Arc discharge detection method and device and electronic equipment

Abstract

The invention relates to the technical field of power system monitoring, and discloses an arc discharge detection method and device and electronic equipment. The arc discharge detection method comprises the following steps: acquiring a visible light image, an infrared image and environmental parameters of a to-be-detected area; respectively determining a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters; obtaining fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient; and inputting the fusion features and the environmental parameters into an arc discharge detection model to obtain a detection result. According to the invention, the overall arc discharge state of the to-be-detected area can be reflected more comprehensively, the probability of false alarm or missing alarm is reduced, and the accuracy is higher.

Description

Technical Field

[0001] The present application relates to the technical field of power system monitoring, for example, to an arc discharge detection method, device and electronic equipment. Background Art

[0002] Arc discharge is a gas discharge phenomenon, which refers to the phenomenon that two electrodes are maintained conductive by gaseous charged particles (such as electrons or ions) under a certain voltage. Due to the influence of various factors such as equipment insulation aging and mechanical damage, arc discharge phenomenon often occurs in power equipment or power places such as cable trenches and distribution rooms. If arc discharge is not controlled, it is easy to cause major safety accidents such as fire or explosion.

[0003] Existing arc discharge monitoring methods include partial discharge detection methods. The partial discharge detection method detects the temperature changes and gas composition changes caused by partial discharge by installing temperature sensors, gas sensors, etc. to determine whether arc discharge occurs. However, the partial discharge detection method is limited by the arrangement position and number of sensors, and it is difficult to fully reflect the overall arc discharge state. It is easy to miss or misreport, so the accuracy is low. Summary of the invention

[0004] In order to provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. The summary is not an extensive review, nor is it intended to identify key / critical components or delineate the scope of protection of these embodiments, but rather serves as a prelude to the detailed description that follows.

[0005] The embodiments of the present disclosure provide an arc discharge detection method, device and electronic device, which can more comprehensively reflect the overall arc discharge state of the area to be detected, reduce the probability of false alarm or missed alarm, and have a higher accuracy rate.

[0006] In some embodiments, an arc discharge detection method is provided, including: acquiring a visible light image, an infrared image, and environmental parameters of an area to be detected; determining, according to the environmental parameters, a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image; obtaining fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient; and inputting the fusion features and the environmental parameters into an arc discharge detection model to obtain a detection result.

[0007] Optionally, according to the first weight coefficient and the second weight coefficient, fusion features of the visible light image and the infrared image are obtained, including: respectively decomposing the visible light image and the infrared image to obtain a first component of the visible light image at different frequency levels and a second component of the infrared image at different frequency levels; according to the first weight coefficient and the second weight coefficient, fusing the first component and the second component at the same frequency level to obtain a fused component; reconstructing the image according to the fused component to obtain a fused image; and extracting features from the fused image to obtain a fusion feature.

[0008] Optionally, according to the first weight coefficient and the second weight coefficient, a fusion feature of the visible light image and the infrared image is obtained, including: inputting the first weight coefficient, the second weight coefficient, the visible light image and the infrared image into a feature extraction model to obtain the fusion feature.

[0009] Optionally, according to the first weight coefficient and the second weight coefficient, fusion features of the visible light image and the infrared image are obtained, including: performing feature extraction on the visible light image to obtain first image features; performing feature extraction on the infrared image to obtain second image features; and fusing the first image features and the second image features according to the first weight coefficient and the second weight coefficient to obtain fusion features.

[0010] Optionally, the detection result includes arc discharge and normal state, and the fused features and environmental parameters are input into the arc discharge detection model to obtain the detection result, including: inputting the fused features and environmental parameters into the arc discharge detection model to obtain a first predicted probability that the detection result is arc discharge and a second predicted probability that the detection result is normal state; using Bayesian theorem to respectively calculate a first posterior probability corresponding to arc discharge and a second posterior probability corresponding to normal state; calculating a weighted average of the first predicted probability and the first posterior probability to obtain a first probability; and calculating a weighted average of the second predicted probability and the second posterior probability to obtain a second probability; when the first probability is greater than the second probability, determining that the detection result is arc discharge; otherwise, determining that the detection result is normal state.

[0011] Optionally, the fused features and environmental parameters are input into an arc discharge detection model to obtain a detection result, including: obtaining power consumption parameters of electrical equipment in the area to be detected; inputting the power consumption parameters, fused features and environmental parameters into the arc discharge detection model to obtain a detection result.

[0012] Optionally, an arc discharge detection model is obtained in the following manner: a training data set is obtained; the training data set is a data set including a visible light image stream, infrared images and environmental parameters containing arc discharge marks, and a visible light image, infrared images and environmental parameters containing normal state marks; an initial model is trained using the training data set to obtain a trained initial model; the initial model includes a recursive neural network or a long short-term memory network; and the trained initial model is used as an arc discharge detection model.

[0013] Optionally, the initial model is trained using a training data set, including: image preprocessing of visible light images and infrared images; image preprocessing including one or more of noise reduction, contrast enhancement, dehazing, color correction and edge enhancement; parameter preprocessing of environmental parameters; parameter preprocessing including filtering and / or denoising; and training the initial model using the visible light images and infrared images after image preprocessing, and the environmental parameters after parameter preprocessing.

[0014] Optionally, the initial model is trained using a training data set, including: determining a first parameter from environmental parameters according to time information of the visible light image and / or infrared image; combining the visible light image, the infrared image and the first parameter to obtain a training timing matrix; and training the initial model using the training timing matrix.

[0015] In some embodiments, an arc discharge detection device is provided, including: an acquisition module, configured to acquire visible light images, infrared images and environmental parameters of an area to be detected; a determination module, configured to determine a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters; a fusion module, configured to obtain fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient; and a detection module, configured to input the fusion features and environmental parameters into an arc discharge detection model to obtain a detection result.

[0016] In some embodiments, an arc discharge detection device is provided, including a processor and a memory storing program instructions, wherein the processor is configured to execute the arc discharge detection method as described in the above embodiments when running the program instructions.

[0017] In some embodiments, an electronic device is provided, comprising: a device body; and an arc discharge detection device as described in the above embodiments, installed on the device body.

