass and high-pass filters to generate four different sub-bands: a low-frequency information sub-image LL1, a horizontal high-frequency edge sub-image HL1, a vertical high-frequency edge sub-image LH1 and a diagonal high-frequency edge sub-image HH1;
[0040] The two-dimensional wavelet transform is:
[0041]
[0042] ψ 3 (x, y) = ψ(x)ψ(y) (9)
[0043] in, Represents a separable two-dimensional scaling function derived from a one-dimensional scaling function At the same time 1 (x,y),ψ 2 (x,y),v 3 (x, y) is a separable two-dimensional wavelet function with "directional perception" characteristics, which corresponds to the changes in the horizontal edge of the image in the column direction, the vertical edge in the row direction, and the edge grayscale in the diagonal direction. is a representation of a one-dimensional wavelet function;
[0044] The features related to the second dynamic change area are obtained by calculating and extracting the energy value of the high frequency sub-band H;
[0045] The energy value is calculated by summing the squares of all pixel values in the subband. The specific calculation formula is:
[0046] ENergy=|LH| 2 +|HL| 2 +|HH| 2
[0047] The process of determining the second dynamic area by threshold determination is:
[0048] Assume that Energy represents the energy value of the image to be detected at the position (x1, y1), and B(x1, y1) represents the energy value of the high-frequency component of the first static change area image at the corresponding position after two-dimensional discrete wavelet transform. If the energy value of the image to be detected is lower than the energy value of the first static area, the image block at this position is determined to be the second dynamic change area.
[0049] The condition for judging whether the second dynamic change area exists is:
[0050] T1×B(x1,y1) <E(x1,y1)<T2×B(x1,y1) (10)
[0051] If the energy value of the image block to be detected is between T1 and T2 times the energy value of the high-frequency component at the corresponding position of the background image, the image block is considered to be the second dynamic change area, and this area is marked as an oil mist area.
[0052] An industrial building indoor oil mist image processing system, comprising:
[0053] The acquisition unit is configured as:
[0054] Used to collect visual information of oil mist particles in the target area and determine the video frames in the visual information;
[0055] The model processing unit is configured as follows:
[0056] Used to describe the background of each pixel in the video frame based on a mixed Gaussian background model, and separate a first dynamically changing area and a first static area based on the background of each pixel;
[0057] The enhanced processing unit is configured to:
[0058] Used to enhance the oil mist image characteristics of the first dynamically changing area by using a morphological method;
[0059] Output unit, configured as:
[0060] The method is used to enhance the image features of the first dynamic area by using a morphological method, extract the energy features of the enhanced first dynamic area by using a wavelet analysis method, set a threshold based on the energy features, determine the second dynamic area by threshold judgment, and use the second dynamic area as the oil mist area.
[0061] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for processing indoor oil mist in industrial buildings are implemented.
[0062] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the industrial building indoor oil mist image processing method are implemented.
[0063] Compared with the prior art, the present invention has the following beneficial technical effects:
[0064] The present invention provides an industrial building indoor oil mist image processing method, system, equipment and medium, including: collecting visual information of oil mist particles in a target area, determining a video frame in the visual information; describing the background of each pixel in the video frame based on a mixed Gaussian background model, and separating a first dynamic change area and a first static area based on the background of each pixel point; using a morphological method to enhance the oil mist image features of the first dynamic change area; using a morphological method to enhance the image features of the first dynamic area, using a wavelet analysis method to extract the energy features of the enhanced first dynamic area, setting a threshold based on the energy features, determining a second dynamic area through threshold judgment, and using the second dynamic area as an oil mist area; the application monitors the emission state of oil mist particles in the target area in real time, can accurately and efficiently identify oil mist particles, and provides a data basis for local exhaust system control. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] Figure 1 This is a flow chart of a method for processing oil mist images in an industrial building interior according to an embodiment of the present invention;
[0066] Figure 2 The results of processing the target image using the algorithm and the conventional algorithm in the embodiment of the present invention are compared;
[0067] Figure 3 This is a schematic diagram of an oil mist particle recognition experiment in an embodiment of the present invention;
[0068] Figure 4 This is the oil mist particle identification result in the embodiment of the present invention;
[0069] Figure 5 Schematic diagram of two-dimensional discrete wavelet decomposition in an embodiment of the present invention.
