CN112631301A

CN112631301A - STM 32-based remote control forest fire extinguishing vehicle

Info

Publication number: CN112631301A
Application number: CN202011560625.5A
Authority: CN
Inventors: 朱沛杰; 邢键; 金林仟; 李姝彬
Original assignee: Northeast Forestry University
Current assignee: Northeast Forestry University
Priority date: 2020-12-25
Filing date: 2020-12-25
Publication date: 2021-04-09

Abstract

The invention discloses a remote control forest fire extinguishing vehicle based on STM32, which comprises a fire extinguishing vehicle based on an STM32F103 single chip microcomputer, fully excavates STM32 resources, and utilizes serial ports and I²The C bus carries out the communication between chip and the sensor, and link to each other main control board and WIFI module, make the dolly can realize by the android applet, the bluetooth, cell-phone APP, the many terminals of PC remote control, as the medium through the WIFI module, can give the host computer with the data of sensor, also can receive control command through the host computer, again by the motion of STM32F105 control whole dolly, use the mecanum wheel as the drive wheel of dolly, make the dolly can adapt to multiple topography, realized water jet equipment's automatic control and can judge whether the distance of spun water can put out a fire according to whether flame changes, first step confirms cloud platform direction according to the flame position and aims, the second step as long as do not discern the water smoke in video information, just increase the water pressure or change the angle of elevation of water pump cloud platform, until the water appears in the video information.

Description

STM 32-based remote control forest fire extinguishing vehicle

Technical Field

The invention relates to the field of intelligent control and the technical field of image recognition, in particular to an STM 32-based remote control forest fire extinguishing vehicle.

Background

In the annual fire protection period, due to the drying out of the climate, it is the high frequency period in which fires occur.

Especially in forest zones, once a fire occurs, a great deal of manpower and material resources are needed to put out the fire, and sometimes firefighters even need to be in danger of life.

Aiming at the problems, the existing device is improved, a remote control forest fire extinguishing vehicle based on STM32 is provided, and the remote control forest fire extinguishing vehicle mainly relates to the analysis and identification of complex terrain, the mechanical and intelligent control of fire extinguishing and fire prevention vehicles in forest zones and the comprehensive identification of flame and smoke based on yolov3 deep learning.

Disclosure of Invention

The invention aims to provide a remote control forest fire extinguishing vehicle based on STM32, aiming at the environment with complex landforms such as forest zones, an unmanned intelligent vehicle is used for replacing a firefighter to enter a fire scene for detection, a camera is used for tracking a fire source, a video picture collected by the camera is transmitted back in real time, the fire source is automatically identified by the video picture, a water gun can automatically aim at the fire source according to image information to spray water for fire extinguishing, thus the optimal purpose can be achieved with the least cost, the life safety of the firefighter can be ensured to a certain extent, and the remote control forest fire extinguishing vehicle can be well applied to urban fires due to the adaptability to complex landforms, the speed of the fire extinguishing vehicle can be controlled in real time by an upper computer, the angle of a cradle head can be adjusted, the condition in front of the current fire extinguishing vehicle can be seen by the camera, and the video information can be transmitted back to the upper computer for image processing in, the stability is achieved, the cost is greatly saved, and the problems in the background art are solved.

In order to achieve the purpose, the invention provides the following technical scheme: remote control forest fire extinguishing vehicle based on STM32, include the fire extinguishing vehicle based on STM32F103 singlechip, the fire extinguishing vehicle based on STM32F103 singlechip includes WIFI transmission module, the host computer, the camera module, yolov3 degree of deep learning module, motion control module and drive module, the fire extinguishing vehicle based on STM32F103 singlechip passes through WIFI module and host computer wireless communication connection, the camera module, power module, display terminal and host computer electromechanical connection, yolov3 degree of deep learning module and host computer electromechanical connection.