[0018] The arc discharge detection method, device and electronic device provided by the embodiments of the present disclosure can achieve the following technical effects:

[0019] In the disclosed embodiments, visible light images and infrared images of the area to be detected are introduced to detect arc discharge phenomena. Visible light images can provide direct visual evidence to identify arc discharge phenomena in the area to be detected. Infrared images can intuitively display the overall temperature distribution of the area to be detected, which helps to quickly locate abnormal hot spots in the area to be detected and find out areas where arc discharge may occur or is occurring. Compared with the local discharge detection method of the related art, the present application uses visible light images, infrared images and environmental parameters to perform arc discharge detection, which can more comprehensively reflect the overall arc discharge state of the area to be detected, reduce the probability of false alarms or missed alarms, and have a higher accuracy rate.

[0020] The above general description and the following description are exemplary and explanatory only and are not intended to limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] One or more embodiments are exemplarily described by corresponding drawings, which do not limit the embodiments. Elements with the same reference numerals in the drawings are shown as similar elements, and the drawings do not constitute a scale limitation, and wherein:

[0022] Figure 1 is a schematic diagram of an electronic device provided by an embodiment of the present disclosure;

[0023] Figure 2 is a schematic diagram of an arc discharge detection method provided by an embodiment of the present disclosure;

[0024] Figure 3 is a schematic diagram of an arc discharge detection method provided by another embodiment of the present disclosure;

[0025] Figure 4 is a schematic diagram of a method for training an arc discharge detection model provided by an embodiment of the present disclosure;

[0026] Figure 5 is a schematic diagram of an arc discharge detection device provided by an embodiment of the present disclosure;

[0027] Figure 6 is a schematic diagram of an arc discharge detection device provided in another embodiment of the present disclosure. DETAILED DESCRIPTION

[0028] In order to be able to understand the features and technical contents of the embodiments of the present disclosure in more detail, the implementation of the embodiments of the present disclosure is described in detail below in conjunction with the accompanying drawings. The attached drawings are for reference only and are not used to limit the embodiments of the present disclosure. In the following technical description, for the convenience of explanation, a full understanding of the disclosed embodiments is provided through multiple details. However, one or more embodiments can still be implemented without these details. In other cases, to simplify the drawings, well-known structures and devices can be simplified for display.

[0029] The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the embodiments of the present disclosure described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0030] Unless otherwise stated, the term "plurality" means two or more.

[0031] In the embodiment of the present disclosure, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B indicates: A or B.

[0032] The term "and / or" is a description of the association relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A, B, A and B.

[0033] The term "correspondence" may refer to an association relationship or a binding relationship. The correspondence between A and B means that there is an association relationship or a binding relationship between A and B.

[0034] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0035] Combination Figure 1 As shown, the embodiment of the present disclosure provides an electronic device 1, comprising a device body 10 and an arc discharge detection device 50 (60). The arc discharge detection device 50 (60) is installed on the device body 10.

[0036] In the electronic device 1 provided in the embodiment of the present disclosure, the arc discharge detection device 50 (60) is installed in the device body 10. The installation relationship described here is not limited to being placed inside the device body 10, but also includes installation connections with other components of the electronic device 1, including but not limited to physical connections, electrical connections or signal transmission connections, etc. It can be understood by those skilled in the art that the arc discharge detection device 50 (60) can be adapted to a feasible electronic device 1, thereby realizing other feasible embodiments.

[0037] Optionally, the arc discharge detection device 60 includes a processor 600. The processor 600 can obtain a visible light image, an infrared image, and environmental parameters of the area to be detected; can determine a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters; can obtain fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient; can input the fusion features and the environmental parameters into the arc discharge detection model to obtain the detection result.

[0038] Combination Figure 1 and Figure 6 The electronic device 1 shown in the figure, the embodiment of the present disclosure provides an arc discharge detection method, such as Figure 2 As shown, including:

[0039] S201, the processor obtains a visible light image, an infrared image and environmental parameters of the area to be detected.

[0040] In this step, the area to be detected refers to a specific spatial range where arc discharge detection is required. The area to be detected is different based on the different targets for arc discharge detection. For example, when arc discharge detection is required inside a distribution box, the area to be detected is the internal space of the distribution box; when arc discharge detection is required on a cable trench, the area to be detected is the pipe space of the cable trench.

[0041] In this step, a visible light camera can be used to acquire visible light images of the area to be detected, and a thermal imaging infrared camera can be used to acquire infrared images of the area to be detected. Sensors (such as temperature sensors, humidity sensors, and air pressure sensors) installed at intervals in the area to be detected can be used to acquire environmental parameters.

[0042] The environmental parameters in this step include one or more of temperature, humidity, and air pressure. During the arc discharge process, it is inevitable to cause changes in the ambient temperature, because the accelerated motion of electrons and ions in the electric field and the collision of gas molecules cause the gas molecules to move violently and collide with each other, thereby causing the gas temperature to rise. In a humid environment, it is easy to form a conductive path, which reduces the insulation performance, thereby increasing the risk of partial discharge. And the arc discharge starting voltage of the uneven electric field is low, the movement speed of free electrons is slow, the energy is small, and it is more susceptible to humidity. Air pressure is also one of the important factors affecting arc discharge, because with the increase of air pressure, the collision of electrons and gas molecules is more frequent, and it is easier to form an ionization channel, and the arc discharge voltage will be reduced, which is more likely to cause arc discharge.

[0043] S202: The processor determines a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters.

[0044] In this step, a correspondence table between environmental parameters and the first weight coefficient and the second weight coefficient is preset. After obtaining the environmental parameters, the first weight coefficient corresponding to the visible light image and the second weight coefficient corresponding to the infrared image under the current environmental parameters can be determined by searching the correspondence table between the environmental parameters and the first weight coefficient and the second weight coefficient, thereby determining the first weight coefficient and the second weight coefficient.

[0045] S203: The processor obtains a fusion feature of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient.

[0046] S204, the processor inputs the fusion features and environmental parameters into the arc discharge detection model to obtain the detection results.

