An intelligent monitoring method and device for preventing electrical fire of charging pile

Abstract

The application provides an intelligent monitoring method and device for preventing electric fire of charging piles, wherein the method comprises the following steps: when a picture shot by a first camera has firelight, acquiring first video data, acquiring maximum value information of the fire, extracting position information and pixel information of the fire according to the maximum value, making subsequent feature information obtained according to the position information and the pixel information, and screening target shot video through the feature information. The application has the beneficial effect that the obtained shot video has relatively complete fire shot pictures, and relevant personnel can fully observe the fire.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence, in particular to an intelligent monitoring method and device for preventing electrical fire of charging pile. BACKGROUND

[0002] The charging pile is a kind of fixed on the ground or wall, installed in public buildings and residential parking lots or charging stations, which can charge various types of electric vehicles according to different voltage levels. However, with the development of science and technology, the existing charging method is generally to charge the vehicle by high power, which will increase the frequency of electrical fire of charging pile.

[0003] In the prior art, the identification of electrical fire of charging pile is mainly through camera recognition. However, the location of charging pile fire is uncertain. If the camera installed for fire shooting is not complete, the camera may have limitations and cannot completely capture the possible fire scene, which may lead to insufficient identification of the fire situation. SUMMARY

[0004] The main purpose of the present application is to provide an intelligent monitoring method and device for preventing electrical fire of charging pile, which aims to solve the problem of insufficient identification of fire situation.

[0005] The present application provides an intelligent monitoring method for preventing electrical fire of charging pile, which is realized through a local area network. The local area network includes a plurality of cameras for shooting charging pile area and a master control center. Each camera is connected with the master control center and is used to upload the shooting video to the master control center. The intelligent monitoring method includes:

[0006] When it is detected that the first camera shoots the charging pile area with firelight, the charging pile area is continuously shot to obtain a first shooting video;

[0007] Obtain the height information of the flame in each frame of the first shooting video, and obtain the maximum value point in the flame height information;

[0008] Calculate the distance between two adjacent maximum value points in each frame of picture, and sum up to obtain the target distance;

[0009] Based on the position of each maximum value point in each frame of picture, find the maximum value point closest to the position of the maximum value point in the last frame of picture, and mark it as the last maximum value point of the maximum value point;

[0010] Calculate the distance and direction of each maximum value point and the corresponding last maximum value point to obtain the target vector; the direction is the opposite direction of the direction from the maximum value point to the corresponding last maximum value point;

[0011] extract a same target vector in each frame of the first shooting video to form a target vector set, wherein a last maximum point of a latter target vector and a maximum point of a former target vector in two adjacent target vectors in the target vector set are a same point;

[0012] calculate a modulus of a vector difference between two adjacent target vectors in the target vector set;

[0013] calculate an average value of all moduli in two adjacent frames of pictures and set the average value as a weight value in a former frame of picture;

[0014] multiply a corresponding target vector in each frame with a corresponding weight value to obtain a vector set corresponding to each frame of picture;

[0015] obtain pixel information of each frame of picture in the first shooting video and set a weight value of the pixel information based on the target distance to obtain target pixel information;

[0016] input the target pixel information and the vector set into a preset double-flow network to obtain feature information of the first shooting video;

[0017] screen a target shooting video from shooting videos shot by each camera according to the feature information to monitor the charging pile area.

[0018] Further, before the step of screening a target shooting video from shooting videos shot by each camera according to the feature information to monitor the charging pile area, the method further comprises:

[0019] obtain a first time point at which the fire light appears in the first shooting video;

[0020] obtain a second time point after a preset time interval based on the first time point;

[0021] obtain a target time period according to the first time point and the second time point;

[0022] cut a shooting video from the videos shot by each camera based on the target time period.

[0023] Further, before the step of detecting that the first camera shoots the charging pile area with fire light and continuously shooting the charging pile area to obtain the first shooting video, the method further comprises:

[0024] shoot the charging pile area from multiple preset angles to obtain multiple pictures of the charging pile area;

[0025] extract multiple first features in the multiple pictures of the charging pile area by a preset feature extraction model to form a feature set;

[0026] acquire a shooting picture of a camera in a region with the charging pile as a center and a set distance as a radius through a preset wireless device;

[0027] extract second features in the shooting picture through the preset feature extraction model;

[0028] determine whether the first features and the second features in the feature set are the same features;

[0029] group the cameras with the same features through the master control center to form the local area network.

[0030] Further, the step of screening a target shooting video from shooting videos captured by each camera according to the feature information comprises:

[0031] extracting a preset number of video frames from each shooting shooting video;

[0032] extracting third features in the video frames through a preset feature extraction module;

[0033] calculating the similarity between the third features and the feature information through a preset similarity measurement algorithm;

[0034] screening the target shooting video according to the similarity.