Further, the WIFI transmission module comprises a WIFI wireless remote communication system and an upper computer software app, the WIFI wireless remote communication system is electrically connected with the upper computer software app, and the upper computer software app is electrically connected with an upper computer.

Further, the yolov3 deep learning module comprises yolov3 deep learning.

Furthermore, the motion control module comprises an ultrasonic sensor, an infrared tracking sensor and small wheels, the ultrasonic sensor adopts an HC-SR04 module, the infrared tracking sensor adopts a TCRT5000 infrared sensor and is mainly used for controlling the rail car to run along a fixed rail route or a preset symbolic path to finally find a running target, the ultrasonic sensor adopts an HC-SR04 module, the professional ultrasonic module is stable in performance, high in accuracy, small in blind area and high in feedback speed, the sensing angle is not more than 15 degrees, the detection distance is 2 cm-450 cm, the accuracy can reach 0.3cm, the blind area is only 2cm and can be used for preventing obstacles and following functional modules of a train, the small wheels adopt Mecanum wheels, and the HC-SR04 module and the TCRT5000 infrared sensor module are electrically connected with an upper computer.

Further, drive module includes cloud platform steering wheel, step motor module and water pump control, and drive module adopts MPU6050 module, and MPU 6050's main function is the gesture fine setting function when realizing the fire prevention car motion, and MPU 6050's the complete check sensing range of angular velocity does: 250, ± 500, ± 1000 and ± 2000 °/sec (dps), can accurately track fast and slow motions, and the user-programmable accelerator full-cell sensing range is: 2g, 4g, 8g and 16g, product transmission can be achieved through I up to 400kHz²C or the SPI (MPU6000, because MPU6050 does not have the SPI) up to 20MHz, the pan-tilt steering wheel adopts MG996R steering wheel pan-tilt module, water pump control L9110S water pump drive module for the work of drive water pump, if discernment has the fire source when carrying out image analysis discernment to the video, will transfer to the assigned position at the pan-tilt after, go to spray water with L9110S drive module.

Furthermore, the stepping motor module adopts LV8731V, a PWMV current control stepping motor driving chip is arranged in the stepping motor module, the stepping motor driving chip comprises 2 driving modes, namely a DCM mode and an STM mode, LV8731V rated working voltage is 9-32V, and output peak current is 2.5A; an ultra-low on-resistance (0.55 omega) with built-in output short-circuit protection and abnormal state warning output functions; 4-gear electrifying current modes and 4 subdivision modes can be set.

Furthermore, the camera module adopts an OV7725 module, and utilizes an OV7725 module, namely an OV7725 module, which mainly provides pins such as D0-D7 digital pixel output signals, a VSYNC frame synchronization signal, an HREF row synchronization signal, a PCLK pixel synchronization signal, an SCCB bus (SDA, SCL) register configuration signal, an XCLK operation clock signal and the like, a 24M active crystal oscillator is designed on the XCLK pin as a clock signal for 0V7725 to operate, and the SCCB bus is similar to the I2C bus, so that 2 pull-up resistors of 10K are designed on the SDA and SCL pins to be connected with 2 general IOs of the C51 singlechip, and parameters such as different image data format outputs, resolution adjustment, automatic exposure control, automatic gain control, automatic white balance, saturation, brightness, contrast and the like of the camera are configured by utilizing a STM32F103VET6 general purpose analog IO B bus protocol.

Compared with the prior art, the invention has the beneficial effects that:

1. according to the remote intelligent forest fire extinguishing vehicle based on STM32, the upper computer can control the speed of the fire prevention vehicle in real time and adjust the angle of the steering engine pan-tilt, the camera can also see the current situation of the fire prevention vehicle in front of the fire prevention vehicle, and the video information can be transmitted back to the upper computer in real time for image processing, so that the stability is high, and the cost is greatly saved.