[0047] The arc discharge detection method provided by the embodiment of the present disclosure can dynamically determine the first weight coefficient corresponding to the visible light image and the second weight coefficient corresponding to the infrared image according to the environmental parameters after obtaining the visible light image, infrared image and environmental parameters of the area to be detected, so as to adjust the contribution of different images in the fusion feature according to the current environmental conditions, so as to adapt to different detection scenarios. Then, the fusion feature of the visible light image and the infrared image is obtained by using the first weight coefficient and the second weight coefficient, and the fusion feature and the environmental parameters are input into the arc discharge detection model to obtain the detection result of the arc discharge detection in the area to be detected. By obtaining the fusion feature of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient, and inputting the fusion feature and the environmental parameters into the arc discharge detection model to detect the arc discharge phenomenon, it is achieved that both the information of the image itself and the influence of the environmental conditions on the detection are considered, the probability of false alarm or missed alarm caused by changes in environmental conditions is reduced, and the accuracy of detection is improved. In addition, visible light images and infrared images each have their own advantages in capturing arc discharge. By obtaining the fusion features of visible light images and infrared images as the input data of the arc discharge detection model, the information complementarity of visible light images and infrared images can be fully utilized to improve the comprehensiveness of detection.

[0048] In the disclosed embodiments, visible light images and infrared images of the area to be detected are introduced to detect arc discharge phenomena. Visible light images can provide direct visual evidence to identify arc discharge phenomena in the area to be detected. Infrared images can intuitively display the overall temperature distribution of the area to be detected, which helps to quickly locate abnormal hot spots in the area to be detected and find out areas where arc discharge may occur or is occurring. Compared with the local discharge detection method of the related art, the present application uses visible light images, infrared images and environmental parameters to perform arc discharge detection, which can more comprehensively reflect the overall arc discharge state of the area to be detected, reduce the probability of false alarms or missed alarms, and have a higher accuracy rate.

[0049] Optionally, according to the first weight coefficient and the second weight coefficient, fusion features of the visible light image and the infrared image are obtained, including: respectively decomposing the visible light image and the infrared image to obtain a first component of the visible light image at different frequency levels and a second component of the infrared image at different frequency levels; according to the first weight coefficient and the second weight coefficient, fusing the first component and the second component at the same frequency level to obtain a fused component; reconstructing the image according to the fused component to obtain a fused image; and extracting features from the fused image to obtain a fusion feature.

[0050] In this embodiment, by first decomposing the visible light image and the infrared image respectively, the image information is divided according to different frequency levels to obtain the first component of the visible light image at different frequency levels and the second component of the infrared image at different frequency levels. Then, according to the determined first weight coefficient and the second weight coefficient, the first component and the second component at the same frequency level are fused. The fused components are used to reconstruct the image to obtain a fused image, so that the fused image has the characteristics of both the visible light image and the infrared image.

[0051] In this embodiment, the visible light image is rich in detail information such as texture and edge, while the infrared image has high brightness and high contrast, which can highlight the target information. Through fusion processing, the information of the visible light image and the infrared image can be complemented to obtain an image with high brightness, high contrast and rich texture details, that is, a fused image. The fused image combines the advantages of the visible light image and the infrared image, making the target information in the image more prominent and the detail information richer, which helps to improve the accuracy of fusion feature extraction, thereby improving the accuracy of arc discharge detection model detection.

[0052] In some embodiments, the visible light image and the infrared image are decomposed respectively to obtain the first component of the visible light image at different frequency levels and the second component of the infrared image at different frequency levels, including: using a multi-resolution algorithm such as wavelet transform, non-subsampled contourlet transform or sparse representation to decompose the visible light image into the first component at different frequency levels; using a multi-resolution algorithm such as wavelet transform, non-subsampled contourlet transform or sparse representation to decompose the infrared image into the second component at different frequency levels. The first component and the second component generally include a low-frequency component (representing the main structure and background of the image) and a high-frequency component (representing the details and edge information of the image).

[0053] Optionally, the frequency level includes a low-frequency level and a high-frequency level. According to the first weight coefficient and the second weight coefficient, the first component and the second component at the same frequency level are fused to obtain a fused component, including: in the low-frequency level part, using the first weight coefficient and the second weight coefficient to calculate the weighted average value of the first component and the second component to obtain the fused component; in the high-frequency level part, using the maximum value selection or sparse coefficient optimization method to fuse the first component and the second component to obtain the fused component.

[0054] In this embodiment, the frequency level corresponds to the low-frequency component, and the high-frequency level corresponds to the high-frequency component. By using the first weight coefficient and the second weight coefficient to calculate the weighted average value of the first component and the second component in the low-frequency level part, the smooth fusion of information is achieved at the low-frequency level, the main structure and background information of the image are retained, and the fused image is visually smoother and more natural. By using the maximum value selection or sparse coefficient optimization method to fuse the first component of the visible light image and the second component of the infrared image in the high-frequency level part, the edge and detail information in the image are highlighted, and the clarity and resolution of the fused image are improved. The maximum value selection method compares the values ​​of the pixels at the same position and selects the larger value as the fused pixel value to highlight the edge and detail information in the image. The sparse coefficient optimization method constructs a sparse representation model and optimizes the solution of the sparse coefficient to achieve fine fusion of high-frequency information.

[0055] It should be noted that reconstructing the image according to the fused components to obtain the fused image is the inverse process of image decomposition, so the reconstruction algorithm corresponding to the decomposition method is used, that is, the inverse algorithm corresponding to the decomposition method.

[0056] Optionally, before decomposing the visible light image and the infrared image respectively, obtaining the fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient, further comprising: performing image registration on the visible light image and the infrared image. The visible light image and the infrared image are spatially aligned through image registration so that the same feature points in the visible light image and the infrared image can correspond one to one, thereby eliminating the image misalignment problem caused by factors such as shooting angle and focal length.

[0057] Optionally, according to the first weight coefficient and the second weight coefficient, a fusion feature of the visible light image and the infrared image is obtained, including: inputting the first weight coefficient, the second weight coefficient, the visible light image and the infrared image into a feature extraction model to obtain the fusion feature.

[0058] In this embodiment, the feature extraction model may be a feature layer of a convolutional neural network. The first weight coefficient, the second weight coefficient, the visible light image, and the infrared image are input into the same feature extraction model to share low-level features, perform feature fusion, and output fusion features.