[0035] Further, the step of calculating the distance and direction of each maximum value point and a corresponding previous maximum value point further comprises:

[0036] calculating the distance of each maximum value and a previous maximum value in adjacent two frames of pictures, and summing up the distances to obtain a total movement amount;

[0037] determining whether the total movement amount is greater than a preset movement amount;

[0038] if the total movement amount is greater than the preset movement amount, an alarm signal is sent.

[0039] The application further provides an intelligent monitoring device for preventing electrical fires of charging piles, which is realized through a local area network, the local area network comprises a plurality of cameras for shooting charging pile regions and a master control center, each camera is connected with the master control center and is used for uploading shooting videos to the master control center, and the intelligent monitoring device comprises:

[0040] a shooting module, which is used for continuously shooting a charging pile region when it is detected that a first camera shoots the charging pile region with firelight, and obtaining a first shooting video;

[0041] a first acquisition module, which is used for acquiring height information of a flame in each frame of picture in the first shooting video, and acquiring maximum value points in the flame height information;

[0042] The first calculation module is configured to calculate distances between two adjacent maximum value points in each frame of picture and sum the distances to obtain a target distance;

[0043] The searching module is configured to find a maximum value point closest to a position of each maximum value point in a previous frame of picture based on the position of each maximum value point in each frame of picture, and record the maximum value point as a previous maximum value point of the maximum value point;

[0044] The second calculation module is configured to calculate distances and directions between each maximum value point and a corresponding previous maximum value point to obtain a target vector; the direction is a reverse direction of a direction from the maximum value point to the corresponding previous maximum value point;

[0045] The extraction module is configured to extract a same target vector in each frame of the first shooting video to form a target vector set; in the target vector set, a previous maximum value point of a latter target vector and a maximum value point of a former target vector are a same point.

[0046] The third calculation module is configured to calculate a modulus of a vector difference between two adjacent target vectors in the target vector set.

[0047] The fourth calculation module is configured to calculate an average value of all moduli in two adjacent frames of picture, and set the average value as a weight value in a previous frame of picture.

[0048] The fifth calculation module is configured to multiply a corresponding target vector of each frame and a corresponding weight value to obtain a vector set corresponding to each frame of picture.

[0049] The second acquisition module is configured to acquire pixel information of each frame of picture in the first shooting video, and set a weight value of the pixel information based on the target distance to obtain target pixel information.

[0050] The input module is configured to input the target pixel information and the vector set into a preset double-flow network to obtain feature information of the first shooting video.

[0051] The screening module is configured to screen a target shooting video from shooting videos shot by each camera according to the feature information, to monitor the charging pile area.

[0052] Further, the intelligent monitoring device comprises:

[0053] The third acquisition module is configured to acquire a first time point at which the first shooting video appears the firelight.

[0054] The fourth acquisition module is configured to acquire a second time point after a preset time interval based on the first time point.

[0055] an acquisition module configured to obtain a target time period according to the first time point and the second time point;

[0056] a clipping module configured to clip a shooting video from the video shot by each camera based on the target time period.

[0057] Further, the intelligent monitoring device further comprises:

[0058] a charging pile area shooting module configured to shoot the charging pile area from a plurality of preset angles to obtain a plurality of charging pile area pictures;

[0059] a first feature extraction module configured to extract a plurality of first features in the plurality of charging pile area pictures by a preset feature extraction model to form a feature set;

[0060] a setting module configured to obtain a shooting picture of a camera in a region with the charging pile as a center and a preset distance as a radius by a preset wireless device;

[0061] a second feature extraction module configured to extract a second feature in the shooting picture by the preset feature extraction model;

[0062] a first feature judgment module configured to judge whether the first feature and the second feature in the feature set are the same feature;

[0063] a networking module configured to network the cameras with the same feature by a master control center to form the local area network.

[0064] Further, the screening module comprises:

[0065] an extraction sub-module configured to extract a preset number of video frames from each shooting video;

[0066] an extraction sub-module configured to extract a third feature in the video frame by a preset feature extraction model;

[0067] a similarity calculation sub-module configured to calculate the similarity between the third feature and the feature information by a preset similarity measurement algorithm;

[0068] a screening sub-module configured to screen the target shooting video according to the size of the similarity.

[0069] Further, the intelligent monitoring device further comprises:

[0070] a distance calculation module configured to calculate the distance between each maximum value and the previous maximum value in two adjacent pictures and sum up the distances to obtain a total movement amount;

[0071] The mobile total amount judgment module is configured to judge whether the mobile total amount is greater than a preset mobile amount.