2. The invention provides a remote intelligent forest fire fighting vehicle based on STM32, wherein a motion control module comprises an ultrasonic sensor, an infrared tracking sensor and a small wheel, the ultrasonic sensor adopts an HC-SR04 module, the infrared tracking sensor adopts a TCRT5000 infrared sensor module, the small wheel adopts a Mecanum wheel, the HC-SR04 module and the TCRT5000 infrared sensor module are electrically and mechanically connected with an upper computer, the fire fighting vehicle can realize basic obstacle avoidance and tracking functions, the obstacle avoidance adopts the HC-SR04 module, when an obstacle is encountered, the fire fighting vehicle can change a path, the tracking adopts the TCRT5000 infrared sensor module, the fire fighting vehicle can run along a planned path, one-dimensional following running can also be realized by using the sensor, the Mecanum wheel for preventing tires of the train can realize advancing of complex landforms and climbing high difficulty and the like, and therefore, the remote intelligent forest fire fighting vehicle can be used for different environments, and different driving strategies can be provided for the current terrain in different environment positions of the fire-proof vehicle, the fire-proof vehicle increases the moving speed of the fire-proof vehicle by adjusting the rotating speed of the stepping motor LV8731V, the time in the journey is reduced, in the control aspect, the fireproof vehicle can be controlled by various modes such as Android APP, a PC machine and the like, the connection mode with the upper computer is WIFI or Bluetooth, the camera module is used for shooting video data in front of the train, the video data are sent to the upper computer through the WIFI module, the video data shot by the camera can be returned in real time through the method in the upper computer and are sent to the upper computer for processing, the position of the fire source is obtained through the automatic algorithm of the fire truck, then transfer MG996R steering wheel cloud platform module and MPU6050 module to adjust the angle of cloud platform, transfer L9110S module at last and control the water pump and spray water and just can realize putting out a fire under the different condition.

3. The invention provides a remote intelligent forest fire extinguishing vehicle based on STM32, which carries out data communication between an upper computer and the fire extinguishing vehicle through a WIFI module, the fire extinguishing vehicle can adapt to running of complex environment zones such as forests and the like, has multiple functions of tracking, obstacle avoidance and the like, road condition image data are returned to the upper computer in real time through a camera, the upper computer continuously trains through yolov3 deep learning, identifies fire sources and fire information, returns the data to the fire extinguishing vehicle for fire extinguishing control, completes fire extinguishing tasks and resource saving, meanwhile, the upper computer software app is used for carrying out instruction control on the fire extinguishing vehicle, the WiFi module is used for wireless data transmission, the motion of the fire extinguishing vehicle is controlled through the app, the real-time road condition of the fire extinguishing vehicle is transmitted to an upper computer information processing system, the image data are conveniently acquired, the yolov3 deep learning module comprises the lov3 deep learning, the fire sources are judged through analysis, the fire source information can be obtained through continuous training, the size of the fire source and the fire behavior can be determined through analyzing image data transmitted by the fire-proof vehicle, and an instruction is continuously sent to the fire-proof vehicle to extinguish the fire.

Drawings

FIG. 1 is a general structural diagram of a remote control forest fire extinguishing vehicle based on STM 32;

FIG. 2 is a STM32F103 schematic diagram of the remote control forest fire extinguishing vehicle based on STM 32;

FIG. 3 is a schematic diagram of an MPU6050 module of an STM 32-based remote control forest fire extinguishing vehicle;

FIG. 4 is a schematic diagram of a remote control forest fire extinguishing vehicle TCRT5000L infrared tracking module based on STM 32;

FIG. 5 is a schematic diagram of an HC-SR04 ultrasonic module of the STM 32-based remote control forest fire extinguishing vehicle;

FIG. 6 is a schematic diagram of a remote control forest fire extinguishing vehicle L9110S water pump driving module based on STM 32;

FIG. 7 is a deep learning flow chart of image data extraction of a remote control forest fire extinguishing vehicle yolo3 based on STM 32;

FIG. 8 is a block diagram of a power supply system of the STM 32-based remote control forest fire extinguishing vehicle;

