WO2021115343A1 - Intelligent fire-fighting system and method for transformer substation - Google Patents

Abstract

An intelligent fire-fighting system and method for a transformer substation. The intelligent fire-fighting system comprises a plurality of fire-fighting robots and fire-fighting medium supply equipment as well as a control center; the fire-fighting robots and the fire-fighting medium supply equipment are all deployed at intervals in each area in the transformer substation; the fire-fighting medium supply equipment can be fixed, and can also be movable; while a fire behavior occurs, ignition point information can be obtained according to a smoke/temperature sensor deployed in each area in the transformer substation, can also be obtained by sensors carried on the fire-fighting robots, and can also be detected by arranging an unmanned aerial vehicle. The control center schedules the fire-fighting robot closest to an ignition point, so that the fire-fighting robot firstly goes to the ignition point to extinguish fire, and meanwhile, other fire-fighting robots closer to the ignition point or the fire-fighting medium supply equipment can be scheduled to go to the ignition point. Timely fire-extinguishing and adaptive fire-extinguishing are implemented.

Description

Intelligent fire-fighting system and method for substation Technical field

The present disclosure belongs to the technical field of smart substations, and specifically relates to a substation smart fire fighting system and method.

Background technique

The statements in this section merely provide background technical information related to the present disclosure, and do not necessarily constitute prior art.

There are a large number of high-voltage and high-current equipment in the substation/converter station. It is easy to cause fires due to equipment failures, line defects, etc., and once a fire occurs, the consequences are very serious. Therefore, the substation/converter station has taken certain fire protection measures. .

According to the inventor, there are mainly two types of fire protection measures in current substations:

(1) Use sensors to detect fire parameters (such as smoke, temperature, etc.), and after determining the fire situation, the on-duty personnel use the fire-fighting facilities stored in various locations in the substation/converter station to perform fire-fighting operations;

(2) The substation/converter station generally puts in a robot loaded with a certain capacity of fire extinguishing medium (such as fire fighting water, etc.).

However, the above-mentioned fire-fighting measures have certain shortcomings. First, although manual rescue can observe the fire situation in real time and help to put out the fire, it has great hidden dangers to personal safety. At the same time, a large number of fire-fighting facilities need to be invested in each substation/converter station. Many fire-fighting facilities (such as fire extinguishers, etc.) need to be replaced regularly. At the same time, there are a large number of high-voltage and high-current equipment, and the demand for fire-fighting media is large, which increases the cost of fire-fighting; The fire scene takes a certain amount of time, and it is impossible to extinguish the fire in the best extinguishing period. A small fire may cause a major fire accident. A large fire accident will have a serious impact on the construction safety and electricity safety of the power system in multiple areas.

At present, small and medium-sized fire-fighting robots carry limited fire extinguishing medium capacity and cannot control the fire at one time, which causes the fire delay or the safety of the robot itself is threatened. Large-scale robots are generally equipped with a large-caliber water cannon, which is supplied by fire trucks, and can only use water jets to extinguish fires. The facilities facing the substation have application limitations. At the same time, the height of the fire water cannon is fixed, which cannot meet the requirements of different height equipment in the substation. Accurate fire extinguishing needs. In addition, if the robot is far from the fire point, it will not be able to reach the fire scene quickly and in time, and it may also delay the fire and cause the fire to expand.

Summary of the invention

In order to solve the above problems, the present disclosure proposes an intelligent fire protection system and method for a substation. The present disclosure can enable multiple fire fighting robots and fire fighting medium supply equipment deployed in the substation to ensure the first Fire fighting at a time to ensure the safety of the substation.

According to some embodiments, the present disclosure adopts the following technical solutions:

An intelligent fire protection system for a substation, comprising a plurality of fire fighting robots and fire fighting medium supply equipment that can be linked and controlled, and a control center. The fire fighting robots and fire fighting medium supply equipment are arranged at intervals in each area of the substation. The fire fighting medium supply equipment It can be detachably connected with the fire hydrant to realize continuous water supply, among which:

The fire-fighting robot includes a robot body on which a fire-fighting water/foam spray mechanism with a first interface is provided; the fire-fighting medium supply equipment carries a water supply mechanism and a foam supply mechanism, and the corresponding fire extinguishing medium is provided through a hose, The end of the hose is provided with a second interface adapted to the first interface;

The control center is configured to receive fire information, and dispatch a corresponding number of fire-fighting robots with a short distance or a corresponding number of fire-fighting robots and fire-fighting medium supply equipment with a short distance according to the fire situation to the vicinity of the fire point, through the first interface and the second The second interface is automatically and quickly connected to implement fire fighting.

As a further limitation, the fire fighting robot includes a robot body, and a fire extinguishing mechanism is provided on the robot body. The fire extinguishing mechanism includes a fire fighting water/foam spraying mechanism, and the fire fighting water/foam spraying mechanism includes at least one water inlet pipe. One end of the water pipe is provided with a first interface, and the other end is connected to a transmission water pipe. The other end of the transmission water pipe is provided with a rotary joint. The rotary joint is provided with a spray nozzle, and the rotary joint is connected with the robot body through a lifting mechanism.

As a further limitation, the fire-fighting medium supply equipment carries a water supply mechanism and a foam supply mechanism, wherein:

The water supply mechanism includes a number of fire-fighting water storage tanks. The fire-fighting water storage tank is connected to a booster pump through a pipeline. The drain pipe of the booster pump is connected to a hose, and the other end of the hose is provided with a first interface. Equipped second interface;

The foam supply mechanism includes a plurality of foam storage tanks, the foam storage tanks are connected to the foam pump through a pipeline, the drain pipe of the foam pump is connected to the hose, and the other end of the hose is provided with a first interface. Adapted second interface.

According to the present disclosure, multiple firefighting robots and firefighting medium supply equipment are set in the substation. The control center can confirm whether to put in several firefighting robots separately according to the size of the fire, or whether to put in firefighting robots + firefighting medium supply equipment, and quickly dispatch nearby according to the location of the fire. , Strive to extinguish the fire in the best extinguishing period, which can effectively control the fire and ensure the safety of the substation.

The present disclosure can use fire fighting robots and fire fighting medium supply equipment to cooperate with each other, combined with the length and flexibility of hoses and transmission pipes, and can make full use of the limited space in the substation, ensure continuous operation, and fully control the fire.

The fire-fighting medium supply equipment can be a movable mechanism such as a car body, or a fixed mechanism.

At the same time, the above technical solutions can achieve fire extinguishing with water and foam, and at least two media. At the same time, it can adapt to the complex environment in the substation and adapt to the different regions and heights of the fire source in the substation by using the mutual cooperation of the mobile chassis, the lifting mechanism and the rotary joint. Sexual adjustment and suppression have great freedom.

As an alternative embodiment, the fire extinguishing mechanism further includes a dry powder spraying mechanism, which is arranged on the mobile chassis of the fire fighting robot, and specifically includes a number of dry powder tanks. The outlet of the dry powder tank is connected to the spray head through a pipeline, and the spray head is arranged On the revolving head, the revolving head is arranged on the mobile chassis through a lifting mechanism, so that the height and angle of the dry powder spray can be adjusted.

As an alternative embodiment, the mobile chassis is a crawler-type mobile chassis.

As an alternative embodiment, a shell is provided on the mobile chassis, and the shell contains the dry powder injection mechanism and the fire-fighting water/foam injection mechanism, and at least the rotary joint, the rotary head and the nozzle are exposed outside the shell. body.

As an alternative embodiment, the front end of the housing is provided with a distance measuring sensor, a camera device, and a lighting lamp.

As an alternative embodiment, the dry powder injection mechanism includes a number of solenoid valves and dry powder tanks, the powder inlet pipe of the solenoid valve is connected to the corresponding dry powder tank, one end of the powder outlet pipe of the solenoid valve is connected to the solenoid valve, and the other end is connected to the multi-way The joint, one end of the multi-way joint is connected to the dry powder nozzle, which is fixed on the rotary head.

As an alternative embodiment, the fire-fighting water/foam spray mechanism includes a main water pipe and a solenoid valve. One end of the main water pipe is connected to the water inlet pipe, and the other end is connected to the spray nozzle. The main water pipe is provided with a solenoid valve, and the fire water is controlled by the solenoid valve. The spraying work of the foam spraying mechanism.