[0070] In the figure, 1. local exhaust hood; 2. camera; 3. oil mist release device; 4. oil mist generator; 5. computer. DETAILED DESCRIPTION
[0071] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.
[0072] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.
[0073] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0074] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection, or a communication; it can be a direct connection, or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0075] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is lower in level than the second feature.
[0076] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.
[0077] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0078] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
[0079] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.
[0080] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0081] Figure 1 A method for processing oil mist images in an industrial building according to an embodiment of the present disclosure is shown. Figure 1 As shown, the following steps are included:
[0082] Step S1: collecting visual information of oil mist particles in the target area and determining a video frame in the visual information;
[0083] Specifically, a fixed camera is used to acquire a video sequence of a target area, and video frames in the video sequence are connected to each other through different types and dependencies to form a complete video content.
[0084] Step S2: describing the background of each pixel in the video frame based on a mixed Gaussian background model, and separating a first dynamically changing area and a first static area based on the background of each pixel;
[0085] Preferably, in the embodiment of the present disclosure, the process of describing the background of each pixel in the video frame is:
[0086] The observed value of the pixel point is taken as a random variable, and the probability distribution of the random variable is composed of multiple Gaussian distributions, which represent the typical value of the pixel in different time periods;
[0087] The mixed Gaussian background model assumes that each pixel observation value is I, and the probability distribution of It for a single sample can be expressed as:
[0088]
[0089] Where K is the number of Gaussian distribution quantities. The higher the K value, the better the system describes the complex environment; ω kt is the weight of the kth Gaussian distribution at time, which is used to reflect the importance of the Gaussian distribution in the model. The sum of all weights is equal to 1; η is the probability density function of the Gaussian distribution, which is defined as:
[0090]
[0091] According to the size of the weight coefficient or the weight coefficient ω kt The ratio of the Gaussian components to the standard deviation σ is used to sort all Gaussian components, determine the importance and order of the Gaussian components in the model, and divide the Gaussian distribution into the first dynamic change area and the first static area;
[0092] When describing the background, several Gaussian distributions whose weighted cumulative values exceed a preset threshold T for the first time are selected as the first static area set, and the remaining Gaussian distributions are classified as the first dynamically changing area set; the first static area set is used as the boundary between the first dynamically changing area and the first static area by accumulating the weights of the sorted Gaussian distributions, and if the accumulated value exceeds the set threshold T;
[0093] The background formula is expressed as:
[0094]
[0095] Where B is the number of Gaussian distributions used for the background.
[0096] Further preferably, in the embodiment of the present disclosure, the process of separating the first dynamically changing area and the first static area based on the background of each pixel point is:
[0097] The mixed Gaussian background model includes K Gaussian distributions, each of which has a mean vector μ kt and standard deviation σ;
[0098] The mixed Gaussian background model calculates the observation value I based on the arrangement order of the K Gaussian distributions t With the mean μ of each Gaussian distribution kt The distance between the observation values I t Match it with one of the Gaussian distributions;
[0099] The matching principle is: when the distance of the Kth Gaussian distribution is less than or equal to λ times the standard deviation λσ of the distribution for the first time, the Gaussian distribution and the observation value I t Matching; the observed value I t is any pixel in the video frame;
[0100] The matching function M of the pixel point it+1 If the value can be assigned to 1, it means the match is successful; otherwise, the matching function remains at 0, indicating a match failure;
[0101] If the match is successful, the pixel point is used as the first static area, otherwise the pixel point is used as the first dynamically changing area.