FIG. 9 is a flow chart of implementation of train protection functions of the STM 32-based remote control forest fire extinguishing vehicle.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Example one

Referring to fig. 1-9, a remote control forest fire fighting vehicle based on STM32 comprises a fire fighting vehicle based on an STM32F103 single chip microcomputer, the fire fighting vehicle based on the STM32F103 single chip microcomputer comprises a WIFI transmission module, an upper computer, a camera module, a yolov3 deep learning module, a motion control module, a driving module, a stepping motor module and an L9110S water pump driving module, the fire fighting vehicle based on the STM32F103 single chip microcomputer is in wireless communication connection with the upper computer through the WIFI module, the camera module, a power supply module and a display terminal are electrically connected with the upper computer, the yolov3 deep learning module is electrically connected with the upper computer, the yolov3 deep learning module comprises yolov3 deep learning, the invention is controlled by an STM32F1 main control board and mainly divided into a motion control system of a fire prevention vehicle, a fire identification system and a fire control system, a 50 battery pack is used for supplying power to the whole system after passing through a main switch on the main board, and meanwhile, a WC70H (DC _ DC5 AM V) + AM117, and supplying power to different modules to enable each module to work under a proper working voltage.

Comprehensive identification of flame and smoke based on yolov3 deep learning: firstly, screening, extracting a network darknet53 from trunk features, extracting features, using a Residual error network Residual, downsampling an input picture, continuously compressing the width and the height, continuously expanding the number of channels, decomposing the channels, and then performing convolution operation, wherein each convolution part of the darknet53 uses a specific darknetv 2D structure, l2 regularization is performed during each convolution, Batchnormalization standardization and LeakyReLU are performed after the convolution is completed, a common ReLU sets all negative values to zero, and a Leaky ReLU gives a non-zero slope to all negative values, which can be expressed in a mathematical way that:

in the characteristic utilization part, yolo3 extracts a plurality of characteristic layers for target detection, three characteristic layers are extracted in total, the three characteristic layers are positioned at different positions of a trunk part darknet53 and are respectively positioned at an intermediate layer, a middle lower layer and a bottom layer, the three characteristic layers are subjected to 5 times of convolution processing, after the processing, the processed part is used for outputting a prediction result corresponding to the characteristic layer, the processed part is used for carrying out deconvolution UmSamplling 2d and then combined with other characteristic layers to divide the whole graph into grids of 13x13, 26x26 and 52x52 respectively by the 3 characteristic layers of yolo3, each grid point is responsible for detecting one area, each grid point is added with x _ fset and y _ fset corresponding to the grid point, the added result is the center of a prediction frame, then the length and the width of the prediction frame are calculated by combining a priori frame with h and w, the position of the whole prediction frame can be obtained, and the final prediction structure is further subjected to score sorting and non-maximum suppression screening to obtain a result, secondly, training, calculating parameters required by loss, judging the position of the real frame in the picture through pred and target, and judging which grid point the real frame belongs to for detection; judging which prior frame has the highest coincidence degree with the real frame; calculating how much prediction result should be obtained for the grid point to obtain a real frame; all real frames are processed as above; the method comprises the steps of obtaining a prediction result which the network should have, comparing the prediction result with an actual prediction result, and finally outputting the content of the network, namely the prediction frame and the type thereof corresponding to each grid point of three characteristic layers, namely the positions, confidence degrees and the types thereof corresponding to three prior frames on each grid point after the three characteristic layers respectively correspond to the grids of which pictures are divided into different sizes.