As an alternative embodiment, the fire-fighting robot is also provided with a self-spraying mechanism, which specifically includes a standpipe and a sprinkler. The water inlet pipe is connected to the standpipe through a pipeline, and the angle between the standpipe and the vertical direction is less than or equal to 15 °, the nozzle is installed on the riser.

Self-spraying is the process of self-spraying of the robot during the water spraying process to lower its temperature and ensure the safety of the robot itself. The self-sprinkling is always in the open state. As long as the standpipe has fire fighting water, it will work.

As an alternative embodiment, one end of the lifting mechanism is fixed to the mobile chassis, and the other end is provided with a revolving base, the revolving base is equipped with a revolving head, and the revolving head is provided with a lighting lamp, a camera device and an infrared Thermal Imager.

As an alternative embodiment, the fire-fighting medium supply equipment is provided with a hose docking mechanism, the hose docking mechanism includes a socket provided on the fire-fighting medium supply equipment, the socket includes a support seat, and the support seat is A plurality of uprights are provided, the uprights are distributed circumferentially, the center can accommodate the water inlet pipe, the upright is provided with an elastic body, the end of the upright is provided with a compression plate, and the compression plate is provided with a relatively rotatable The pressing piece is connected to the second interface by a movable card.

When the second interface is docked with the first interface, the second interface is set in the socket to ensure the stability of the second interface. At the same time, when facing outwards, the first interface on the robot side can be quickly connected to ensure timeliness and speed. , The elastic body can also offset the impulse when the two joints are connected to a certain extent.

As an alternative embodiment, the second interface and the first interface are both quick-connect plugs.

As an alternative embodiment, the fire fighting robot body and the fire fighting medium supply equipment are both provided with at least one image acquisition device for real-time monitoring of the docking state of the second interface and the first interface.

As a further limitation, the fire-fighting medium supply equipment is also provided with a hose retracting mechanism, which specifically includes a reel, a driving member, and a support frame. The driving member is connected to the reel through the transmission member to drive the reel to rotate. The hose is wound on the reel, and an image acquisition device is provided on the reel for real-time acquisition of the state images of the fire hose.

As a further limitation, a processor is provided on the fire-fighting robot or fire-fighting medium supply equipment, and the processor is connected to the control system of the fire-fighting robot. The state image is processed and recognized. When the connection between the reel and the hose appears in the image, it is assumed that the hose has been fully expanded, and the processor controls the robot to stop the action; when the image shows the joint between the hose and the robot, it is recognized The hose has been completely retracted, and the processor controls the reel to stop.

As a further limitation, a first sensor is provided on the reel. When the hose is unrolled: the number of turns of the reel is monitored in real time. When the number of turns reaches a preset value, it is determined that the hose has been fully unrolled, and the first sensor Send a fully unfolding control signal to the processor, and the processor will control the robot to stop after receiving the fully unfolded control signal; when the hose is retracted: monitor the number of turns of the reel in real time, and when the number of turns reaches the preset value, It is determined that the hose has been completely retracted, and the first sensor sends a complete retracting control signal to the processor, and the processor controls the reel to stop the action after receiving the complete retracting control signal.

As a further limitation, the first sensor is a rotary encoder or a rotation sensor or other revolution measurement sensors.

As a further limitation, a pressure sensor is provided at the position of the fire hose or the second interface or the first interface, which is used to detect the water pressure in real time and transmit it to the processor, and compare it with the stored preset pressure threshold. When the pressure is lower than the preset pressure threshold, it is prompted that the pressure is insufficient, and the processor determines that there is a pipeline leak/hose damage.

As a further limitation, the processor is also configured to obtain the relative position of the second interface and the first interface by extracting more obvious feature values in the image, and to determine the current docking status by continuously capturing images and performing image processing.

As a further limitation, the processor is also configured to: while performing image processing, the second interface and/or the first interface is equipped with a first sensor, and when the second interface and the first interface are successfully connected, the first sensor is installed on the second interface and/or the first interface. The sensor sends a docking success signal to the processor.

As an optional implementation manner, a multi-eye vision device is provided on the fire fighting robot for collecting visual image information and infrared image information of the on-site environment.

As an alternative embodiment, at least one unmanned aerial vehicle is also deployed in the substation to collect image information from different perspectives of the equipment in the substation and communicate with the control center to assist the substation fire fighting robot to establish a three-dimensional model of the station. Assist the robot to determine its position coordinates in the station.

Specifically, the drone collects image information from different perspectives of the equipment in the station, and assists the substation fire fighting robot to establish a three-dimensional model of the station;

During the operation of the fire-fighting robot, the drone collects image information in the substation in real time and assists the robot in determining its position coordinates in the station;

The drone collects the image information of the burning equipment and determines the location of the fire point;

According to the position of the robot and the position of the ignition point, control the robot to plan the path and adjust the spray angle;

During the robot operation, the drone collects images of the location of the fire point in real time, determines the current status information of the fire point and transmits it to the robot, and the robot adjusts the fire fighting strategy according to the received current status information of the fire point.

The operation method of the above-mentioned intelligent fire fighting system is to establish a three-dimensional model in the station. When a fire occurs, determine the fire point and the position of each firefighting robot, and control at least one nearest firefighting robot to drive to the vicinity of the fire point according to the position of the robot and the position of the fire point;

Carry recognition and positioning of the equipment ignition point, analyze the position of the ignition point in the three-dimensional coordinate system, and combine the fire situation to adjust the ejection elimination curve based on multi-eye vision, calculate the ejection angle and ejection flow rate of the ejection device, and select the fire-fighting medium according to the type of equipment on fire; The injection angle, the injection flow rate and the fire fighting medium are fire extinguishing parameters;

When there is a fire, dispatch the nearest fire-fighting robot to dock with the corresponding fire-fighting medium supply equipment, and perform fire-fighting operations according to the determined fire-fighting parameters.

After the fire extinguishing operation is completed, the robot completes the stripping and the hose is recovered.

As an optional implementation, the specific process of calculating the injection angle and injection flow rate of the injection device in combination with the fire situation includes:

Obtain the visual image information and infrared image information of the scene environment collected by the multi-eye vision device;

Preprocess the obtained visual image information and infrared image information respectively;

Determine the fire area according to the preprocessing results of visual image information and infrared image information;

According to the fire area, establish the spray curve model, identify the drop point of the water column, and determine the best spray angle and spray flow;

Analyze the status of the fire equipment in the fire area and determine the best fire fighting location and distance;

Judge the size of the fire of the burning equipment and select the best spray mode.

When the selected fire-fighting medium is dry powder or fine water mist, it is sufficient that the spray coverage area includes the fire point.

As a further limitation, the step of preprocessing the visual image information includes:

Preprocess the visual image;

Perform grayscale processing and motion detection on the pre-processed image to determine whether there is a suspicious flame area in the visual image;

Filtering the area of suspicious flames, extracting the color histogram of the filtered image, extracting image feature values, and performing matching processing to determine the suspicious fire area in the visual image;

The suspicious fire area is divided and normalized.

As a further limitation, the steps of preprocessing infrared image information include:

The infrared image is preprocessed by grayscale and then segmented, the image feature value after segmentation is extracted, and the extracted image feature value is input into the trained neural network model for recognition, and the suspicious fire area of the infrared image is obtained.

As a further limitation, the method for determining the fire area is:

The suspicious fire area obtained after visual image processing is compared with the suspicious fire area obtained after infrared image processing. The overlapping suspicious fire area is regarded as the credible fire area, and the suspicious fire area that does not overlap is regarded as the suspected fire area. The non-suspected fire area is determined to be an area where no fire has occurred.

As a further limitation, the method for determining the optimal injection angle and injection flow rate is:

Take the bottom of the fire area as the target area to establish the injection curve model;

Obtain the spray image of the fire-fighting robot, and process the spray image to identify the drop point of the sprayed water column;

Determine the best spray angle according to the coordinate difference between the drop point of the water column and the fire area; adjust the spray flow rate according to the area ratio of the credible fire area and the suspected fire area in the fire area.

As a further limitation, the steps of analyzing the status of the fire equipment in the fire area and determining the best fire extinguishing position and distance include:

Preprocess the image of the fire area;

Extract the feature value of the image of the fire area after preprocessing;

Input the extracted eigenvalues into the neural network image recognition model to identify the burning equipment;

Choose the angle with the least obscuration among all directions of the fire equipment as the best extinguishing position;

According to the proportion of the fire area occupying the entire image, adjust the distance between the fire fighting robot and the fire equipment.