[0059] Optionally, according to the first weight coefficient and the second weight coefficient, fusion features of the visible light image and the infrared image are obtained, including: performing feature extraction on the visible light image to obtain first image features; performing feature extraction on the infrared image to obtain second image features; and fusing the first image features and the second image features according to the first weight coefficient and the second weight coefficient to obtain fusion features.

[0060] In this embodiment, feature extraction is first performed on the visible light image and the infrared image respectively to obtain the first image feature (such as edge, corner point, texture, brightness, etc. in the visible light image) and the second image feature (such as shape contour, temperature distribution, brightness in the infrared image), and then the first image feature and the second image feature are fused to ensure the comprehensiveness of the extraction, thereby improving the accuracy of the arc discharge detection model. In this embodiment, the specific method of fusing the second image feature and the second image feature can be achieved by weighted summation or feature splicing, and the specific method of fusion is determined according to the specific second image feature and the second image feature type. Exemplarily, when the first image feature and the second image feature are both brightness, the weighted summation method can be used for fusion; when the first image feature is an edge and the second image feature is a shape contour, the feature splicing method can be used for fusion.

[0061] In some embodiments, feature extraction is performed on visible light images and infrared images to obtain fusion features of the visible light images and infrared images, including: decomposing the visible light image and the infrared image respectively to obtain a first component of the visible light image at different frequency levels and a second component of the infrared image at different frequency levels; fusing the first component and the second component at the same frequency level according to a first weight coefficient and a second weight coefficient to obtain a fused component; reconstructing an image according to the fused component to obtain a fused image; extracting features from the fused image to obtain a first intermediate feature; inputting the first weight coefficient, the second weight coefficient, the visible light image and the infrared image into a feature extraction model to obtain a second intermediate feature; extracting features from the visible light image to obtain a first image feature; extracting features from the infrared image to obtain a second image feature; fusing the first image feature and the second image feature according to the first weight coefficient and the second weight coefficient to obtain a third intermediate feature; fusing the first intermediate feature, the second intermediate feature and the third intermediate feature to obtain a fusion feature of the visible light image and the infrared image.

[0062] In this embodiment, the fusion features of the visible light image and the infrared image are obtained by fusing the first intermediate feature, the second intermediate feature and the third intermediate feature. In this embodiment, the fusion process of the first intermediate feature, the second intermediate feature and the third intermediate feature can be achieved by weighted summation, maximum value selection or feature splicing. In this embodiment, the fusion features of the visible light image and the infrared image are obtained based on the fusion of the first intermediate feature, the second intermediate feature and the third intermediate feature. Therefore, the technical effects of the first intermediate feature, the second intermediate feature and the third intermediate feature are all possessed by the embodiments of the present disclosure and will not be repeated here. For the technical effects of the first intermediate feature, the second intermediate feature and the third intermediate feature, refer to the corresponding specific embodiments above.

[0063] Optionally, the detection result includes arc discharge and normal state. Inputting the fusion features and environmental parameters into the arc discharge detection model to obtain the detection result includes: inputting the fusion features and environmental parameters into the arc discharge detection model to obtain a first predicted probability that the detection result is arc discharge and a second predicted probability that the detection result is normal state; using Bayes' theorem to calculate the first posterior probability corresponding to the arc discharge and the second posterior probability corresponding to the normal state respectively; calculating the weighted average of the first predicted probability and the first posterior probability to obtain the first probability; and calculating the weighted average of the second predicted probability and the second posterior probability to obtain the second probability; when the first probability is greater than the second probability, determining that the detection result is arc discharge; otherwise, determining that the detection result is normal state.

[0064] In this embodiment, the output results of the arc discharge detection model and the Bayesian decision theory can be combined to comprehensively judge the results of arc discharge and normal state, so as to more accurately judge whether there is an arc discharge phenomenon, reduce false alarms, and improve the accuracy of the judgment results.

[0065] It should be noted that the weight coefficients corresponding to the first predicted probability and the first posterior probability, as well as the weight coefficients corresponding to the second predicted probability and the second posterior probability in the embodiment of the present disclosure are pre-set by technical personnel according to specific application requirements, and this application does not make specific limitations here.

[0066] Optionally, Bayesian theorem is used to calculate a first a posteriori probability corresponding to arc discharge and a second a posteriori probability corresponding to a normal state, respectively, including: obtaining a first a priori probability corresponding to arc discharge and a second a priori probability corresponding to a normal state; respectively obtaining a first likelihood function and a second likelihood function that the fusion features belong to arc discharge and a normal state; respectively obtaining a third likelihood function and a fourth likelihood function that the environmental parameters belong to arc discharge and a normal state; and using Bayesian theorem to calculate a first a posteriori probability corresponding to arc discharge and a second a posteriori probability corresponding to a normal state according to the first a priori probability, the second a priori probability, the first likelihood function, the second likelihood function, the third likelihood function and the fourth likelihood function.

[0067] In this embodiment, the first prior probability corresponding to arc discharge and the second prior probability corresponding to normal state can be pre-set according to historical data or expert experience. The first prior probability and the second prior probability respectively reflect the possibility of arc discharge and normal state respectively when no characteristic information is observed. The likelihood function refers to the probability of the observed data occurring under given model parameters (or assumptions), that is, the likelihood function of arc discharge: P(X|arc discharge), which indicates the probability of observing data X under the assumption of arc discharge; the likelihood function of normal state: P(X|normal state), which indicates the probability of observing data X under the assumption of normal state. In this embodiment, the likelihood function can be obtained by the output of the arc discharge detection model.

[0068] Using Bayesian theorem, combined with likelihood function and prior probability, the posterior probabilities corresponding to arc discharge and normal state are calculated respectively. The posterior probability reflects the possibility of arc discharge and normal state after the characteristic information is observed. The specific formula of Bayesian theorem is as follows:

[0069]

[0070] Among them, P(H|E) is the posterior probability, which indicates the probability that hypothesis H is true after observing event E; P(E|H) is the likelihood function, which indicates the probability of observing event E when hypothesis H is true; P(H) is the prior probability, which indicates the probability that hypothesis H is true before observing any data; P(E) is the total probability of observing event E, which can be calculated by summing all possible hypotheses.