[0072] The alarm signal sending module is configured to send an alarm signal if the mobile total amount is greater than the preset mobile amount.

[0073] The present application has the beneficial effects that when the picture captured by the first camera has a firelight, the first video data is acquired, the maximum value information of the fire is acquired, the position information and the pixel information of the fire are extracted according to the maximum value, the feature information is obtained according to the position information and the pixel information, the target shooting video is screened through the feature information, so that the shooting video obtained can have a full fire shooting picture, which is convenient for being used as a basis material for tracing the cause of the fire and is convenient for relevant personnel to fully observe the fire. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 is a flowchart of an intelligent monitoring method for preventing electrical fire of a charging pile according to an embodiment of the present application;

[0075] Figure 2 is a structural schematic block diagram of an intelligent monitoring device for preventing electrical fire of a charging pile according to an embodiment of the present application;

[0076] Figure 3 is a structural schematic block diagram of a computer device according to an embodiment of the present application.

[0077] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0078] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0079] It should be noted that all directionality indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative position relationship, movement condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directionality indications also change accordingly. The connection can be direct connection or indirect connection.

[0080] The term "and / or" in this document is only used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and B can represent three cases of existence of A alone, existence of A and B simultaneously, and existence of B alone.

[0081] In addition, the description such as "first", "second" and the like in the present application is only for the purpose of description and cannot be understood as indicating or implying the relative importance of the technical features indicated or implying the number of technical features indicated. Therefore, the features defined as "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the technical solutions of various embodiments can be combined with each other, but it must be based on the realization of ordinary skilled in the art, when the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, also not within the protection scope required by the present application.

[0082] Referring to Figure 1 , the present application proposes an intelligent monitoring method for preventing electrical fire of charging pile, which is realized through local area network, the local area network includes multiple cameras shooting charging pile area and master control center, each camera is connected with the master control center respectively, and is used for uploading shooting video to the master control center, the intelligent monitoring method includes:

[0083] S1: when detecting that the first camera shoots the charging pile area with fire light, continuously shooting the charging pile area, and obtaining the first shooting video;

[0084] S2: obtaining the height information of flame in each frame of the first shooting video, and obtaining the maximum value point in the flame height information;

[0085] S3: calculating the distance between adjacent two maximum value points in each frame of picture, and summing to obtain target distance;

[0086] S4: based on the position of each maximum value point in each frame of picture, finding out the maximum value point closest to the position of the maximum value point in the last frame of picture, and recording as the last maximum value point of the maximum value point;

[0087] S5: calculating the distance and direction of each maximum value point and corresponding last maximum value point, and obtaining target vector; the direction is the opposite direction of the direction from the maximum value point to the corresponding last maximum value point;

[0088] S6: extracting the same target vector in each frame of the first shooting video, and constituting target vector set; wherein, in adjacent two target vectors of the target vector set, the last maximum value point of the latter target vector and the maximum value point of the former vector are the same point;

[0089] S7: calculating the modulus of the vector difference of adjacent two target vectors in the target vector set;

[0090] S8: calculating the average value of all moduli in adjacent two frames of pictures, and setting it as the weight value in the previous frame of picture;

[0091] S9: multiplying the target vector corresponding to each frame with the corresponding weight value to obtain a vector set corresponding to each frame of picture;

[0092] S10: obtaining pixel information of each frame of picture in the first shooting video, and setting a weight value of the pixel information based on the target distance to obtain target pixel information;

[0093] S11: inputting the target pixel information and the vector set into a preset double-flow network to obtain feature information of the first shooting video;

[0094] S12: screening a target shooting video from the shooting videos shot by each camera according to the feature information, to monitor the charging pile area.

[0095] As described in step S1, when it is detected that the first camera shoots the charging pile area with firelight, the charging pile area is continuously shot to obtain the first shooting video. Specifically, the first camera can be any camera in the local area network. It should be noted that the local area network includes a plurality of cameras, each camera shoots the camera from a different angle. The camera can be designed when the charging pile is installed, or it can be a camera for monitoring. It can shoot the charging pile area. The charging pile area is a pre-set area, which can include the charging pile and the area where the electric vehicle is parked. After shooting the video, image processing and analysis techniques can be used to detect fire, smoke or other fire signs. These techniques can identify fire signs by analyzing pixels, colors, dynamic changes, etc. in the video, and then continuously shoot them to obtain the first shooting video.

[0096] As described in step S2, the height information of the flame in each frame of the first shooting video is obtained, and the maximum value point in the flame height information is obtained. That is, the first shooting video is divided into single pictures, and then the image processing technique is used to extract the position and height information of the flame. In each frame of picture, the flame can be separated from the surrounding environment by threshold segmentation, and then the height of the flame can be calculated. The position difference of the upper and lower edges of the flame can be measured, and then the pixel unit is regarded as the height unit, and the maximum value point is found.