[0102] Furthermore, the mixed Gaussian background model is used to adapt to background changes by learning and updating the parameters of each Gaussian distribution;
[0103] The parameters of the Gaussian distribution include weight, mean and variance;
[0104] The rule for the mixed Gaussian background model to learn and update the parameters of each Gaussian distribution is:
[0105] At time t, if there is no Gaussian distribution and observation value I t The mixed Gaussian background model will introduce a new Gaussian distribution, and the mean of the new Gaussian distribution is set to the current observation value I t , and set a larger initial variance and a smaller weight, and remove the existing distribution with the lowest weight;
[0106] If the observed value I t Matching with a specific Gaussian distribution, the mixed Gaussian background model adjusts the weight, mean and variance according to the update mode of formula (5);
[0107]
[0108] In the formula, ω kt+1 is the weight coefficient of the kth Gaussian distribution after update at time t+1; α is the learning rate, ranging from 0<α<1; M kt Indicator function indicating successful matching, which is 1 if the matching is successful, otherwise 0; μ kt+1 is the updated mean of the kth Gaussian distribution at time t+1; ρ kt is the adaptive learning rate, indicating the observed value I t The degree of influence on the update of the mean of the Gaussian distribution; is the probability density function of the Gaussian distribution, and calculates the observed value I t The probability of the k-th Gaussian distribution; It is the updated variance of the kth Gaussian distribution at time t+1.
[0109] Step S3: using a morphological method to enhance the oil mist image features of the first dynamically changing area; preferably, the morphological method includes expansion, corrosion, opening operation and closing operation methods; specifically, in morphological transformation, the expansion operation is usually used to highlight the details and features in the image, and by expanding the edge of the object outward, the expansion operation can enhance certain areas in the image, making these areas more prominent visually; in morphological transformation, the corrosion operation is usually used to eliminate noise and small details in the image, and by shrinking the edge of the object inward, the corrosion operation can remove small areas in the image, such as small noise points or connected domains; the opening operation is also called an opening operation, and the opening operation can remove the image The closing operation can fill up small lakes, i.e. small holes, and close small cracks in the image, while the overall position and shape will not change. It is usually used to fill holes and small black spots in the image, eliminate small objects, fill holes and connect two separated objects, so as to make the image more continuous and smooth, which is convenient for subsequent processing.
[0110] Step S4: using a morphological method to enhance the image features of the first dynamic area, using a wavelet analysis method to extract the energy features of the enhanced first dynamic area, setting a threshold based on the energy features, determining the second dynamic area through threshold judgment, and using the second dynamic area as the oil mist area.
[0111] Preferably, in the embodiment of the present disclosure, the process of extracting the energy characteristics of the enhanced first dynamic region by using the wavelet analysis method is:
[0112] Decomposing the enhanced first dynamic region along one direction to generate two sub-bands: a low-frequency sub-band L and a high-frequency sub-band H;
[0113] The wavelet analysis method is based on two-dimensional wavelet transform and decomposes the image by applying low-pass and high-pass filters to generate four different sub-bands: low-frequency information sub-image LL1, horizontal high-frequency edge sub-image HL1, vertical high-frequency edge sub-image LH1 and diagonal high-frequency edge sub-image HH1, as shown in Figure 5 As shown;
[0114] The two-dimensional wavelet transform is:
[0115]
[0116] in, Represents a separable two-dimensional scaling function derived from a one-dimensional scaling function At the same time 1 (x,y),ψ 2 (x,y),ψ 3 (x, y) is a separable two-dimensional wavelet function with "directional perception" characteristics, which corresponds to the changes in the horizontal edge of the image in the column direction, the vertical edge in the row direction, and the edge grayscale in the diagonal direction. is a representation of a one-dimensional wavelet function;
[0117] The features related to the second dynamic change area are obtained by calculating and extracting the energy value of the high frequency sub-band H;
[0118] The energy value is calculated by summing the squares of all pixel values in the subband. The specific calculation formula is:
[0119] Energy=|LH| 2 +|HL| 2 +|HH| 2
[0120] The process of determining the second dynamic area by threshold determination is as follows:
[0121] Assume that Energy represents the energy value of the image to be detected at the position (x1, y1), and B(x1, y1) represents the energy value of the high-frequency component of the first static change area image at the corresponding position after two-dimensional discrete wavelet transform. If the energy value of the image to be detected is lower than the energy value of the first static area, the image block at this position is determined to be the second dynamic change area.