[0071] In this embodiment, arc discharge and normal state may be regarded as two hypotheses H1 and H2 respectively, and the observation data E may be a combination of fusion features and environmental parameters.

[0072] In some embodiments, electrical equipment, such as a distribution box, is provided in the area to be detected. Figure 3 As shown, the embodiment of the present disclosure provides another arc discharge detection method, comprising:

[0073] S301, the processor obtains a visible light image, an infrared image and environmental parameters of the area to be detected.

[0074] S302: The processor determines a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters.

[0075] S303: The processor obtains a fusion feature of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient.

[0076] S304: The processor obtains power consumption parameters of electrical equipment in the area to be detected.

[0077] The power parameters in this step include current and / or voltage, which can reflect the operating state of the electrical equipment and the possibility of arc discharge.

[0078] S305, the processor inputs the power consumption parameters, fusion characteristics and environmental parameters into the arc discharge detection model to obtain the detection results.

[0079] The arc discharge detection method provided by the embodiment of the present disclosure can further combine the power consumption parameters of the electrical equipment in the area to be detected, consider the operating status of the electrical equipment during operation, and perform arc discharge phenomenon detection. By fusing multiple feature information (including the fusion features of visible light images and infrared images, electrical parameters, and environmental parameters), the arc discharge detection model can more comprehensively identify arc discharges, thereby further improving the accuracy of detection.

[0080] It should be noted that in the embodiments of the present disclosure, feature extraction of image data (such as fused images, visible light images, and infrared images) includes feature extraction through computer vision technology, such as scale-invariant feature transformation, accelerated robust features, or feature layers in convolutional neural networks, to extract multi-scale features such as morphology, texture, and color in the image.

[0081] In this embodiment, the Bayesian theorem is used to calculate the first a posteriori probability corresponding to the arc discharge and the second a posteriori probability corresponding to the normal state, including: obtaining the first a priori probability corresponding to the arc discharge and the second a priori probability corresponding to the normal state; respectively obtaining the first likelihood function and the second likelihood function of the fusion features belonging to the arc discharge and the normal state; respectively obtaining the third likelihood function and the fourth likelihood function of the environmental parameters belonging to the arc discharge and the normal state; respectively obtaining the fifth likelihood function and the sixth likelihood function of the power consumption parameters belonging to the arc discharge and the normal state; and according to the first a priori probability, the second a priori probability, the first likelihood function, the second likelihood function, the third likelihood function, the fourth likelihood function, the fifth likelihood function and the sixth likelihood function, using the Bayesian theorem to calculate the first a posteriori probability corresponding to the arc discharge and the second a posteriori probability corresponding to the normal state.

[0082] In some embodiments, the arc discharge detection model is obtained in the following manner: Figure 4 As shown, the embodiment of the present disclosure provides a method for training an arc discharge detection model, comprising:

[0083] S401: The processor obtains a training data set.

[0084] The training data set includes visible light images, infrared images and environmental parameters containing arc discharge marks, and visible light images, infrared images and environmental parameters containing normal state marks.

[0085] S402: The processor trains the initial model using the training data set to obtain a trained initial model.

[0086] Initial models include recurrent neural networks or long short-term memory networks.

[0087] Recurrent Neural Networks (RNN) are artificial neural networks with a tree-like hierarchical structure and network nodes that recursively process input information in the order in which they are connected. Recurrent neural networks loop the hierarchy of the neural network so that neurons in the same layer can share weights and states, thereby being able to process long sequence data, that is, the dependencies between data in time or space. Recurrent neural networks use cyclic connections to take the output of the previous moment as the input of the next moment, thereby capturing dynamic information in the sequence.

[0088] Long Short-Term Memory (LSTM) is an improved version of recursive neural network. LSTM introduces gating mechanism (input gate, forget gate and output gate) to control the flow of information, which can solve the shortcomings of recursive neural network in long-term dependency problem.

[0089] S403: The processor uses the trained initial model as an arc discharge detection model.

[0090] In the training method of the arc discharge detection model provided in the embodiment of the present disclosure, the arc discharge detection model is obtained by training a recursive neural network or a long short-term memory network. By utilizing the dynamic information capture capability of the recursive neural network or the long short-term memory network, the arc discharge detection model obtained by training can realize the pre-judgment of the arc discharge phenomenon, that is, when the arc discharge phenomenon has not occurred, it can judge whether the arc discharge phenomenon will occur in the future according to the current input data.

[0091] In the embodiment of the present disclosure, visible light images, infrared images and environmental parameters containing arc discharge marks, as well as visible light images, infrared images and environmental parameters containing normal state marks are used to train recursive neural networks or long short-term memory networks to obtain arc discharge detection models, so that the arc discharge detection model can learn the changing rules of dynamic characteristics of input data that change over time, thereby pre-judging arc discharge phenomena. Compared with the local discharge detection method in the related art, the arc discharge detection model trained in the embodiment of the present disclosure can realize the pre-judgment of arc discharge phenomena and improve the sensitivity of arc discharge phenomenon detection.

[0092] In some embodiments, the training data set also includes power parameters of electrical equipment in the area to be detected containing arc discharge marks, and power parameters of electrical equipment in the area to be detected containing normal state marks. By inputting environmental parameters and power parameters as input data into the initial model for training, the initial model can capture the dynamic characteristics of environmental parameters and power parameters that change over time, such as temperature gradient, humidity gradient, air pressure fluctuation amplitude, current fluctuation amplitude, voltage mutation, etc., learn the changing rules therein, and thus have the ability to predict future environmental conditions.