[0097] As described in step S3, the distance between adjacent two maximum value points in each frame of picture is calculated, and the sum is obtained to obtain the target distance. First, the positions of the maximum value points are extracted from the flame height information obtained before. These positions represent the frames where the flame height reaches the maximum value. For each maximum value point, the distance between it and the adjacent maximum value point is calculated, and the sum of the distances between all adjacent maximum value points is calculated to obtain the target distance.

[0098] As described in step S4 above, based on the position of each maximum point in each frame of picture, find the maximum point closest to the position of the maximum point in the previous frame of picture, and record it as the previous maximum point of the maximum point, specifically, traverse each maximum point in each frame. Take the maximum point of the first frame as the starting point. For each maximum point, find the closest maximum point in the previous frame. The Euclidean distance or other distance measure between the maximum points can be calculated to determine the nearest point, i.e. the previous maximum point.

[0099] As described in step S5 above, calculate the distance and direction of each of the maximum points from the corresponding previous maximum point to obtain the target vector; the direction is the opposite direction of the direction from the maximum point to the corresponding previous maximum point. First, calculate the vector from the previous maximum point to the current maximum point, and then take its opposite direction. The direction of the target vector can be obtained by calculating the difference vector between the two points and then taking the opposite direction. Combine the distance and direction to form the target vector.

[0100] As described in step S6 above, extract the same target vector in each frame of the first shot video to form a target vector set; wherein the previous maximum point of the next target vector and the maximum point of the previous vector in the adjacent two target vectors in the target vector set are the same point, specifically, for each frame in the first shot video, extract the target vector. Store the target vector of each frame into the target vector set. Traverse each target vector in the target vector set. Check the next target vector of each target vector, i.e. the target vector in the next frame. Ensure that the previous maximum point of the next target vector and the maximum point of the previous vector are the same point. Thus, the same target vector in each frame of the first shot video is extracted to form a target vector set. At the same time, such screening process can help reduce errors and ensure the continuity and stability of the target vector.

[0101] As described in step S7 above, calculate the modulus of the vector difference between the adjacent two target vectors in the target vector set. For each pair of adjacent target vectors in the target vector set, calculate the vector difference between them. The vector difference can be obtained by subtracting the previous target vector from the next target vector. Calculate the modulus of the vector difference. The modulus of the vector difference refers to the length or size of the vector difference, which can be obtained by calculating the Euclidean distance of the vector difference. Repeat the above steps until the modulus of the vector difference between all adjacent two target vectors in the target vector set is calculated.

[0102] As described in step S8 above, the average of all the norms in the two adjacent frames of pictures is calculated and set as the weight value in the previous frame of picture. For each pair of adjacent target vectors in the target vector set, the norm of the vector difference between them is calculated and the values of these norms are stored. The average of all the norms in the two adjacent frames of pictures is calculated. The values of all the norms are added up and then divided by the number of norms to obtain the average. The calculated average is used as the weight value in the previous frame of picture.

[0103] As described in step S9 above, the target vector corresponding to each frame is multiplied by the corresponding weight value to obtain the vector set corresponding to each frame of picture. The target vector and the weight value corresponding to each frame of picture are obtained. For each frame of picture, the target vector and the weight value corresponding to it are multiplied. This can be achieved by multiplying the coordinates in each target vector by the corresponding weight value. The result of the multiplication is taken as the vector set corresponding to each frame of picture.

[0104] As described in step S10 above, the pixel information of each frame of picture in the first captured video is obtained and the weight value of the pixel information is set based on the target distance to obtain target pixel information. The pixel information of each frame of picture is obtained from the first captured video. This can be achieved by reading the video file or obtaining the video stream frame by frame. For each frame of picture, the weight value of the pixel information is set according to the target distance. You can assign a corresponding weight value to each pixel in the pixel information according to the size of the target distance. The weight value is applied to the pixel information to obtain the target pixel information. The adjusted target pixel value can be obtained by multiplying the original value of each pixel by the corresponding weight value.

[0105] As described in step S11, the target pixel information and the vector set are input into a preset dual-stream network to obtain the feature information of the first shooting video. The target pixel information is input into a pixel stream. The target pixel information of each frame of picture is transmitted to the pixel stream, so that the network can extract features from the image. The vector set is input into a vector stream. The vector set of each frame of picture is input into the vector stream, so that the network can extract features from the vector. The outputs of the pixel stream and the vector stream are connected. The output features of the two streams are connected to form a complete feature representation. The connected features are input into a classifier or other task-related module of the network. A common dual-stream network is a combination of an optical flow network and a convolutional neural network (CNN). The optical flow network is used to process the optical flow field, that is, the motion information of pixels between consecutive frames, and the CNN is used to extract the static features of the image. Such a dual-stream network can produce better performance in video action recognition, optical flow estimation and other tasks. In this application, the vector set is used as the motion information of the fire to input into the optical flow network as an alternative.