As a further limitation, the step of judging the size of the fire of the burning equipment and selecting the best injection mode includes:

Establish a sample library containing the fire extinguishing distance of the fired equipment and the basis for the judgment of the fire situation;

Obtain the fire extinguishing distance and fire judgment basis of the burning equipment from the sample library;

Compare the area of the burning area with the area of the burning equipment, and judge the fire size of the burning equipment according to the fire judgment basis of the burning equipment;

According to the size of the fire of the burning equipment, select the best spray mode.

Compared with the prior art, the beneficial effects of the present disclosure are:

The present disclosure proposes a multi-source heterogeneous fire-fighting equipment linkage control technology. Multiple fire-fighting robots and fire-fighting medium supply equipment are deployed in a substation. The fire-fighting robots and fire-fighting medium supply equipment are used to cooperate with each other, and the hose, inlet pipe, and transmission water pipe The length and flexibility can make full use of the limited space in the substation to form a dynamic pairing and temporary team formation. The fire-fighting robot and the fire-fighting medium supply equipment can be dynamically matched according to the location of the fire and the fire situation. , Realize early warning, ensure the first time to put out the fire, eliminate the fire at the embryonic stage, and ensure the safety of the substation.

This disclosure proposes a multi-sensor fusion fire hose docking monitoring technology, which uses real-time monitoring of sensors to accurately confirm the docking status of the fire-fighting robot and the fire-fighting medium supply equipment, ensuring the intelligence of the docking process, and accurately grasping the progress of the operation to facilitate the docking Deal with the error as soon as it occurs; at the same time, monitor the process of expanding and retracting the hose to assist the robot in finding the best operating point.

The present disclosure can realize the fire extinguishing of various media such as fire fighting water, foam, and dry powder. At the same time, it can adapt to the complex environment in the substation and target different areas and heights of the substation by using the mutual cooperation of the mobile chassis and the lifting mechanism, the rotary head and the rotary joint. The source is adaptively adjusted and extinguished, which has a great degree of freedom.

The fire-fighting robot of the present disclosure is provided with a self-spraying mechanism, and the robot self-sprays during the water spraying process to lower its temperature and ensure the safety of the robot itself.

The present disclosure proposes a rapid identification technology for the ignition point of substation equipment that integrates multiple images, combines the collected data, converts flame image information into flame space coordinate information, overcomes the difficult problem of not being able to accurately locate the flame position, and realizes the flame The fast and accurate positioning. The ability to quickly identify and analyze the fire point of the equipment is improved, and it provides data support for the selection of the best fire extinguishing method. Combined with the spray curve calculation algorithm, it can realize the automatic aiming of the spray device and quickly and effectively extinguish the fire source.

According to the present disclosure, according to the environment, path and fire intensity in the substation, when the fire intensity is small, the firefighting robot at the closest substation dispatching distance can carry fire extinguishing equipment and various sensors to the spot immediately, and use the quick docking mechanism to complete and firefighting medium supply equipment. The docking of the supply mechanism to achieve long-term continuous operation of fire extinguishing, and strive to achieve the best and fastest fire extinguishing.

The integrated intelligent fire-fighting robot for substations provided by the present disclosure can also be combined with existing fire early warning systems (such as sensors provided at intervals between substations) or sensors carried by each fire-fighting robot, or drone images to monitor the environment in the substation in real time. The fire is discovered at a time, and the fire-fighting robot is dispatched as soon as possible, and the fire-fighting robot is first carried out. During the work of the fire-fighting robot, the robots in other positions rush to the fire point according to the size of the fire, which saves the response time to the greatest extent and guarantees the fire fighting Effectiveness.

The present disclosure uses drones to cooperate in the air, combined with the three-dimensional registration fusion model of the substation, solves the problem of the fire detection algorithm's misrecognition of the interference light source, and realizes the rapid and accurate guidance of the fire fighting robot to the fire location in the event of a fire.

Description of the drawings

The accompanying drawings of the specification constituting a part of the present disclosure are used to provide a further understanding of the present disclosure, and the exemplary embodiments and descriptions of the present disclosure are used to explain the present disclosure, and do not constitute an improper limitation of the present disclosure.

Figure 1 is a side view of the fire-fighting robot of the present disclosure;

Figure 2 is a rear view of the fire-fighting robot of the present disclosure;

Figure 3 is a side view of the internal structure of the fire-fighting robot of the present disclosure;

Figure 4 is a top view of the internal structure of the fire-fighting robot of the present disclosure;

Figure 5 is a system block diagram of the fire-fighting robot of the present disclosure;

Figure 6 is a schematic diagram of the structure of the fire-fighting robot of the present disclosure;

Figure 7 is a schematic diagram of the socket structure of the present disclosure;

Figure 8 is a top view of the fire-fighting medium supply equipment of the present disclosure;

Figure 9 is a side view of the fire-fighting medium supply equipment of the present disclosure;

Figure 10 is a rear view of the fire-fighting medium supply equipment of the present disclosure;

Figure 11 is a side view of the fire-fighting medium supply equipment of the present disclosure;

Figure 12 is a block diagram of the fire protection medium supply equipment of the present disclosure;

Figure 13 is a specific flow chart of fire extinguishing in the present disclosure;

Figure 14 is an image processing flowchart of the docking process of the present disclosure;

Figure 15 is a schematic diagram of the substation intelligent fire protection system of the present disclosure.

Detailed ways:

The present disclosure will be further described below in conjunction with the drawings and embodiments.

It should be noted that the following detailed descriptions are all illustrative, and are intended to provide further descriptions of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present disclosure belongs.

It should be noted that the terms used here are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should also be understood that when the terms "comprising" and/or "including" are used in this specification, they indicate There are features, steps, operations, devices, components, and/or combinations thereof.

As shown in Figure 15, a substation intelligent fire fighting system includes multiple fire fighting robots and fire fighting medium supply equipment, as well as a control center, fire fighting robots (represented by circles in the figure), and fire fighting medium supply equipment (shown in boxes on the way Instead of representation) are deployed in each area of the substation (indicated by a hexagon) at intervals, and each has a positioning mechanism.

The fire-fighting medium supply equipment can be fixed, mobile, or both.

Of course, the figure is only one case, and it is a schematic diagram. The actual size and quantity can be changed adaptively according to the size and level of the substation.

When a fire occurs, the ignition point (indicated by a star icon) can be obtained according to the smoke/temperature sensors deployed in each area of the substation, or it can be obtained by the sensor carried on the firefighting robot, or it can be detected by configuring the drone .

The drone collects image information from different perspectives of the equipment in the station, and assists the substation fire fighting robot to establish a three-dimensional model of the station;

The substation fire-fighting robot uses the multi-eye vision equipment onboard, and takes the structural characteristics of the equipment in the station as the constraint, and uses multi-view reconstruction to obtain a primary model of the substation as a whole; the specific process is:

The parallax of two images is captured by multiple cameras (two in some embodiments) to construct a three-dimensional scene. After the target is detected, the three-dimensional information of the target is obtained by calculating the position deviation between corresponding points of the image.

By analyzing the picture returned by the camera, it is possible to analyze various distance information such as the distance between the device and the device in the picture, the distance from the shooting point to the device, and so on. The equipment in the station usually has a fixed specification, and the length and height information are known, so it can be used as a reference when measuring the distance.

The disadvantage of establishing a model through stereo vision technology is that the accuracy is not high enough, so the primary model is established.

The drone is equipped with multiple sensors and uses tilt photography technology to collect images from five different angles: vertical, front view, rear view, left view, and right view at the same time, to obtain rich high resolution of the top and side views of the building Rate texture. Through this method, the situation of the ground objects can be truly reflected, the texture information of the object side can be obtained with high precision, and a real three-dimensional model can be generated through positioning, fusion, modeling and other technologies.

Combined with the primary model established by the multi-eye vision device, the real 3D model and the primary model are closely matched to generate an accurate 3D vision model.

Dense matching of multi-view images can obtain high-precision and high-resolution digital surface models. The matching method is:

Operators are used to detect corner points, and then feature descriptors are used to describe the detected corner points, and the image feature points are matched according to the corresponding matching criteria. The specific process is:

Use the improved PMVS algorithm (a three-dimensional multi-view stereo vision algorithm based on a patch), including the following steps:

Initial feature matching, face element diffusion, face element filtering.