[0093] In some embodiments, data such as visible light images, infrared images, environmental parameters, and power consumption parameters can be collected through preliminary experiments to obtain a training data set. For example, take the cable trench as the area to be detected. Install the visible light camera and the infrared camera at key positions in the cable trench to ensure that the visible light camera and the infrared camera can cover all areas where arc discharge may occur. The visible light camera is used to capture visible light images or videos in the cable trench tunnel. The infrared camera is used to capture infrared images or videos in the cable trench tunnel. Different sensors, such as voltage sensors, current sensors, humidity sensors, temperature sensors, and air pressure sensors, are arranged at intervals in the cable trench to collect environmental parameters and power consumption parameters. The arc discharge phenomenon is randomly triggered in the cable trench. Control the visible light camera, infrared camera, and different sensors to synchronously collect visible light images, infrared images, environmental parameters, and power consumption parameters of the cable trench. According to the actual situation in the cable trench (i.e., whether arc discharge occurs), the collected visible light images, infrared images, environmental parameters, and electricity usage parameters are marked. The visible light images, infrared images, environmental parameters, and electricity usage parameters where arc discharge occurs are marked as arc discharge, and the visible light images, infrared images, environmental parameters, and electricity usage parameters where no arc discharge occurs are marked as normal, to obtain a training data set.

[0094] When the visible light camera and / or infrared camera captures a video stream, multiple visible light images and / or multiple infrared images can be exported from the video stream captured by the visible light camera and the infrared camera according to a preset frame rate to ensure the temporal continuity of the multiple visible light images and / or multiple infrared images.

[0095] In some embodiments, the training data set can also be constructed by accessing the Internet or applet to obtain public data (i.e., including visible light images, infrared images, environmental parameters and power parameters containing arc discharge marks, and visible light images, infrared images, environmental parameters and power parameters containing normal state marks). The visible light images, infrared images, environmental parameters and power parameters are continuous at the time of acquisition, so as to facilitate the training of the initial model in the embodiment of the present disclosure, so that the initial model has the ability to predict future environmental conditions.

[0096] Optionally, the initial model is trained using a training data set, including: image preprocessing of visible light images and infrared images; image preprocessing including one or more of noise reduction, contrast enhancement, dehazing, color correction and edge enhancement; parameter preprocessing of environmental parameters; parameter preprocessing including filtering and / or denoising; and training the initial model using the visible light images and infrared images after image preprocessing, and the environmental parameters after environmental parameter preprocessing.

[0097] In this embodiment, the visible light image and the infrared image, as well as the environmental parameters, can be preprocessed separately to effectively improve the data quality, reduce noise interference, and provide high-quality input data for model training.

[0098] Among them, noise reduction processing includes median filtering and Gaussian filtering. Median filtering sets the gray value of each pixel to the median of the gray values ​​of all pixels in a certain neighborhood window of the point. Through sorting statistics theory, the surrounding pixel values ​​are made close to the true value, thereby eliminating isolated noise points. Gaussian filtering performs weighted averaging on the entire image. The value of each pixel is obtained by weighted averaging the value of itself and other pixels in the neighborhood to remove noise in the image. Removing random noise in the image through noise reduction processing can improve image clarity.

[0099] Contrast enhancement is achieved through histogram equalization. Histogram equalization improves the visual effect of the image by changing the grayscale distribution of the image to make the image brightness distribution more uniform. Through contrast enhancement, the image contrast is adjusted to make the arc discharge area more prominent. Defogging is achieved through dark channel prior technology and atmospheric scattering model. Defogging is performed to ensure the true color and details of the image. Color correction is achieved through white balance adjustment. Through white balance adjustment, it is ensured that the white objects in the image appear truly white, improve the color restoration of the entire image, and ensure the accuracy of color information. Edge enhancement is achieved by detecting edge points in the image through the Sobel operator and enhancing the edge points. The enhancement can be achieved by increasing the pixel value, changing the pixel value contrast, etc. The specific method of enhancement can be selected according to actual needs. Through edge enhancement, the edge information in the image is highlighted.

[0100] Environmental parameters are filtered and / or denoised to remove outliers and / or noise.

[0101] In some embodiments, the training data set also includes power usage parameters. Using the training data set to train the initial model includes: performing image preprocessing on visible light images and infrared images; image preprocessing includes one or more of noise reduction, contrast enhancement, defogging, color correction, and edge enhancement; performing parameter preprocessing on environmental parameters and power usage parameters; parameter preprocessing includes filtering and / or denoising; using the preprocessed visible light images and infrared images, as well as the preprocessed environmental parameters and power usage parameters to train the initial model.

[0102] Optionally, the initial model is trained using a training data set, including: determining a first parameter from environmental parameters according to time information of the visible light image and / or infrared image; combining the visible light image, the infrared image and the first parameter to obtain a training timing matrix; and training the initial model using the training timing matrix.

[0103] In this embodiment, the time information of the visible light image and / or the infrared image includes the acquisition time of the visible light image and / or the infrared image. Determining the first parameter from the environmental parameter according to the time information of the visible light image and / or the infrared image includes: taking the environmental parameter with the same acquisition time as the visible light image and / or the infrared image as the first parameter. Then, the visible light image, the infrared image and the first parameter are combined to construct a time series matrix containing multiple information sources, that is, a training time series matrix. Among them, each row of the training time series matrix represents a different time node (the time node of each data in the same row, or the acquisition time is the same), each column represents different data (that is, the visible light image, the infrared image and the first parameter), and the elements in the training time series matrix represent the values ​​of different data at a specific time node. The initial model is trained based on the constructed training time series matrix, so that the initial model can learn the time series characteristics and laws in the data set, so as to have the ability to predict future environmental conditions.

[0104] In some embodiments, the training data set also includes power usage parameters. Using the training data set to train the initial model includes: determining a first parameter from environmental parameters according to time information of the visible light image and / or infrared image; determining a second parameter from power usage parameters according to time information of the visible light image and / or infrared image; combining the visible light image, the infrared image, the first parameter and the second parameter to obtain a training time sequence matrix; and using the training time sequence matrix to train the initial model.

[0105] In some embodiments, the detection result includes arc discharge or normal state. In the case where the detection result is arc discharge, the detection method further includes: obtaining arc discharge intensity; and determining and executing a target adjustment strategy according to the arc discharge.