[0106] As described in step S12, the target shooting video is selected from the shooting videos shot by each camera according to the feature information, so as to monitor the charging pile area. The feature information is used to determine whether other video frames have the same feature information, so as to select the target shooting video. The selected video is displayed, so that the obtained shooting video has a relatively complete fire shooting picture, and relevant personnel can fully observe the fire.

[0107] In one embodiment, before the step S12 of selecting the target shooting video from the shooting videos shot by each camera according to the feature information to monitor the charging pile area, the method further comprises:

[0108] S1101: obtaining a first time point at which the first shooting video appears the firelight;

[0109] S1102: obtaining a second time point after a preset time interval based on the first time point;

[0110] S1103: obtaining a target time period according to the first time point and the second time point;

[0111] S1104: cutting a shooting video from the videos shot by each camera based on the target time period.

[0112] As described in steps S1101-S1104 above, the acquisition of the shooting video is realized, when the firelight appears, the current time point, i.e. the first time point, is recorded, the second time point after the preset time interval based on the first time point is acquired, the preset time interval is a pre-set time interval point, and the present application does not limit this, then the target time period can be acquired according to the first time point and the second time point, and the shooting video of each camera is cut according to the target time period, so as to obtain the shooting video.

[0113] In one embodiment, before the step S1 of continuously shooting the charging pile area to obtain the first shooting video when it is detected that the first camera shoots the charging pile area with firelight, it further includes:

[0114] S001: shooting the charging pile area from multiple preset angles to obtain multiple charging pile area pictures;

[0115] S002: extracting multiple first features in the multiple charging pile area pictures by a preset feature extraction model to form a feature set;

[0116] S003: acquiring shooting pictures of cameras in an area with the charging pile as the origin and a set distance as the radius by a preset wireless device;

[0117] S004: extracting second features in the shooting pictures by the preset feature extraction model;

[0118] S005: judging whether the first feature and the second feature in the feature set are the same feature;

[0119] S006: networking the cameras with the same feature by the master control center to form the local area network.

[0120] As described in steps S001-S002 above, the charging pile area is shot from multiple preset angles to obtain multiple charging pile area pictures, and multiple first features in the multiple charging pile area pictures are extracted by a preset feature extraction model to form a feature set; wherein the shooting mode can be that a related person shoots by moving the camera, for example, shooting by a mobile phone, and then multiple first features in the multiple charging pile area pictures are extracted by a preset feature extraction model, which can be a pre-trained convolutional neural network model to extract the first features.

[0121] The step S003-S006 described above is obtained by the preset wireless device to obtain the camera shooting pictures in the area with the charging pile as the origin and the set distance as the radius; the second feature in the shooting picture is extracted by the preset feature extraction model; it is judged whether the first feature and the second feature in the feature set are the same feature; the cameras with the same feature are networked by the master control center to form the local area network. The pictures taken by the cameras in the area near the charging pile are obtained by the preset wireless device. This step involves using a preset and configured wireless device, placing it near the charging pile, and then obtaining the pictures taken by the cameras in the nearby area. The second feature in the shooting picture is extracted using the preset feature extraction model. The preset feature extraction model (which may be the same model as the first feature extraction model) is imported, and the pictures taken by the cameras are passed to the model to extract the second feature. It is judged whether the first feature and the second feature in the feature set are the same feature. The first feature and the second feature are compared to determine whether they are the same. The cameras with the same feature are networked by the master control center to form the local area network. A master control center (which can be an intelligent center or a server) is used to network the cameras with the same feature to build a local area network that can share data and perform various tasks.

[0122] In one embodiment, the step S12 of screening the target shooting video from the shooting videos taken by each camera according to the feature information comprises:

[0123] S1201: Extract a preset number of video frames from each shooting video;

[0124] S1202: Extract the third feature in the video frame by the preset feature extraction module;

[0125] S1203: Calculate the similarity between the third feature and the feature information by the preset similarity measurement algorithm;

[0126] S1204: Screen the target shooting video according to the size of the similarity.

[0127] As described in steps S1201-S1204, the selection of the target shooting video is realized. Specifically, a preset number of video frames are extracted from each shot shooting video, and the extraction can be random extraction or extraction according to a preset time interval, which is not limited in the present application. The target features in the video frames are extracted by a preset feature extraction module, which is described in detail above, and will not be repeated here. The similarity between the third features and the feature information is calculated by a preset similarity measurement algorithm. It should be noted that since a second video data has multiple third features, the present application only takes the similarity value with the largest similarity as the similarity. The preset similarity measurement algorithm can be any one of Euclidean distance, cosine similarity, Hamming distance, edit distance, etc.