[0122] The condition for judging whether the second dynamic change area exists is:
[0123] T1×B(x1,y1) <E(x1,y1)<T2×B(x1,y1) (10)
[0124] If the energy value of the image block to be detected is between T1 and T2 times the energy value of the high-frequency component at the corresponding position of the background image, the image block is considered to be the second dynamic change area, and this area is marked as an oil mist area.
[0125] The oil mist recognition performance of the conventional image processing algorithm in the prior art and the algorithm proposed in the present invention is compared. The comparison data set adopts the data set published by Prof. Yuan Feiniu, such as Figure 2 As shown, Figure 2 (a) shows the oil mist scene image taken. Figure 2 (b) shows the recognition of oil mist particles using the frame difference method. Figure 2 (c) shows the identification of oil mist particles using the background difference method. Figure 2(d) shows the recognition of oil mist particles obtained by using an industrial building indoor oil mist image processing method provided by the present application. It can be seen that the method provided by the present application has better recognition accuracy.
[0126] In order to verify the detection of oil mist particles actually emitted by this algorithm, an oil mist emission test bench was further built to restore the oil mist emission of the aluminum rolling process in actual industrial buildings. Figure 3 The experimental platform is composed of an oil mist generator 4, an oil mist release device 3, a camera 2, a local exhaust hood 1 and a computer 5.
[0127] In the experimental setting, a special oil mist escape monitoring area was planned in the experimental area. This area is located in the upper right corner of the video monitoring screen, such as Figure 4 The establishment of this area is based on the preliminary observation and analysis of the oil mist escape behavior, with the aim of ensuring a clearly defined visual reference, thereby facilitating the accurate monitoring and analysis of the dynamic process of the oil mist, and ensuring the immediate identification and judgment of the potential escape of the oil mist during the experiment.
[0128] As for the experimental results, when the camera receives the frame image of oil mist emission, the image is transmitted to the computer for analysis and processing. When the algorithm detects oil mist in the ROI area, it will output the result, displaying 'smoke detected', and mark the oil mist particles detected in the target area in red. Furthermore, it can also output the range information of the oil mist particles in the image.
[0129] It should be noted that the operating system used in this embodiment is Windows 10, the CPU model is i9-13900HX, the GPU model is RTX4060, the video memory size is 48GB, and the memory size is 64GB. All models used are based on OpenCV, the cuda version is 11.7, and the cudnn version is 7.6.5.
[0130] The present disclosure also provides an industrial building indoor oil mist image processing system, including:
[0131] The acquisition unit is configured as:
[0132] Used to collect visual information of oil mist particles in the target area and determine the video frames in the visual information;
[0133] The model processing unit is configured as follows:
[0134] Used to describe the background of each pixel in the video frame based on a mixed Gaussian background model, and separate a first dynamically changing area and a first static area based on the background of each pixel;
[0135] The enhanced processing unit is configured to:
[0136] Used to enhance the oil mist image characteristics of the first dynamically changing area by using a morphological method;
[0137] Output unit, configured as:
[0138] The method is used to enhance the image features of the first dynamic area by using a morphological method, extract the energy features of the enhanced first dynamic area by using a wavelet analysis method, set a threshold based on the energy features, determine the second dynamic area by threshold judgment, and use the second dynamic area as the oil mist area.
[0139] In another embodiment of the present invention, a computer device is provided, the computer device includes a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of the indoor oil mist image processing method of industrial buildings.
[0140] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the industrial building indoor oil mist image processing method in the above embodiment.
[0141] 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.
[0142] The present invention is described with reference to flowcharts and / or block diagrams of methods, 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.
[0143] 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.
[0144] 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.
[0145] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.
[0146] In addition, it should be understood that although this specification is described in accordance with the implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that can be understood by those skilled in the art. The above content is only to illustrate the technical idea of the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.