The purpose of initial feature matching is to generate a series of sparse surface elements, which are used as seed points for diffusion. The process of diffusion and filtering is repeated many times, and the sparse seed points are diffused to generate dense point clouds and delete error point clouds.

Initial feature matching: The first is feature detection. Commonly used feature extraction operators include Harris corner extraction operator and DoG operator. After feature detection, image matching is performed.

When performing image matching, on the basis of the original algorithm, some constraints are added to obtain more accurate matching points, which are used as seed points for subsequent operations. According to the preliminary model obtained before, the candidate space points with higher confidence are selected Points are used as seed points.

During the operation of the fire-fighting robot, the drone collects image information in the substation in real time and assists the robot in determining its position coordinates in the station;

When a fire occurs, the robot is affected by smoke and other harsh environments. The visual positioning technology may cause inaccurate positioning, which can easily lead to errors in the subsequent process of determining the position of the water supply point, path planning, and spraying angle, which is not conducive to the fire fighting robot. Work autonomously.

In this embodiment, when the robot is moving in the station to the scene of the fire, the drone takes images in the air to obtain the location of the burning equipment, the distribution of peripheral equipment, and the relative position information of the robot and other equipment; the image data is transmitted back The background is used to detect the on-site environment.

At the same time, it uses algorithms to process the returned images, build a simple model, and compare it with a pre-established accurate model to assist the robot in determining its own position coordinates in the station.

When the robot is operating in the station, the drone will send back an image in the air, then model the image based on the image returned by the drone. At this time, the robot exists in the model as a mobile device in the station and will have the model The next coordinate. At the same time, the sensor installed by the robot reads the precise model built in advance, and also measures its own coordinates under the fine model.

The fine model and the model established by the drone are both models in the same substation. Therefore, the same coordinate system can be used for measurement. The two coordinate values in the same coordinate system can be compared to assist the robot in determining its own position in the station. Position coordinates.

The above method enables the robot to accurately perform coordinate positioning even when the vision positioning technology is inaccurate; it overcomes the influence of smoke or spraying water mist on the positioning of the robot itself.

The control center dispatches the fire-fighting robot a closest to the fire point to go to the fire point first to fight the fire. At the same time, it can dispatch the fire-fighting robot b closer to the fire point, or when the fire-fighting medium supply equipment is equipped with a movable type, as shown in the figure. Show the fire-fighting medium supply equipment i, and dispatch it to the fire point.

When the fire-fighting robot a reaches the fire point, first use its own fire extinguishing mechanism to extinguish the fire. If it is judged that the fire is small, you can stop calling other fire-fighting robots and fire-fighting medium supply equipment. If the fire is large, continue to dispatch closer to the fire point. Fire-fighting robot b, or fire-fighting medium supply equipment i, go to the fire point, or dispatch more fire-fighting robots and fire-fighting medium supply equipment to the fire point. To ensure a one-time fire extinguishment. Save response time to the greatest extent, while ensuring the effectiveness of rescue work.

Of course, if the fire-fighting medium supply equipment is fixed, the control center only dispatches the actions of the fire-fighting robots.

Specifically, the fire-fighting robot mentioned, as shown in Figure 6, includes a crawler-type mobile chassis assembly 601, a housing assembly 602, a lifting and turning device 603, a dry powder injection device 604, a fire-fighting water/foam injection device 605, a control device 606, etc. ,among them:

As shown in Figure 1, the crawler-type mobile chassis assembly includes upper 1 crawler, 2 driving wheels, 3 single rollers, 4 supporting rollers, 1, 5 double supporting wheels, 6 supporting rollers, 2, 7 guide wheels , 8 chassis shell and other components, the motor reducer component is installed in the 8 chassis shell, 2 driving wheels are installed on the motor reducer component, 3 single rollers, 4 supporting pulleys, 5 double supporting wheels, 6 supporting wheels Pulleys 2, 7 guide wheels are respectively installed on the 8 chassis shell and 1 crawler is installed on them;

Shell components, including the upper 9 range sensor, 10 visible light camera, 11 screws, 12 headlights, 13 front housing, 14 screws, 15 rear housing, 16 visible light camera, 9 range sensor, 10 visible light camera, Two 12 headlights are installed on the front shell 13 respectively, the front shell 13 is installed on the 8 chassis shell with 11 screws, the 16 visible light camera is installed on the rear shell 15, and the rear shell 15 is installed on the rear shell 15 with 14 screws. 13 On the front housing;

As shown in Figure 2, Figure 3 and Figure 4, the lifting and turning device includes 17 lifting mechanism, 18 screws, 19 turning base, 20 turning head, 21 screws, 22 lighting headlights, 23 visible light camera, 24 infrared thermal imager The 17 lifting mechanism is mounted on the 8 chassis shell through 21 screws, the 19 revolving base is mounted on the 17 lifting mechanism through 18 screws, the 20 revolving head is mounted on the 19 revolving base, 22 lighting headlights, 23 visible light cameras, The 24 infrared thermal imager is installed on the 20 rotary head.

Dry powder spray device, including 25 dry powder tank, 26 solenoid valve powder inlet pipe, 27 solenoid valve, 28 solenoid valve powder outlet pipe, 29 three-way joint, 30 powder outlet pipe, 31 dry powder nozzle, 25 dry powder tank and 27 solenoid valve Fixed on the 8 chassis shell, one end of the 26 solenoid valve powder inlet pipe is connected to the 25 dry powder tank, the other end is connected to the 27 solenoid valve, one end of the 28 solenoid valve powder pipe is connected to the 27 solenoid valve, the other end is connected to the 29 three-way connector, and the other end is connected to the powder outlet. One end of the pipe is connected with a 29 tee joint, and the other end is connected with a 31 dry powder nozzle, which is fixed on a 20 rotary head.

As shown in Figure 4, the fire-fighting water/foam spray device includes 32 hose joints, 33 water pipes, 34 main water pipes, 35 solenoid valves, 36 rotary joints, 37 solenoid valve water pipes, 38 rotary joints, 39 water pipes, and 40 solenoid valve water pipes. , 41 fire sprinklers, 42 self-sprinkler device, 43 standpipe, 44 standpipe, 45 self-sprinkler device, etc., 34 main water pipe and 35 solenoid valve are fixed on the 8 chassis shell, 34 main water pipe one end is connected to 32 water With connector, the other end is connected to 35 solenoid valve, 37 solenoid valve water pipe and 40 solenoid valve water pipe are respectively installed on 35 solenoid valve, one end of 43 riser pipe is connected to 37 solenoid valve water pipe, the other end is connected to 36 rotary joint, and one end of riser pipe is connected to 40 The other end of the solenoid valve water pipe is connected to 38 rotary joint, one end of 33 water pipe is connected to 36 rotary joint, the other end is connected to 41 fire water sprinklers, one end of 39 water pipe is connected to 38 rotary joints, and the other end is connected to 41 fire water sprinklers. 42 self-spraying device is installed in On 43 standpipes, 45 self-spraying devices are installed on 44 standpipes.

31 dry powder sprinklers and 41 fire water sprinklers can be arranged in a shell to ensure the safety of the sprinklers. And there are holes in the nozzles of the shell, so that the two nozzles can work. 41 Fire water sprinklers can spray water jets or fine water mist.

As shown in Figure 7, the 32 hose connector can be quickly connected to the fire-fighting water supply equipment. The 32 hose connector is the first connector and is matched with the socket of the fire-fighting medium supply equipment. The socket includes a support seat, and the support The seat is provided with a plurality of uprights, the uprights are distributed around the circumference, and the water inlet pipe can be accommodated in the center, the upright is provided with an elastic body, the end of the upright is provided with a pressing plate, and the pressing plate is provided with opposing The rotating pressing member is movably clamped to the first joint.