[0128] In one embodiment, after the step S5 of calculating the distance and direction of each maximum point from the corresponding previous maximum point, the method further comprises:

[0129] S601: Calculate the distance of each maximum value from the previous maximum value in the adjacent two frames of pictures, and sum up the distances to obtain the total movement amount;

[0130] S602: Determine whether the total movement amount is greater than a preset movement amount;

[0131] S603: If the total movement amount is greater than the preset movement amount, an alarm signal is sent.

[0132] As described in steps S601-S603, in the present embodiment, since the spread of the fire is large, it may not wait for the relevant personnel to handle, or the handling capacity of the relevant personnel is limited, therefore, the total movement amount of the maximum value can be used for judgment, and the spread of the fire can be determined according to the total movement amount. When the total movement amount is greater than the preset displacement amount, it can be determined that it is uncontrollable, and an alarm signal can be sent. The alarm signal can include the position information of the charging pile area, so that the firefighters can know the position information in time. If the total movement amount is less than or equal to the preset displacement amount, the fire is still controllable, and can be uploaded to the relevant personnel for timely fire extinguishing or alarm, which is not limited in the present application and can be determined by the relevant personnel.

[0133] The present application has the following advantages: when the picture captured by the first camera has firelight, the first video data is obtained, the maximum value information of the flame is obtained, the position information and the pixel information of the flame are extracted from the maximum value, the feature information is obtained according to the position information and the pixel information, and the target shooting video is screened through the feature information, so that the obtained shooting video has a relatively complete fire shooting picture, and the relevant personnel can observe the fire sufficiently.

[0134] Reference Figure 2The application further provides an intelligent monitoring device for preventing electric fire of a charging pile, which is realized through a local area network, wherein the local area network comprises a plurality of cameras for shooting a charging pile area and a master control center, each camera is connected with the master control center and is used for uploading a shooting video to the master control center, and the intelligent monitoring device comprises:

[0135] A shooting module 10 is configured to continuously shoot the charging pile area when it is detected that the first camera shoots the charging pile area with firelight, so as to obtain a first shooting video;

[0136] A first acquisition module 20 is configured to acquire height information of a flame in each frame of the first shooting video and acquire a maximum point in the flame height information;

[0137] A first calculation module 30 is configured to calculate a distance between two adjacent maximum points in each frame of picture and sum up the distance to obtain a target distance;

[0138] A finding module 40 is configured to find a maximum point closest to a position of each maximum point in a previous frame of picture based on the position of each maximum point in each frame of picture, and record the maximum point as a previous maximum point of the maximum point;

[0139] A second calculation module 50 is configured to calculate a distance and a direction of each maximum point and a corresponding previous maximum point to obtain a target vector, wherein the direction is a reverse direction of a direction from the maximum point to the corresponding previous maximum point;

[0140] An extraction module 60 is configured to extract a same target vector in each frame of the first shooting video to form a target vector set, wherein a previous maximum point of a latter target vector and a maximum point of a former vector in two adjacent target vectors in the target vector set are a same point;

[0141] A third calculation module 70 is configured to calculate a modulus of a vector difference between two adjacent target vectors in the target vector set;

[0142] A fourth calculation module 80 is configured to calculate an average value of all moduli in two adjacent frames of picture and set the average value as a weight value in a previous frame of picture;

[0143] A fifth calculation module 90 is configured to multiply a corresponding target vector of each frame and a corresponding weight value to obtain a vector set corresponding to each frame of picture;

[0144] A second acquisition module 100 is configured to acquire pixel information of each frame of picture in the first shooting video and set a weight value of the pixel information based on the target distance to obtain target pixel information;

[0145] The input module 110 is configured to input the target pixel information and the vector set into a preset double-flow network to obtain feature information of the first shooting video.

[0146] The screening module 120 is configured to screen a target shooting video from shooting videos shot by each camera according to the feature information, to monitor the charging pile area.

[0147] In one embodiment, the intelligent monitoring device comprises:

[0148] The third acquisition module is configured to acquire a first time point at which the first shooting video appears the firelight;

[0149] The fourth acquisition module is configured to acquire a second time point after a preset time interval based on the first time point;

[0150] The acquisition module is configured to obtain a target time period according to the first time point and the second time point;

[0151] The interception module is configured to intercept a shooting video from videos shot by each camera based on the target time period.