Claims
1. A method for processing oil mist images in industrial buildings, characterized in that: The following steps are involved: Collect visual information of oil mist particles in the target area and determine video frames in the visual information; Describing the background of each pixel in the video frame based on a mixed Gaussian background model, and separating a first dynamically changing area and a first static area based on the background of each pixel; Using morphological method to enhance the oil mist image features in the first dynamic change area; A morphological method is used to enhance the image features of the first dynamic area, and a wavelet analysis method is used to extract the energy features of the enhanced first dynamic area. A threshold is set based on the energy features, and the second dynamic area is determined by threshold judgment, and the second dynamic area is used as the oil mist area.
2. The method for processing indoor oil mist images of industrial buildings according to claim 1, characterized in that: The process of describing the background of each pixel in the video frame is: The observed value of the pixel point is taken as a random variable, and the probability distribution of the random variable is composed of multiple Gaussian distributions, which represent the typical value of the pixel in different time periods; The mixed Gaussian background model assumes that each pixel observation value is I, and the probability distribution of It for a single sample can be expressed as: Where K is the number of Gaussian distribution quantities. The higher the K value, the better the system describes the complex environment; ω kt is the weight of the kth Gaussian distribution at time, which is used to reflect the importance of the Gaussian distribution in the model. The sum of all weights is equal to 1; η is the probability density function of the Gaussian distribution, which is defined as: According to the size of the weight coefficient or the weight coefficient ω kt The ratio of the Gaussian components to the standard deviation σ is used to sort all Gaussian components, determine the importance and order of the Gaussian components in the model, and divide the Gaussian distribution into the first dynamic change area and the first static area.
3. The method for processing indoor oil mist images of industrial buildings according to claim 1, characterized in that: When describing the background, several Gaussian distributions whose weighted cumulative values exceed a preset threshold T for the first time are selected as the first static area set, and the remaining Gaussian distributions are classified as the first dynamically changing area set; the first static area set is used as the boundary between the first dynamically changing area and the first static area by accumulating the weights of the sorted Gaussian distributions, and if the accumulated value exceeds the set threshold T; The background formula is expressed as: Where B is the number of Gaussian distributions used for the background.
4. The method for processing indoor oil mist images of industrial buildings according to claim 1, characterized in that: The process of separating the first dynamically changing area and the first static area based on the background of each pixel point is as follows: The mixed Gaussian background model includes K Gaussian distributions, each of which has a mean vector μ kt and standard deviation σ; The mixed Gaussian background model calculates the observation value I based on the arrangement order of the K Gaussian distributions t With the mean μ of each Gaussian distribution kt The distance between the observation values I t Match it with one of the Gaussian distributions; The matching principle is: when the distance of the Kth Gaussian distribution is less than or equal to λ times the standard deviation λσ of the distribution for the first time, the Gaussian distribution and the observation value I t Matching; the observed value I t is any pixel in the video frame; The matching function M of the pixel point kt If the value can be assigned to 1, it means the match is successful; otherwise, the matching function remains at 0, indicating a match failure; If the match is successful, the pixel point is used as the first static area, otherwise the pixel point is used as the first dynamically changing area.
5. The method for processing indoor oil mist images of industrial buildings according to claim 1, characterized in that: The mixed Gaussian background model is used to adapt to background changes by learning and updating the parameters of each Gaussian distribution; The parameters of the Gaussian distribution include weight, mean and variance; The rule for the mixed Gaussian background model to learn and update the parameters of each Gaussian distribution is: At time t, if there is no Gaussian distribution and observation value I t The mixed Gaussian background model will introduce a new Gaussian distribution, and the mean of the new Gaussian distribution is set to the current observation value I t , and set a larger initial variance and a smaller weight, and remove the existing distribution with the lowest weight; If the observed value I t Matching with a specific Gaussian distribution, the mixed Gaussian background model adjusts the weight, mean and variance according to the update mode of formula (5); In the formula, ω kt+1 is the weight coefficient of the kth Gaussian distribution after update at time t+1; α is the learning rate, ranging from 0<α<1; M kt Indicator function indicating successful matching, which is 1 if the matching is successful, otherwise 0; μ kt+1 is the updated mean of the kth Gaussian distribution at time t+1; ρ kt is the adaptive learning rate, indicating the observed value I t The degree of influence on the update of the mean of the Gaussian distribution; is the probability density function of the Gaussian distribution, and calculates the observed value I t The probability of the k-th Gaussian distribution; It is the updated variance of the kth Gaussian distribution at time t+1.