As shown in Figure 5, the control device includes 46 robot master control module, 47 power supply module, 48 power display module, 49 relay module 1, 50 screen splitter, 51 water cannon camera, 52 tilt sensor, 53 temperature sensor, 54 sound Light alarm module, 55 self-sprinkler control module, 56GPS positioning module, 57 relay module 2, 58 relay module 3, 59 motor driver 1, 60 motor driver 2, 61 robot motion motor and so on. The 46 robot master control module is connected to the 47 power supply module through the 48 power display module; the 46 robot master control module is connected to the 12 lighting headlights and the 22 lighting headlights through the 49 relay module 1; the 46 robot master control module is connected to the 50 screen splitter respectively Connected with 10 visible light cameras, 16 visible light cameras, 51 water cannon cameras; 46 robot master control modules are respectively connected with 24 infrared thermal imagers, 9 distance sensors, 52 inclination sensors, 53 temperature sensors, 54 sound and light alarm modules, and 56 GPS positioning Modules are connected; 46 robot master control modules are connected to 42 automatic spray devices and 45 automatic spray devices through 55 automatic spray device control modules; 46 robot master control modules are connected through 57 relay modules 2 and 27 solenoid valves; 46 robot master The control module is connected to 35 solenoid valve through 58 relay module 3; 46 robot master control module is connected to 17 lifting mechanism through 59 motor driver 1; 46 robot master control module is connected to 61 robot motion motor through 60 motor driver 2. In the foregoing embodiment of the present invention, the 46 robot master control module may be a controller, a processing chip (CPU), or a single-chip microcomputer.

The substation fire-fighting robot can be remotely controlled by the remote control. The operator can observe the fire scene through the visible light camera, the water cannon camera, and the infrared thermal imager; the real-time measurement of the robot's performance through the ranging sensor, the inclination sensor, the temperature sensor, and the GPS positioning module Posture and surrounding environment information, when the robot encounters danger, the sound and light alarm module can send out sound and light alarm prompts; through the remote control, you can control and choose to use various fire extinguishing media such as dry powder, fire water and foam for fire extinguishing and cooling operations.

As shown in Figures 8-11, the fire-fighting medium supply equipment includes hose components, hose docking devices, hose retraction devices, water storage pipe components, booster pump components, mobile chassis components, foam pump components, foam storage tank components, and fire fighting Water storage tank components, control components, etc.

Hose components, including 2-1 quick-plug hose connector, 2-2 hose, 2-3 threaded hose connector, 2-1 quick-plug hose connector and 2-2 hose buckle together, 2-3 The threaded hose connector and the 2-2 hose are buckled together, the 2-3 threaded hose connector is fixed on the hose retracting device at one end, and the 2-1 quick-plug hose connector is placed on the hose docking device 2-19 reel ；

Hose docking device, including 2-4 hose support seat, 2-5 shaft screw, 2-6 compression plate, 2-7 screw, 2-8 support plate, 2-9 pillar, 2-10 spring, 2- 11 block female, 2-12 base, 2-13 screws, etc., two 2-6 compression plates are fixed on the 2-4 hose support seat by two 2-5 shaft screws, as shown in Figure 10, The tight plate fixes the hose on the hose support seat for positioning. When the robot pulls out the hose, the pressure plate can move during the process of lifting the hose without hindering the movement of the hose. During the docking process, the spring It can slow down the impact force during the docking process of the robot. The 4 2-10 springs are respectively sleeved on the 2-9 pillars. One end of the 4 pillars is fixed on the 2-4 hose support seat, and the other end passes through the 2-8 support. Plate, 2-12 base and 2-11 block nut are fixed together, 2-8 support plate and 2-12 base are connected together by 2-7 screws, 2-12 base is connected with 2 of the mobile chassis assembly by 2-13 screws -39 on the bottom plate.

The hose retracting device, including 2-14 screws, 2-15 reel motors, 2-16 pinions, 12-7 screws, 2-18 frame body, 2-19 reels, 2-20 central shafts, etc., 2 -15 reel motor is fixed on the 2-39 bottom plate of the mobile chassis assembly by 2-14 screws, 2-16 pinion is fixed on the output shaft of 2-15 reel motor by 12-7 screws, pinion and reel motor The output shafts are fixed together. The output shaft of the reel motor rotates to drive the pinion gear to rotate. There is a ring gear on the reel that meshes with the pinion gear. The pinion gear rotates to drive the reel to rotate. The 2-18 frame is welded to the mobile chassis assembly. On the 2-39 bottom plate, one end of the 2-20 central axis is connected with the water storage pipe assembly through a 2-23 connecting pipe, and the other end passes through the 2-19 reel and is fixed on the 2-18 frame.

Water storage pipe assembly, including 2-21 booster pump drain pipe I, 2-22 booster pump drain pipe II, 2-23 connecting pipe, 2-24 overflow valve, 2-25 overflow valve return pipe, 2-26 Screw, 2-27 booster pump inlet pipe I, 2-28 three-way joint, 2-29 booster pump inlet pipe II, 2-30 booster pump main inlet pipe, 2-31 foam pump discharge pipe, 2- 32 foam pump inlet pipe, 2-33 drain ball valve, etc. The 2-23 connecting pipe is fixed on the 2-39 bottom plate of the mobile chassis assembly with 2-26 screws, and the 2-24 overflow valve is connected to the 2-23 connecting pipe. On the upper part, one end of the 2-21 booster pump drain pipe I and the 2-22 booster pump drain pipe is connected to the 2-34 booster pump, the other end is connected to the 2-23 connecting pipe, and the 2-25 overflow valve return pipe One end is connected to the 2-24 overflow valve, the other end is connected to the 2-30 booster pump main inlet pipe, 2-27 booster pump inlet pipe I, and 2-29 booster pump inlet pipe II are connected at one end to 2 On the -34 booster pump, the other end is connected to the 2-28 three-way joint. One end of the 2-30 booster pump’s main water inlet pipe is connected to the 2-28 three-way joint, and the other end is connected to the 2-33 drain ball valve. The -33 drain ball valve is connected to the fire water storage tank.

Booster pump components, including 2-34 booster pump, 2-35 booster pump bracket, 2-36 screws, 2-37 bolts, 2-38 nuts, etc. 2-34 booster pump passes 4 sets of 2-37 The bolts, 2-38 nuts and 2-35 booster pump bracket are connected together, and the 2-35 booster pump bracket is connected to the 2-39 bottom plate of the mobile chassis assembly through 2-36 screws.

Mobile chassis components, including 2-39 bottom plate, 2-40 movable wheels, etc., two 2-40 movable wheels and 2-39 bottom plate are welded together.

Foam pump components, including 2-41 foam pump bracket, 2-42 screws, 2-43 screws, 2-44 foam pumps, etc. The 2-41 foam pump bracket is fixed to 2-39 of the mobile chassis assembly by 2-43 screws On the bottom plate, the 2-44 foam pump is fixed on the 2-41 foam pump bracket with 2-42 screws. The foam pump discharges the foam from the foam storage tank into the foam pump discharge pipe. The 2-31 foam pump discharge pipe is connected to the 2-30 booster pump main water inlet pipe. The foam mixed with the fire-fighting water from the foam pump enters the booster pump together. .

Foam storage tank components, including 2-45 ball valve, 2-46 screw, 2-47 foam level indicator, 2-48 foam storage tank, 2-49 screw plug, etc. 2-45 ball valve is installed on 2-48 foam On the storage tank, the 2-49 screw plug is installed on the 2-48 foam storage tank as an inlet for filling foam, the 2-47 foam level indicator is installed on the 2-48 foam storage tank, and the 2-48 foam storage tank passes through Four 2-46 screws are fixed on the 2-51 fire water storage tank.

Fire water storage tank components, including 2-50 pressure relief valve, 2-51 fire water storage tank, 2-52 water inlet connector, 53 fire water level indicator, 2-50 pressure relief valve and 2-52 inlet The water joint is welded to the 2-51 fire water storage tank, the 53 fire water level indicator is installed on the 2-51 fire water storage tank, and the 2-51 fire water storage tank is welded to the 2-39 bottom plate of the mobile chassis assembly.

The fire water storage tank assembly is connected to the fire hydrant through a 2-52 water inlet connector and a control valve, and the control valve can be remotely controlled.

As shown in Figure 12, the control components include 2-54 boost switch, 2-55 emergency stop switch, 2-56 foam mixing switch, 2-57 reel switch, 2-58 wireless transceiver module, 2-59 master control System, 2-60 power module 24V, 2-61 power module 48V, 2-62 booster pump soft starter, 2-63 reel motor soft starter, 2-64 foam pump soft starter, including 2-54 Booster switch, 2-55 emergency stop switch, 2-56 foam mixing switch, 2-57 reel switch, 2-58 wireless transceiver module, 2-60 power module 24V, 2-61 power module 48V, 2-62 increase The pressure pump soft starter, the 2-63 reel motor soft starter, and the 2-64 foam pump soft starter are respectively connected to the 2-59 master control system.