Example two

Referring to fig. 1-9, the remote control forest fire extinguishing vehicle based on the STM32 comprises a fire extinguishing vehicle based on an STM32F103 single chip microcomputer, the fire extinguishing vehicle based on the STM32F103 single chip microcomputer comprises a WIFI transmission module, an upper computer, a camera module, a yolov3 deep learning module, a motion control module, a driving module, a stepping motor module and an L9110S water pump driving module, the fire extinguishing vehicle based on the STM32F103 single chip microcomputer is in wireless communication with the upper computer through the WIFI module, the camera module, a power supply module, a display terminal is in electrical connection with the upper computer, the yolov3 deep learning module is in electrical connection with the upper computer, the remote control forest fire extinguishing vehicle based on the STM32 comprises the fire extinguishing vehicle based on the STM32F103 single chip microcomputer, the fire extinguishing vehicle based on the STM32F103 single chip microcomputer comprises the WIFI transmission module, the upper computer, the camera module, the yolov3 deep learning module, the motion control module, the driving module, the stepping motor module and the L9110S water pump driving module, the camera module, the power supply module and the display terminal are electrically connected with an upper computer, the yolov3 deep learning module is electrically connected with the upper computer, and the stepping motor module and the tracking module are additionally arranged on the vehicle bodyThe fire-fighting vehicle comprises a block, an ultrasonic module, an infrared sensing module, a water pump module, a camera module and a steering engine holder module, wherein a power supply system is used for independently supplying power to a vehicle body and each module, an STM32F103 development board is used as a main control chip, an upper computer is used for image recognition and command control of a fire-fighting vehicle, the fire-fighting vehicle receives road condition and fire information through a camera and feeds back the road condition and the fire information to the upper computer in real time, the judgment of the information is controlled by the upper computer through water spraying, a data processing server learns deeply through yolov3 of the upper computer and continuously trains to master the fire information, when the fire-fighting vehicle runs, the image data is fed back by the camera in real time, after the image data is received by the upper computer, yolov3 deep learning is carried out to search for the size; the data processed in the remote control unit includes: the fire-proof vehicle real-time image data, the fire-proof vehicle terrain and ground condition data, the ultrasonic ranging data, the fire-proof vehicle condition data and the fire-proof vehicle response information are used for controlling a fire-proof vehicle through a wireless signal remote end, each instruction corresponding to each action is predefined in upper computer programming software, different voltage changes among the instructions are identified through an STM32F103 chip and a C51 program in the fire-proof vehicle, so that the fire-proof vehicle is controlled to make a corresponding mechanical motion response, the fire-proof vehicle response information is fed back, a sensor network is a WIFI communication network, the conditions of the fire-out vehicle can be detected remotely and at low power consumption and low cost, various actions of the fire-out vehicle are controlled through upper computer software app, fire source information can be obtained through continuous training through deep learning of yolv 3, the sizes of a fire source and the fire behavior are determined through analyzing image data transmitted by the fire-proof vehicle, and the instructions are, based on STM32F105 main control chip, the resource of STM32 has been fully excavated, serial ports, I are utilized²The C bus carries out the communication between chip and the sensor, and link to each other main control board and WIFI module, make the dolly can realize by the android applet, the bluetooth, cell-phone APP, PC multi-terminal remote control, in addition, as the medium through the WIFI module, can give the host computer with the data of sensor, also can receive control command through the host computer, again by the motion of the whole dolly of STM32F105 control, the dolly uses the mecanum wheel as the drive wheel of dolly, make littleThe vehicle can adapt to various terrains, in a flame identification part, image identification processing is carried out on video information by adopting a yolov3 deep learning-based method, flame information and smoke information in a video transmitted back by a camera can be identified, an image identification algorithm based on mixing of flame and water mist is provided, namely, the flame and the smoke are used for carrying out comprehensive judgment on the flame, automatic control of a water spraying device is realized, whether the distance of sprayed water can extinguish a fire or not can be judged according to whether the flame changes or not, the unmanned intelligent function of a fire extinguishing vehicle is improved, the specific function is realized by two-step automatic control, the direction of a holder is determined to aim according to the position of the flame in the first step, and as long as the water mist is not identified in the video information in the second step, the elevation angle of a water pressure is increased or the water pump is changed until the water mist appears in.