6. The method for processing oil mist images in industrial buildings according to claim 1, characterized in that: The morphological methods include dilation, erosion, opening and closing methods.
7. The method for processing oil mist images in industrial buildings according to claim 1, characterized in that: The process of extracting the energy characteristics of the enhanced first dynamic region by using the wavelet analysis method is as follows: Decomposing the enhanced first dynamic region along one direction to generate two sub-bands: a low-frequency sub-band L and a high-frequency sub-band H; The wavelet analysis method is based on two-dimensional wavelet transform and decomposes the image by applying low-pass and high-pass filters to generate four different sub-bands: a low-frequency information sub-image LL1, a horizontal high-frequency edge sub-image HL1, a vertical high-frequency edge sub-image LH1 and a diagonal high-frequency edge sub-image HH1; The two-dimensional wavelet transform is: in, Represents a separable two-dimensional scaling function derived from a one-dimensional scaling function At the same time 1 (x,y),φ 2 (x,y),φ 3 (x, y) is a separable two-dimensional wavelet function with "directional perception" characteristics, which corresponds to the changes in the horizontal edge of the image in the column direction, the vertical edge in the row direction, and the edge grayscale in the diagonal direction. is a representation of a one-dimensional wavelet function; The features related to the second dynamic change area are obtained by calculating and extracting the energy value of the high frequency sub-band H; The energy value is calculated by summing the squares of all pixel values in the subband. The specific calculation formula is: Energy=|LH| 2 +|HL| 2 +|HH| 2 The process of determining the second dynamic area by threshold determination is as follows: Assume that Energy represents the energy value of the image to be detected at the position (x1, y1), and B(x1, y1) represents the high-frequency component energy value of the first static change region image at the corresponding position after two-dimensional discrete wavelet transform. If the energy value of the image to be detected is lower than the energy value of the first static region, the image block at the position is determined to be the second dynamic change region. The condition for judging whether the second dynamic change area exists is: T1×T(x1,y1) <E(x1,y1)<T2×B(x1,y1) (10) If the energy value of the image block to be detected is between T1 and T2 times the energy value of the high-frequency component at the corresponding position of the background image, the image block is considered to be the second dynamic change area, and this area is marked as an oil mist area.
8. An industrial building indoor oil mist image processing system, characterized in that: The method for processing indoor oil mist images of industrial buildings according to any one of claims 1 to 7 comprises: The acquisition unit is configured as: Used to collect visual information of oil mist particles in the target area and determine the video frames in the visual information; The model processing unit is configured as follows: Used to describe the background of each pixel in the video frame based on a mixed Gaussian background model, and separate a first dynamically changing area and a first static area based on the background of each pixel; The enhanced processing unit is configured to: Used to enhance the oil mist image characteristics of the first dynamically changing area by using a morphological method; Output unit, configured as: The method is used to enhance the image features of the first dynamic area by using a morphological method, extract the energy features of the enhanced first dynamic area by using a wavelet analysis method, set a threshold based on the energy features, determine the second dynamic area by threshold judgment, and use the second dynamic area as the oil mist area.
9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for processing indoor oil mist images of industrial buildings as described in any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for processing indoor oil mist images of industrial buildings as described in any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Video smoke detection method based on mixed Gaussian model and morphological characteristics
CN101908141A
Method for extracting suspected smoke area from dynamic smoke
CN112258403A
Elevator car smoke detection method and device, electronic equipment and storage medium
CN115311622A
Fire smoke detection method and system and electronic equipment
CN115456976A
Method, device and equipment for extracting moving target under dynamic background and storage medium
CN115830059A
Cited By
Bag-type dust collector leakage detection method based on background difference method
CN120997174A
An openCV-based stitch clarity monitoring method and system
CN122694873A