When the 2-54 booster switch is pressed, the 2-59 master control system controls the 2-34 booster pump to work through the 2-62 booster pump soft starter, and the fire water passes through the 2-51 fire water storage tank. 30 booster pump main inlet pipe, 2-27 booster pump inlet pipe I, 2-28 three-way joint, 2-29 booster pump inlet pipe II, 2-21 booster pump drain pipe I, 2-22 booster Pump drain pipe II, 2-23 interconnection pipe, 2-19 reel, 2-2 hose, 2-1 quick-insertion hose connector to supply water to the robot; when the 2-54 boost switch is pressed again, 2-59 total The control system controls the 2-34 booster pump to stop working through the 2-62 booster pump soft starter.

When the 2-56 foam mixing switch is pressed, the 2-59 master control system controls the 2-44 foam pump to start working through the 2-64 foam pump soft starter, and the foam passes through the 2-32 foam pump inlet pipe and 2-31 foam The pump discharge pipe enters the fire fighting water; when the 2-56 foam mixing switch is pressed again, the 2-59 master control system controls the 2-44 foam pump to stop working through the 2-64 foam pump soft starter.

If foam is needed during the fire extinguishing process, the foam pump is turned on and the fire water is mixed to extinguish the fire. If no foam is required, the foam pump is not turned on and only fire water is used to extinguish the fire. Choose freely according to the fire extinguishing characteristics of the fire equipment.

When the 2-57 reel switch is pressed, the 2-59 master control system controls the 2-15 reel motor to start working through the 2-63 reel motor soft starter, and the reel boiled water is recovered; when the 2-57 reel is pressed again After the reel switch, the 2-59 master control system controls the 2-15 reel motor to stop working through the 2-63 reel motor soft starter.

When a fire occurs, the fire-fighting robot quickly arrives at the fire site, uses the infrared thermal imaging system to identify and locate the fire point of the equipment, analyzes the position of the fire point in the three-dimensional coordinate system, and calculates the jet angle and jet flow rate of the jet device based on the fire equipment. Choose the fire-fighting medium such as dry powder, water jet or fine water mist. After calculating the various fire-fighting parameters, if the robot needs to be connected with the fire-fighting medium supply equipment, the operator can choose 3L/s, 6L/s and 9L/s fire fighting water supply speed, foam ratio can be selected 3%, 5%, 10% and 20% according to the demand. After the parameters are set, start the fire fighting medium supply equipment to provide fire fighting water or foam mixture to the fire fighting robot. Carry out fire-fighting operations. After the fire-fighting operations are completed, the robot automatic stripping device completes the automatic stripping, and the fire-fighting medium supply equipment can realize the automatic recovery of the hose.

The operator can control the robot to realize the automatic docking of the fire hose through the docking camera of the fire fighting robot. The docking device equipped with the fire fighting media supply equipment has a multi-directional buffer function, which can realize automatic adjustment of small errors during the docking process, and quickly realize the docking function.

The first interface is set at the rear of the robot. The robot first stops at a position within the set error range, and reverses backwards. The elastic docking device performs error correction, and the second interface of the hose is docked to the robot;

When it is monitored that the first interface and the second interface cannot be docked due to an excessive error, the fire control platform analyzes the range of the error through the collected images, and feeds the result back to the robot control system to control the robot to move forward and move forward. Adjust the angle to reverse again, reduce the error and re-docking.

Specifically, the docking process is shown in Figure 14. The fire hose automatic docking monitoring is divided into two parts: image processing and image recognition. The fire control platform performs image processing on the images collected by the image acquisition device. Specifically: Denoising, smoothing, and transforming images through image preprocessing, enhancing important features of images, performing image segmentation, edge detection, and image refinement on preprocessed images. Among them, the image segmentation part is based on regional features. Segmentation methods, segmentation methods based on correlation matching, and segmentation methods based on boundary features.

Perform image recognition on the processed image, specifically: use the neural network image recognition model fused with genetic algorithm and BP network to recognize the image, extract important feature values in the processed image, and obtain according to the extracted feature values The current state of the equipment in the image, such as when judging whether the docking is successful, mainly extracts the characteristics of the hose connector, recognizes the hose connector and the robot connector, and calculates their relative position; when judging whether the hose is fully deployed, it is targeted at the hose Train the characteristics of the connection with the hose reel to identify whether the connection appears in the image. When judging whether the hose is completely retracted, extract the characteristics of the hose connector for training, and identify whether the hose connector appears in the image .

As shown in Figure 13, the method for adjusting the spray curve of a fire-fighting robot includes the following steps:

S101: Acquire visual image information and infrared image information of the on-site environment collected by the multi-eye vision device.

The image information of the scene environment is collected through the ordinary vision camera of the multi-eye vision device, including the image information of the equipment in the scene environment, the fire image information in the scene environment, and the smoke concentration information in the scene environment. If it is at the fire scene, visual image information such as the fire equipment, the size of the fire, and the smoke concentration can be collected through multi-eye vision equipment.

The infrared image of the scene environment is collected through the infrared camera of the multi-eye vision device, which mainly includes the temperature, the highest temperature, the position where the highest temperature occurs, and the shape of the flame in the scene environment. If you are at a fire scene, you can collect information such as the highest temperature in the scene environment, the location where the highest temperature occurs, and the shape of the flame.

S102: Perform preprocessing on the obtained visual image information and infrared image information respectively, and determine a corresponding suspicious fire area.

In the step 102, an image processing algorithm is used to perform image graying, segmentation, filtering and other processing on the image obtained in step 101, and the corresponding suspicious fire areas are respectively determined.

Specifically, in the step 102, the specific implementation process of preprocessing the visual image information is as follows:

First, perform color detection on the image, such as large orange-red or black, and do preliminary processing such as calculation of specific gravity.

Then, perform gray-scale processing and motion detection on the preliminarily processed image to determine whether there is a suspicious flame area in the image.

Filtering the area of suspicious flames, extracting the color histogram of the filtered image, extracting image feature values, and performing matching processing to determine the area of suspicious fires in the image.

Finally, the suspicious fire area is divided and normalized as the basic unit for subsequent research and judgment.

The processing of the acquired infrared image is relatively simple. The infrared image is preprocessed by image grayscale and then segmented, the image feature value after segmentation is extracted, and the extracted image feature value is input into the trained neural network model for recognition, and then you can get Suspicious fire area in infrared image.

S103: According to the preprocessing result of the visual image information and the infrared image information, locate the fire area.

In this embodiment, the fire area includes a trusted fire area and a suspected fire area.

Specifically, in the step 103, the suspicious fire area after visual image processing is compared with the suspicious fire area after infrared image processing, and the overlapping suspicious fire area is regarded as the credible fire area. For suspected fire areas, if the overlapping unsuspected fire areas are judged to be non-fire areas, they are areas where no fire has occurred.

S104: Establish an injection curve model according to the fire area, identify the drop point of the water column, and determine the optimal injection angle and injection flow rate.

Specifically, in step 104, aiming is performed after determining the fire area. According to the obtained credible fire area, the bottom of the credible fire area is taken as the target area. Since the water jet curve and the landing point of the equipment are relatively fixed, the jet can be established. Curve model, adjust the angle and height of the pan-tilt, so that the point of the curve model falls within the credible fire area. After spraying, the spraying picture returned by other cameras on the robot is used, and the algorithm is called for image processing, and the spraying is identified in the image The drop of water column.

The specific implementation process of processing the jet image and identifying the drop point of the jetted water column is as follows:

Preprocess the jet image, including operations such as denoising, smoothing, and transformation;

Extract the characteristic value of the jetted water column in the image after preprocessing;

The extracted characteristic values of the jetted water column are input into the neural network image recognition model to identify the drop point of the jetted water column.

When there is no credible fire area, based on the obtained suspected fire area, take the bottom of the suspected fire area as the target area, establish a spray curve model, adjust the angle and height of the pan-tilt, and make the fall point of the curve model fall in the credible fire area Inside, after spraying, the images returned by other cameras on the robot are called, and the algorithm is called for image processing. The position of the sprayed water column is identified in the image, and the best spray angle is determined according to the coordinate difference between the drop point of the water column and the suspected fire area.