In summary, the following steps: the remote intelligent forest fire fighting vehicle based on STM32 provided by the invention can realize basic obstacle avoidance and tracking functions, the obstacle avoidance uses an HC-SR04 module, when an obstacle is encountered, the fire fighting vehicle can change a path, the tracking uses a TCRT5000 infrared sensor module, the fire fighting vehicle can run along a planned route, one-dimensional following running can also be realized by using the sensor, the fire fighting vehicle can run with people in one dimension, the tyre of the fire fighting vehicle uses a Mecanum wheel, the fire fighting vehicle can realize the advancing of complex landforms, climbing and other high difficulty running, therefore, the fire fighting vehicle can cope with fire road conditions in different environments, different running strategies can be provided for the current landforms in different environment positions, the moving speed of the fire fighting vehicle can be increased by adjusting the rotating speed of a stepping motor LV8731V, the time in the course can be reduced, the fire fighting vehicle can realize the control in various modes such as Android APP, PC and the like on a control level, and the connection mode with the host computer has WIFI and bluetooth two kinds, the camera module is used for shooting the video data in the front of the train, and give the video data to the host computer through the WIFI module, can realize the video data that the camera was shot through aforementioned method in the host computer, return in real time, give the host computer to handle, the fire prevention car is automatic to use the algorithm to obtain the position of fire source, then transfer MG99 996R steering wheel cloud platform module and MPU6050 module to adjust the angle of cloud platform, call L9110S module at last and control the water pump and spray water and just can realize putting out a fire under different circumstances, can be through the real-time speed of the real-time control of host computer train and the angle of regulation steering wheel cloud platform, can see the current situation in the front of the train of preventing through the camera, and can transmit video information back to the host computer in real time and image processing, not only there is stability but also.

It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be able to cover the technical solutions and the inventive concepts of the present invention within the technical scope of the present invention.

Claims

1. Remote control forest fire extinguishing vehicle based on STM32, including the fire extinguishing vehicle based on STM32F103 singlechip, its characterized in that: fire extinguishing vehicle based on STM32F103 singlechip includes WIFI transmission module, host computer, camera module, yolov3 degree of deep learning module, motion control module and drive module, and the fire extinguishing vehicle based on STM32F103 singlechip passes through WIFI module and host computer wireless communication connection, and camera module, power module, display terminal and host computer electrical connection, yolov3 degree of deep learning module and host computer electrical connection.

2. An STM 32-based remote control forest fire fighting vehicle as defined in claim 1, wherein: the WIFI transmission module comprises a WIFI wireless remote communication system and an upper computer software app, the WIFI wireless remote communication system is electrically connected with the upper computer software app, and the upper computer software app is electrically connected with an upper computer.

3. An STM 32-based remote control forest fire fighting vehicle as defined in claim 1, wherein: the yolov3 deep learning module comprises yolov3 deep learning.

4. An STM 32-based remote control forest fire fighting vehicle as defined in claim 1, wherein: the motion control module comprises an ultrasonic sensor, an infrared tracking sensor and a small wheel, the ultrasonic sensor adopts an HC-SR04 module, the infrared tracking sensor adopts a TCRT5000 infrared sensor module, the small wheel adopts a Mecanum wheel, and the HC-SR04 module and the TCRT5000 infrared sensor module are electrically connected with an upper computer.

5. An STM 32-based remote control forest fire fighting vehicle as defined in claim 1, wherein: the drive module includes cloud platform steering wheel, step motor module and water pump control, and drive module adopts the MPU6050 module, and the cloud platform steering wheel adopts MG996R steering wheel cloud platform module, water pump control L9110S water pump drive module.

6. An STM 32-based remote control forest fire fighting vehicle as defined in claim 1, wherein: the stepper motor module employs LV 8731V.

7. An STM 32-based remote control forest fire fighting vehicle as defined in claim 1, wherein: the camera module adopts an ov7725 module.