[0106] In this embodiment, the arc discharge intensity can be obtained from the output of the arc discharge detection model. In the arc discharge detection model training process, the training data containing the arc discharge mark can be further marked, including marking the arc discharge intensity, so that the trained arc discharge detection model has the ability to recognize the arc discharge intensity. It is also possible to pre-set multiple brightness thresholds and the corresponding relationship between the comprehensive brightness of the visible light image and the infrared image and the arc discharge intensity (such as the comprehensive brightness of the visible light image and the infrared image is proportional to the arc discharge intensity), extract the brightness information of the visible light image and the infrared image of the area to be detected and calculate the weighted average value, obtain the comprehensive brightness of the visible light image and the infrared image, and compare the comprehensive brightness with multiple brightness thresholds to judge the arc discharge intensity.

[0107] The obtained arc discharge intensity can be expressed in numerical values ​​or words (such as mild, moderate or severe). In this embodiment, the correspondence between different arc discharge intensities and different adjustment strategies is preset. After the arc discharge intensity is obtained, the target adjustment strategy can be determined and executed according to the correspondence between different arc discharge intensities and different adjustment strategies to avoid the occurrence of arc discharge or reduce the arc discharge intensity.

[0108] Exemplarily, when the arc discharge intensity is severe, the target adjustment strategy is to immediately cut off the circuit through a relay to prevent the accident from expanding; when the arc discharge intensity is mild, the target adjustment strategy is to adjust the current or voltage to suppress the development of the arc and avoid unnecessary comprehensive power outages that affect normal operations.

[0109] Combination Figure 5 As shown, the embodiment of the present disclosure provides an arc discharge detection device 50, including an acquisition module 510, a determination module 520, a fusion module 530 and a detection module 540. The acquisition module 510 is configured to acquire a visible light image, an infrared image and environmental parameters of the area to be detected; the determination module 520 is configured to determine a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters; the fusion module 530 is configured to obtain fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient; the detection module 540 is configured to input the fusion features and the environmental parameters into the arc discharge detection model to obtain the detection results.

[0110] The arc discharge detection device 50 provided in the embodiment of the present disclosure can execute the arc discharge detection method described in the above embodiment. Therefore, the technical effects possessed by the arc discharge detection method described in the above embodiment are also possessed by the embodiment of the present disclosure and will not be described in detail here.

[0111] Optionally, the determination module 520 is also configured to decompose the visible light image and the infrared image respectively to obtain the first component of the visible light image at different frequency levels and the second component of the infrared image at different frequency levels; fuse the first component and the second component at the same frequency level according to the first weight coefficient and the second weight coefficient to obtain a fused component; reconstruct the image according to the fused component to obtain a fused image; and extract features from the fused image to obtain fused features.

[0112] Optionally, the determination module 520 is also configured to obtain fusion features of the visible light image and the infrared image based on the first weight coefficient and the second weight coefficient, including: inputting the first weight coefficient, the second weight coefficient, the visible light image and the infrared image into a feature extraction model to obtain the fusion features.

[0113] Optionally, the determination module 520 is also configured to obtain fusion features of the visible light image and the infrared image based on the first weight coefficient and the second weight coefficient, including: performing feature extraction on the visible light image to obtain first image features; performing feature extraction on the infrared image to obtain second image features; and fusing the first image features and the second image features according to the first weight coefficient and the second weight coefficient to obtain fusion features.

[0114] Optionally, the detection result includes arc discharge and normal state. The detection module 540 is also configured to input the fusion features and environmental parameters into the arc discharge detection model to obtain a first predicted probability that the detection result is arc discharge and a second predicted probability that the detection result is normal state; use Bayes' theorem to calculate the first posterior probability corresponding to the arc discharge and the second posterior probability corresponding to the normal state respectively; calculate the weighted average of the first predicted probability and the first posterior probability to obtain the first probability; and calculate the weighted average of the second predicted probability and the second posterior probability to obtain the second probability; when the first probability is greater than the second probability, determine that the detection result is arc discharge; otherwise, determine that the detection result is normal state.

[0115] Optionally, the detection module 540 is further configured to obtain power usage parameters of electrical equipment in the area to be detected; input the power usage parameters, fusion features and environmental parameters into the arc discharge detection model to obtain the detection results.

[0116] Optionally, the arc discharge detection device 50 further includes a training module 550. The training module 550 is configured to obtain a training data set; the training data set is a data set including a visible light image stream, an infrared image and environmental parameters containing arc discharge marks, and a visible light image, an infrared image and environmental parameters containing normal state marks; the training data set is used to train the initial model to obtain a trained initial model; the initial model includes a recursive neural network or a long short-term memory network; and the trained initial model is used as the arc discharge detection model.

[0117] Optionally, the training module 550 is also configured to perform image preprocessing on visible light images and infrared images; image preprocessing includes one or more of noise reduction, contrast enhancement, dehazing, color correction and edge enhancement; parameter preprocessing is performed on environmental parameters; parameter preprocessing includes filtering and / or denoising; and the initial model is trained using the visible light images and infrared images after image preprocessing, and the environmental parameters after parameter preprocessing.

[0118] Optionally, the training module 550 is also configured to determine a first parameter from environmental parameters according to time information of the visible light image and / or infrared image; combine the visible light image, infrared image and first parameter to obtain a training timing matrix; and train the initial model using the training timing matrix.

[0119] Combination Figure 6 As shown, an embodiment of the present disclosure provides an arc discharge detection device 60, including a processor (processor) 600 and a memory (memory) 601. Optionally, the device 60 may also include a communication interface (Communication Interface) 602 and a bus 603. Among them, the processor 600, the communication interface 602, and the memory 601 can communicate with each other through the bus 603. The communication interface 602 can be used for information transmission. The processor 600 can call the logic instructions in the memory 601 to execute the arc discharge detection method of the above embodiment.

[0120] In addition, the logic instructions in the memory 601 described above can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.

[0121] The memory 601 is a computer-readable storage medium that can be used to store software programs and computer executable programs, such as program instructions / modules corresponding to the method in the embodiment of the present disclosure. The processor 600 executes the function application and data processing by running the program instructions / modules stored in the memory 601, that is, implementing the arc discharge detection method in the above embodiment.

[0122] The memory 601 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 601 may include a high-speed random access memory and may also include a non-volatile memory.

[0123] An embodiment of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to execute the above-mentioned arc discharge detection method.