In this embodiment, the steps 101-103 are always executed to analyze the fire situation in the scene image in real time. After the credible fire area is reduced and disappeared, spray the suspected fire area until the camera returns the image Until there is no fire area.

In this embodiment, the injection flow rate is divided into three levels from large to small, and adjusted according to the area of the credible fire area and the suspected fire area, usually the maximum flow. When the credible fire area occupies less than the suspected fire area, it is the medium flow. , Use small traffic when there is no credible fire zone.

In this embodiment, according to the judgment result, the injection curve can be adjusted, the ignition point can be accurately targeted, and the injection flow rate and angle can be selected.

S105: Analyze the condition of the fired equipment and determine the best fire-extinguishing position and distance.

Specifically, the method for analyzing the condition of the burning equipment is:

Preprocess the image of the burning area, including operations such as denoising, smoothing, and transformation;

Extract the feature value of the burning equipment in the image of the burning area after preprocessing;

The extracted feature values of the burning equipment are input into the neural network image recognition model to identify the burning equipment.

Determine the best extinguishing position, including selecting the angle, that is, the angle with the least obscuration among all directions of the fire equipment. Determining the distance: The fire area occupies about 1/3 of the camera screen. If it occupies a small area, it will be close, and if it occupies a large area, it will be far. The robot will give priority to whether it will hit an obstacle when adjusting the distance.

S106: Analyze the on-site fire situation and select the best injection mode.

The specific implementation process of step 106 is:

Studying and judging the fire situation at the scene mainly depends on the relative size of the flame and comparing the area ratio of the fire area to the entire equipment to determine the fire situation of the equipment on site. Different equipment has different rules for on-site fire judgment. For example, for power equipment with a length, width and height of approximately 1m, if the fire area occupies more than half of the area of the equipment surface design, it is considered a large fire, and about one-third is a medium fire. , Less than one-third is a small fire. For power equipment with a length, width and height of about 3m, then one-third of the area is considered a big fire.

According to the different equipment in the station, establish different sample libraries. When the robot recognizes that the equipment is on fire or receives alarm information (such as "xx equipment on fire"), various information such as suitable fire fighting distance and fire judgment basis are available. Obtained directly from the library, the robot performs operations through real-time judgment based on various information such as the fire extinguishing distance and the basis for fire judgment obtained from the sample library.

The information in the sample library is obtained through training in advance, and the real-time judgment results obtained by the robot during each operation will also be stored in the sample library.

In this embodiment, the injection mode includes three injection modes: large, medium, and small. At the beginning of the operation, high fire will be selected. As the fire intensity decreases, medium fire and small fire will be selected. It is to reduce the temperature, and then choose medium heat or low heat at the beginning.

According to the fire size of the field equipment and the area of the entire equipment, select the corresponding spray mode.

The above descriptions are only preferred embodiments of the present disclosure, and are not used to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Although the specific embodiments of the present disclosure are described above in conjunction with the accompanying drawings, they do not limit the scope of protection of the present disclosure. Those skilled in the art should understand that on the basis of the technical solutions of the present disclosure, those skilled in the art do not need to make creative efforts. Various modifications or deformations that can be made are still within the protection scope of the present disclosure.

Claims (27)

1. An intelligent fire protection system for a substation, which is characterized in that it includes a plurality of fire fighting robots and fire fighting medium supply equipment capable of linkage control, and a control center. The fire fighting robots and fire fighting medium supply equipment are arranged at intervals in each area of the substation. The fire-fighting medium supply equipment can be detachably connected with the fire hydrant to realize continuous water supply, among which:

   The fire-fighting robot includes a robot body on which a fire-fighting water/foam spray mechanism with a first interface is provided; the fire-fighting medium supply equipment carries a water supply mechanism and a foam supply mechanism, and the corresponding fire extinguishing medium is provided through a hose, The end of the hose is provided with a second interface adapted to the first interface;

   The control center is configured to receive fire information, and dispatch a corresponding number of fire-fighting robots with a short distance or a corresponding number of fire-fighting robots and fire-fighting medium supply equipment with a short distance according to the fire situation to the vicinity of the fire point, through the first interface and the second The second interface is automatically and quickly connected to implement fire fighting.
2. The intelligent fire protection system for a substation according to claim 1, wherein at least one unmanned aerial vehicle is also deployed in the substation to collect image information from different perspectives of the equipment in the substation and communicate with the control center. It is used to assist the substation fire-fighting robot to establish a three-dimensional model of the station, and to assist the fire-fighting robot to determine its position coordinates in the station.
3. The intelligent fire-fighting system of a substation according to claim 1, characterized in that: the fire-fighting robot includes a robot body, and a fire-extinguishing mechanism is arranged on the robot body, and the fire-extinguishing mechanism includes a fire-fighting water/foam spray mechanism, and the fire-fighting water /Foam spray mechanism includes at least one water inlet pipe, one end of the water inlet pipe is provided with a first interface, the other end is connected to a transmission water pipe, the other end of the transmission water pipe is provided with a rotary joint, and the rotary joint is provided with a spray nozzle, and The rotary joint is connected with the robot body through a lifting mechanism.
4. The intelligent fire-fighting system for a substation according to claim 1, characterized in that: the fire-fighting medium supply equipment carries a water supply mechanism and a foam supply mechanism, wherein:

   The water supply mechanism includes a number of fire-fighting water storage tanks. The fire-fighting water storage tank is connected to a booster pump through a pipeline. The drain pipe of the booster pump is connected to a hose, and the other end of the hose is provided with a first interface. Equipped second interface;

   The foam supply mechanism includes a plurality of foam storage tanks, the foam storage tanks are connected to the foam pump through a pipeline, the drain pipe of the foam pump is connected to the hose, and the other end of the hose is provided with a first interface. Adapted second interface.
5. The intelligent fire protection system of a substation according to claim 3, characterized in that: the fire extinguishing mechanism further includes a dry powder spraying mechanism, which is arranged on the mobile chassis of the fire fighting robot, and specifically includes a number of dry powder tanks and outlets for the dry powder tanks. It is connected to a spray head through a pipeline, and the spray head is arranged on a rotary head, and the rotary head is arranged on a mobile chassis through a lifting mechanism, so that the height and angle of the dry powder spray can be adjusted.
6. A substation intelligent fire protection system according to claim 5, characterized in that: the dry powder injection mechanism includes a number of solenoid valves and a number of dry powder tanks, the powder inlet pipe of the solenoid valve is connected to the corresponding dry powder tank, and the outlet of the solenoid valve One end of the powder pipe is connected with a solenoid valve, the other end is connected with a multi-way joint, and one end of the multi-way joint is connected with a dry powder nozzle, which is fixed on the rotary head.
7. The intelligent fire protection system of a substation according to claim 3, characterized in that: the fire water/foam spray mechanism includes a main water pipe and a solenoid valve, one end of the main water pipe is connected to the water inlet pipe, the other end is connected to the spray nozzle, and the main water pipe is provided There is a solenoid valve, through which the spraying work of the fire water/foam spraying mechanism is controlled.
8. A substation intelligent fire protection system according to claim 3, characterized in that: the fire protection robot is also provided with a self-spraying mechanism, which specifically includes a standpipe and a sprinkler, the water inlet pipe is connected to the standpipe through a pipeline, so The angle between the standpipe and the vertical direction is less than or equal to 15°, and the nozzle is installed on the standpipe.
9. The intelligent fire protection system for a substation as claimed in claim 3, characterized in that:

   The fire-fighting medium supply equipment is provided with a hose docking mechanism, the hose docking mechanism includes a socket provided on the fire-fighting medium supply equipment, the socket includes a support base, and a plurality of uprights are provided on the support base. Circumferential distribution, the water inlet pipe can be accommodated in the center, the upright column is provided with an elastic body, the end of the upright column is provided with a pressing plate, and the pressing plate is provided with a relatively rotatable pressing member for the movable clamping Connect to the second interface.
10. The intelligent fire protection system of a substation according to claim 1, characterized in that: the fire protection robot body and the fire protection medium supply equipment are each provided with at least one image acquisition device for real-time monitoring of the second interface and the first interface. Docking status.
11. The intelligent fire protection system of a substation according to claim 10, characterized in that: the fire protection medium supply equipment is also provided with a hose retraction mechanism, which specifically includes a reel, a driving part and a support frame, and the driving part is transmitted through a transmission The parts are connected with the reel to drive the reel to rotate, the hose is wound on the reel, and the reel is provided with an image acquisition device for real-time acquisition of the status image of the fire hose.
12. A substation intelligent fire protection system according to claim 11, characterized in that: a processor is provided on the fire fighting robot or fire fighting medium supply equipment, and the processor is connected to the control system of the fire fighting robot, and the processor receives the image collection The information of the equipment is processed and recognized on the real-time captured images of the state of the fire hose. When the connection between the reel and the hose appears in the image, it is determined that the hose has been fully expanded, and the processor controls the robot to stop its action; When a joint between the hose and the robot appears in the image, it is determined that the hose has been completely retracted, and the processor controls the reel to stop.
13. The intelligent fire protection system for a substation according to claim 12, characterized in that: the reel is provided with a first sensor, when the hose is unrolled: the number of turns of the reel is monitored in real time, and the number of turns reaches a preset value When the hose is fully unfolded, the first sensor sends a complete unfolding control signal to the processor, and the processor controls the robot to stop after receiving the fully unfolded control signal; when the hose is retracted: monitor the rotation of the reel in real time When the number of turns reaches the preset value, it is determined that the hose has been fully retracted, and the first sensor sends a fully retracted control signal to the processor, and the processor controls the reel to stop after receiving the fully retracted control signal.
14. A substation intelligent fire protection system according to claim 12, characterized in that: the fire hose or the second interface or the position of the first interface is provided with a pressure sensor for real-time detection of water pressure and transmission to the processor, It is compared with the stored preset pressure threshold. When the water pressure is lower than the preset pressure threshold, it is prompted that the pressure is insufficient, and the processor determines that there is a pipeline leak/water belt damage.
15. A substation intelligent fire protection system according to claim 12, characterized in that: the processor is further configured to: obtain the relative position of the second interface and the first interface by extracting the more obvious feature values in the image, and Continue to capture pictures and perform image processing to determine the current docking status;

    Or, the processor is further configured to: while performing image processing, a first sensor is installed on the second interface and/or the first interface, and when the second interface and the first interface are successfully docked, the first sensor is configured to process The device sends a docking success signal.
16. An operation method of an intelligent fire fighting system, which is characterized by: establishing a three-dimensional model of the station, determining the fire point and the position of each fire fighting robot after a fire occurs, and controlling at least one fire fighting robot nearest to the fire point according to the robot position and the fire point position nearby;

    Carry recognition and positioning of the equipment ignition point, analyze the position of the ignition point in the three-dimensional coordinate system, and combine the fire situation to adjust the ejection elimination curve based on multi-eye vision, calculate the ejection angle and ejection flow rate of the ejection device, and select the fire-fighting medium according to the type of equipment on fire; The injection angle, the injection flow rate and the fire fighting medium are fire extinguishing parameters;

    When a fire occurs, dispatch the fire-fighting medium supply equipment corresponding to the most recently selected fire-fighting medium to rush to the vicinity of the fire point, control the docking of the fire-fighting robot with the fire-fighting medium supply equipment, and perform fire-fighting operations according to the determined fire-fighting parameters.
17. The operation method of the intelligent fire protection system according to claim 16, characterized in that: the specific process of calculating the injection angle and the injection flow rate of the injection device in combination with the fire situation includes:

    Obtain the visual image information and infrared image information of the scene environment collected by the multi-eye vision device;

    Preprocess the obtained visual image information and infrared image information respectively;

    Determine the fire area according to the preprocessing results of visual image information and infrared image information;

    According to the fire area, establish the spray curve model, identify the drop point of the water column, and determine the best spray angle and spray flow;

    Analyze the status of the fire equipment in the fire area and determine the best fire fighting location and distance;

    Judge the size of the fire of the burning equipment and select the best spray mode.
18. The operation method of the intelligent fire protection system according to claim 17, characterized in that: the step of preprocessing the visual image information comprises:

    Preprocess the visual image;

    Perform grayscale processing and motion detection on the pre-processed image to determine whether there is a suspicious flame area in the visual image;

    Filtering the area of suspicious flames, extracting the color histogram of the filtered image, extracting image feature values, and performing matching processing to determine the suspicious fire area in the visual image;

    The suspicious fire area is divided and normalized.
19. The operation method of the intelligent fire protection system according to claim 18, characterized in that: the step of preprocessing the infrared image information comprises:

    The infrared image is preprocessed by grayscale and then segmented, the image feature value after segmentation is extracted, and the extracted image feature value is input into the trained neural network model for recognition, and the suspicious fire area of the infrared image is obtained.
20. The operation method of the intelligent fire protection system according to claim 19, characterized in that: the method for determining the fire area is:

    The suspicious fire area obtained after visual image processing is compared with the suspicious fire area obtained after infrared image processing. The overlapping suspicious fire area is regarded as the credible fire area, and the suspicious fire area that does not overlap is regarded as the suspected fire area. The non-suspected fire area is determined to be an area where no fire has occurred.
21. The operation method of the intelligent fire protection system according to claim 17, wherein the method for determining the optimal injection angle and injection flow rate is:

    Take the bottom of the fire area as the target area to establish the injection curve model;

    Obtain the spray image of the fire-fighting robot, and process the spray image to identify the drop point of the sprayed water column;

    Determine the best spray angle according to the coordinate difference between the drop point of the water column and the fire area; adjust the spray flow rate according to the area ratio of the credible fire area and the suspected fire area in the fire area.
22. The operation method of the intelligent fire protection system according to claim 17, characterized in that: the step of analyzing the condition of the fire equipment in the fire area and determining the best fire extinguishing position and distance comprises:

    Preprocess the image of the fire area;

    Extract the feature value of the image of the fire area after preprocessing;

    Input the extracted eigenvalues into the neural network image recognition model to identify the burning equipment;

    Choose the angle with the least obscuration among all directions of the fire equipment as the best extinguishing position;

    According to the proportion of the fire area occupying the entire image, adjust the distance between the fire fighting robot and the fire equipment.
23. The operation method of the intelligent fire protection system according to claim 17, characterized in that: the step of judging the size of the fire of the burning equipment and selecting the best spray mode comprises:

    Establish a sample library containing the fire extinguishing distance of the fired equipment and the basis for the judgment of the fire situation;

    Obtain the fire extinguishing distance and fire judgment basis of the burning equipment from the sample library;

    Compare the area of the burning area with the area of the burning equipment, and judge the fire size of the burning equipment according to the fire judgment basis of the burning equipment;

    According to the size of the fire of the burning equipment, select the best spray mode.
24. The operation method of the intelligent fire protection system according to claim 17, characterized in that: the step of judging the size of the fire of the burning equipment and selecting the best spray mode comprises: when the selected fire fighting medium is dry powder or water mist, controlling The spray coverage area only needs to include the ignition equipment.
25. The operation method according to claim 17, characterized in that: when the three-dimensional model is established, the UAV collects image information of the equipment in the station from different perspectives, and assists the substation fire fighting robot to establish the three-dimensional model of the station;

    During the operation of the fire-fighting robot, the drone collects image information in the substation in real time and assists the robot in determining its position coordinates in the station;

    The drone collects the image information of the burning equipment and determines the location of the fire point;

    According to the position of the robot and the position of the ignition point, control the robot to plan the path and adjust the spray angle;

    During the robot operation, the drone collects images of the location of the fire point in real time, determines the current status information of the fire point and transmits it to the robot, and the robot adjusts the fire fighting strategy according to the received current status information of the fire point.
26. The operation method according to claim 25, characterized in that: the UAV collects the image information of the equipment in the station from different perspectives, and assists the substation fire fighting robot to establish a three-dimensional model of the station, specifically:

    Using the multi-eye vision equipment carried by the fire-fighting robot, using the structural characteristics of the equipment in the station as a constraint, using multi-view reconstruction to obtain a primary model of the entire substation;

    UAVs collect images of substations from different angles, and intensively match them with the established primary models to generate accurate three-dimensional vision models.
27. The operation method according to claim 26, wherein the intensive matching with the established primary model is specifically:

    Operators are used to detect corner points, and then feature descriptors are used to describe the detected corner points, and the image feature points are matched according to the corresponding matching criteria.

Images