[0152] In one embodiment, the intelligent monitoring device further comprises:

[0153] The charging pile area shooting module is configured to shoot the charging pile area from multiple preset angles to obtain multiple charging pile area pictures;

[0154] The first feature extraction module is configured to extract multiple first features in the multiple charging pile area pictures by a preset feature extraction model to form a feature set;

[0155] The setting module is configured to acquire, by a preset wireless device, a shooting picture of a camera in an area with the charging pile as a center and a preset distance as a radius;

[0156] The second feature extraction module is configured to extract second features in the shooting picture by the preset feature extraction model;

[0157] The first feature judgment module is configured to judge whether the first features and the second features in the feature set are the same features;

[0158] The networking module is configured to network cameras with the same features by a master control center to form the local area network.

[0159] In one embodiment, the screening module 120 comprises:

[0160] The extraction sub-module is configured to extract a preset number of video frames from each shot shooting video;

[0161] extracting a third feature in the video frame by a preset feature extraction module;

[0162] calculating a similarity between the third feature and the feature information by a preset similarity measurement algorithm;

[0163] screening the target shooting video according to the size of the similarity.

[0164] In one embodiment, the intelligent monitoring device further comprises:

[0165] a distance calculation module configured to calculate a distance between each maximum value and a previous maximum value in adjacent two frames of pictures, and sum up the distances to obtain a total movement amount;

[0166] a total movement amount judgment module configured to judge whether the total movement amount is greater than a preset movement amount;

[0167] an alarm signal sending module configured to send an alarm signal if the total movement amount is greater than the preset movement amount.

[0168] Reference Figure 3 The embodiments of the present application also provide a computer device which can be a server, and the internal structure of the computer device can be as shown in Figure 3 The computer device comprises a processor, a memory, a network interface and a database connected through a system bus. The processor of the computer device is configured to provide calculation and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store various shooting videos. The network interface of the computer device is configured to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement the intelligent monitoring method for preventing electrical fire of a charging pile according to any one of the above embodiments.

[0169] Those skilled in the art can understand that Figure 3 The structure shown in the above embodiments is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied.

[0170] The embodiments of the present application also provide a computer readable storage medium having a computer program stored thereon. The computer program is executed by the processor to implement the intelligent monitoring method for preventing electrical fire of a charging pile according to any one of the above embodiments.

[0171] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments can be included. Any reference to memory, storage, databases, or other media in this application and in examples provided herein can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0172] It should be noted that in this paper, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, device, article or method including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, device, article or method. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0173] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. Among them, artificial intelligence (Artificial Intelligence, AI) is to use digital computers or digital computer controlled machines to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. Theory, method, technology and application system.

[0174] The basic technology of artificial intelligence generally includes technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics, etc. Artificial intelligence software technology mainly includes computer vision technology, robot technology, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning, etc. Several major directions.

[0175] The above merely provides the preferred embodiments of the present application, and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the principles and technical scope of the present application shall fall into the scope of claims of the present application.

Claims

1. A method for intelligent monitoring to prevent electrical fire in charging piles, characterized in that, The method is implemented through a local area network including a plurality of cameras for shooting charging pile areas and a master control center, each camera is connected with the master control center and is used for uploading shooting videos to the master control center, and the intelligent monitoring method includes the following steps: When it is detected that the first camera shoots a charging pile area with a fire, the charging pile area is continuously shot to obtain a first shooting video; Height information of the flame in each frame of the first shooting video is obtained, and a maximum value point in the flame height information is obtained; A distance between adjacent two maximum value points in each frame of picture is calculated, and a target distance is obtained by summation; Based on the position of each maximum value point in each frame of picture, a maximum value point closest to the position of the maximum value point in the last frame of picture is found out and recorded as a last maximum value point of the maximum value point; A distance and a direction of each maximum value point and a corresponding last maximum value point are calculated to obtain a target vector, and the direction is a reverse direction of the direction from the maximum value point to the corresponding last maximum value point; The same target vector in each frame of the first shooting video is extracted to form a target vector set, wherein in adjacent two target vectors of the target vector set, a last maximum value point of a latter target vector and a maximum value point of a former vector are the same point; A module of a vector difference between adjacent two target vectors in the target vector set is calculated; An average value of all modules in adjacent two frames of pictures is calculated, and the average value is set as a weight value in a previous frame of picture; A corresponding target vector of each frame is multiplied by a corresponding weight value to obtain a vector set corresponding to each frame of picture; Pixel information of each frame of picture in the first shooting video is obtained, and a weight value of the pixel information is set based on the target distance to obtain target pixel information; The target pixel information and the vector set are input into a preset double-flow network to obtain feature information of the first shooting video; According to the feature information, a target shooting video is screened out from shooting videos shot by each camera to monitor the charging pile area.