[0124] The technical solution of the embodiment of the present disclosure can be embodied in the form of a software product, which is stored in a storage medium and includes one or more instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiment of the present disclosure. The aforementioned storage medium may be a non-transient storage medium, such as: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes.

[0125] The above description and the accompanying drawings fully illustrate the embodiments of the present disclosure so that those skilled in the art can practice them. Other embodiments may include structural, logical, electrical, process and other changes. The embodiments represent only possible changes. Unless explicitly required, separate components and functions are optional, and the order of operation may vary. The parts and features of some embodiments may be included in or replace the parts and features of other embodiments. Moreover, the words used in this application are only used to describe the embodiments and are not used to limit the claims. As used in the description of the embodiments and the claims, unless the context clearly indicates, the singular forms of "a", "an" and "the" are intended to include plural forms as well. Similarly, the term "and / or" as used in this application refers to any and all possible combinations of listings containing one or more associated ones. In addition, when used in the present application, the term "comprise" and its variants "comprises" and / or comprising refer to the presence of stated features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the presence of other identical elements in the process, method or device comprising the elements. In this article, each embodiment may focus on the differences from other embodiments, and the same and similar parts between the various embodiments may refer to each other. For the methods, products, etc. disclosed in the embodiments, if they correspond to the method part disclosed in the embodiments, then the relevant parts can refer to the description of the method part.

[0126] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software may depend on the specific application and design constraints of the technical solution. The technicians may use different methods for each specific application to implement the described functions, but such implementations should not be considered to exceed the scope of the embodiments of the present disclosure. The technicians may clearly understand that, for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above may refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0127] In the embodiments disclosed herein, the disclosed methods and products (including but not limited to devices, equipment, etc.) can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units can be only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to implement this embodiment. In addition, each functional unit in the embodiment of the present disclosure may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0128] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

Claims

1. A method for detecting arc discharge, characterized in that: include: Obtain visible light images, infrared images and environmental parameters of the area to be detected; Determine, according to the environmental parameters, a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image; Obtaining fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient; The fused features and environmental parameters are input into the arc discharge detection model to obtain the detection results.

2. The detection method according to claim 1, characterized in that: Obtaining fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient, including: decomposing the visible light image and the infrared image respectively to obtain first components of the visible light image at different frequency levels and second components of the infrared image at different frequency levels; fusing the first component and the second component at the same frequency level according to the first weight coefficient and the second weight coefficient to obtain a fused component; reconstructing the image according to the fused component to obtain a fused image; extracting features from the fused image to obtain fusion features; and / or, Obtaining fusion features of the visible light image and the infrared image according to the first weight coefficient and the second weight coefficient, including: inputting the first weight coefficient, the second weight coefficient, the visible light image and the infrared image into a feature extraction model to obtain the fusion features; and / or, According to the first weight coefficient and the second weight coefficient, fusion features of the visible light image and the infrared image are obtained, including: extracting features from the visible light image to obtain first image features; extracting features from the infrared image to obtain second image features; and fusing the first image features and the second image features according to the first weight coefficient and the second weight coefficient to obtain fusion features.

3. The detection method according to claim 1 or 2, characterized in that: The detection results include arc discharge and normal state. The fusion features and environmental parameters are input into the arc discharge detection model to obtain the detection results, including: Inputting the fused features and the environmental parameters into an arc discharge detection model to obtain a first prediction probability that the detection result is an arc discharge and a second prediction probability that the detection result is a normal state; The Bayesian theorem is used to calculate the first a posteriori probability corresponding to the arc discharge and the second a posteriori probability corresponding to the normal state respectively; Calculate the weighted average of the first predicted probability and the first posterior probability to obtain the first probability; and calculate the weighted average of the second predicted probability and the second posterior probability to obtain the second probability; When the first probability is greater than the second probability, the detection result is determined to be an arc discharge; otherwise, the detection result is determined to be a normal state.

4. The detection method according to claim 1 or 2, characterized in that: The fusion features and environmental parameters are input into the arc discharge detection model to obtain the detection results, including: Obtaining power consumption parameters of electrical equipment in the area to be detected; The power consumption parameters, fusion characteristics and environmental parameters are input into the arc discharge detection model to obtain the detection results.

5. The detection method according to claim 1 or 2, characterized in that: The arc discharge detection model is obtained as follows: Obtaining a training data set; the training data set is a data set including a visible light image stream, an infrared image and environmental parameters containing arc discharge marks, and a visible light image, an infrared image and environmental parameters containing normal state marks; The initial model is trained using a training data set to obtain a trained initial model; the initial model includes a recursive neural network or a long short-term memory network; The trained initial model is used as the arc discharge detection model.

6. The detection method according to claim 5, characterized in that: The initial model is trained using the training data set, including: Performing image preprocessing on visible light images and infrared images; image preprocessing includes one or more of noise reduction, contrast enhancement, defogging, color correction and edge enhancement; Performing parameter preprocessing on environmental parameters; parameter preprocessing includes filtering and / or denoising; The initial model is trained using visible light images and infrared images after image preprocessing, and environmental parameters after parameter preprocessing.

7. The detection method according to claim 5, characterized in that: The initial model is trained using the training data set, including: Determine a first parameter from environmental parameters according to time information of the visible light image and / or the infrared image; Combining the visible light image, the infrared image and the first parameter to obtain a training time series matrix; The initial model is trained using the training timing matrix.

8. An arc discharge detection device, characterized in that: include: An acquisition module is configured to acquire visible light images, infrared images and environmental parameters of the area to be detected; A determination module is configured to determine a first weight coefficient corresponding to the visible light image and a second weight coefficient corresponding to the infrared image according to the environmental parameters; A fusion module is configured to obtain a fusion feature of the visible light image and the infrared image according to a first weight coefficient and a second weight coefficient; The detection module is configured to input the fusion features and environmental parameters into the arc discharge detection model to obtain the detection result.

9. An arc discharge detection device, comprising a processor and a memory storing program instructions, characterized in that: The processor is configured to execute the arc discharge detection method according to any one of claims 1 to 7 when running the program instructions.

10. An electronic device, characterized in that: include: Equipment body; The arc discharge detection device according to claim 8 or 9 is installed on the equipment body.

Images