2. The intelligent monitoring method for preventing electrical fire of charging pile according to claim 1, wherein, Before the step of screening out the target shooting video from the shooting videos shot by each camera according to the feature information to monitor the charging pile area, the following steps are included: A first time point when the first shooting video appears the fire is obtained; A second time point after a preset time interval based on the first time point is obtained; A target time period is obtained according to the first time point and the second time point; A shooting video is intercepted from the videos shot by each camera based on the target time period.

3. The intelligent monitoring method for preventing electrical fire of charging pile according to claim 1, wherein, The step of screening out the target shooting video from the shooting videos shot by each camera according to the feature information includes the following steps: A preset number of video frames are extracted from each shot shooting video; A third feature in the video frame is extracted through a preset feature extraction module; Similarity between the third feature and the feature information is calculated through a preset similarity measurement algorithm; The target shooting video is screened out according to the size of the similarity.

4. The intelligent monitoring method for preventing electrical fire of charging pile according to claim 1, wherein, After the step of calculating the distance and the direction of each maximum value point and a corresponding last maximum value point, the following steps are included: Calculate the distance between each maximum value and the previous maximum value in two adjacent frames of pictures, and sum up the distances to obtain the total movement amount; Determine whether the total movement amount is greater than a preset movement amount; If greater than the preset movement amount, an alarm signal is issued.

5. An intelligent monitoring device for preventing electrical fire of charging piles, characterized in that, The local area network includes a plurality of cameras for shooting charging pile areas and a master control center, each camera is connected with the master control center, and is used for uploading shooting videos to the master control center, and the intelligent monitoring device includes: The shooting module is used for continuously shooting the charging pile area when it is detected that the first camera shoots the charging pile area with firelight, to obtain a first shooting video; The first acquisition module is used for acquiring height information of the flame in each frame of the first shooting video, and acquiring maximum value points in the flame height information; The first calculation module is used for calculating the distance between two adjacent maximum value points in each frame of picture, and summing up to obtain a target distance; The finding module is used for finding the maximum value point closest to the position of each maximum value point in the previous frame of picture based on the position of each maximum value point in each frame of picture, and recording it as the previous maximum value point of the maximum value point; The second calculation module is used for calculating the distance and direction of each maximum value point and the corresponding previous maximum value point to obtain a target vector; the direction is the opposite direction of the direction from the maximum value point to the corresponding previous maximum value point; The extraction module is used for extracting the same target vector in each frame of the first shooting video to form a target vector set; wherein, in adjacent two target vectors of the target vector set, the previous maximum value point of the latter target vector and the maximum value point of the former vector are the same point; The third calculation module is used for calculating the modulus of the vector difference between adjacent two target vectors in the target vector set; The fourth calculation module is used for calculating the average value of all moduli in adjacent two frames of pictures, and setting it as the weight value in the previous frame of picture; The fifth calculation module is used for multiplying each corresponding target vector and the corresponding weight value to obtain a vector set corresponding to each frame of picture; The second acquisition module is used for acquiring pixel information of each frame of picture in the first shooting video, and setting a weight value of the pixel information based on the target distance to obtain target pixel information; The input module is used for inputting the target pixel information and the vector set into a preset double-flow network to obtain feature information of the first shooting video; The screening module is used for screening a target shooting video from the shooting videos shot by each camera according to the feature information, to monitor the charging pile area.

6. The intelligent monitoring device for preventing electrical fire of charging pile according to claim 5, wherein, The intelligent monitoring device includes: The third acquisition module is used for acquiring a first time point when the firelight appears in the first shooting video; The fourth acquisition module is used for acquiring a second time point after a preset time interval based on the first time point; The target time period acquisition module is used for obtaining a target time period according to the first time point and the second time point; The intercepting module is used for intercepting a shooting video from the videos shot by each camera based on the target time period.

7. The intelligent monitoring device for preventing electrical fire of charging pile according to claim 5, wherein, The screening module includes: The extraction sub-module is used for extracting a preset number of video frames from each shot shooting video; The extraction submodule is configured to extract a third feature in the video frame by using a preset feature extraction module; The similarity calculation submodule is configured to calculate a similarity between the third feature and the feature information by using a preset similarity measurement algorithm; The screening submodule is configured to screen the target shooting video according to the size of the similarity.

8. The intelligent monitoring device for preventing electrical fire of charging pile according to claim 5, wherein, The intelligent monitoring device further comprises: The distance calculation module is configured to calculate a distance between each maximum value and a previous maximum value in two adjacent frames of pictures, and sum up the distances to obtain a total movement amount; The total movement amount judgment module is configured to judge whether the total movement amount is greater than a preset movement amount; The alarm signal sending module is configured to send an alarm signal if the total movement amount is greater than the preset movement amount.

Images