1. An intelligent remote monitoring and early warning management system is characterized by comprising a cloud server, a client, a terminal collector and a sensing assembly, wherein the sensing assembly collects relevant parameters of an indoor environment, the parameters at least comprise combustible gas data, water flow data, current/voltage data, temperature data, humidity data, equipment on-off state data, image data and position data, the parameters are transmitted to the terminal collector in a wireless or wired mode, the terminal collector carries out local filtering pretreatment on the received relevant parameters of the indoor environment and transmits the parameters to the cloud server at regular time or when an uploading instruction is received, the cloud server is used for storing the received data according to a time axis, and report data are obtained through cumulative statistics and time-sharing statistics and are transmitted to the client for display; the cloud server is further used for training an intelligent recognition model aiming at a target position according to customer requirements and transmitting the intelligent recognition model to the terminal collector, the terminal collector conducts target recognition preprocessing on image data based on the intelligent recognition model, if a target is recognized, the corresponding image data is transmitted to the cloud server, the cloud server further processes the received image data, generates a target track according to the customer requirements and transmits the target track to the client side for displaying; the cloud server is also used for carrying out trend pre-judging analysis on big data on the received related parameters of the indoor environment, judging whether the related parameters meet early warning conditions or not, and if so, sending early warning information to the client for displaying and controlling the sensing assembly to carry out linkage alarm and early warning processing; the client is also used for the user to input the defense setting parameters and provide the defense setting scheme.

2. The intelligent remote monitoring and early warning management system according to claim 1, wherein the sensing component comprises a combustible gas sensor for collecting and analyzing data of combustible and explosive gas leakage, a water flow sensor for monitoring water flow of the water inlet pipe, a current/voltage sensor for collecting current/voltage data, a temperature sensor for collecting temperature data, a humidity sensor for collecting humidity data, an equipment state sensor for judging whether indoor electric equipment is in a specified working state, a network camera for collecting image data required by intelligent visual analysis, and a position switch sensor for collecting door and window switch signals.

3. The system of claim 1, wherein the terminal collector comprises a core processing module implemented by ARM + NPU + SOC, a wireless communication module connected to the cloud server via TCP/UDP interface, a wired communication module connected to the cloud server via RJ45 network interface/optical fiber interface, an extended interface data collection control module connected to the UART interface, a status display module, and a data exchange module, and the extended interface data collection control module is connected to the sensing component via wire or wireless.

4. The intelligent remote monitoring and early warning management system of claim 3, wherein the expansion interface data acquisition control module comprises an expansion interface and an MCU control module connected with the expansion interface, and the expansion interface comprises an IO input interface, an IO output interface, an ADC input interface, a radio frequency communication circuit, a UART serial port/RS 485 interface, an I2C data bus and a PWM pulse circuit.

5. The intelligent remote monitoring and early warning management system as claimed in claim 3, wherein the wireless communication module comprises a 4G/5G module and a WiFi module.

6. The intelligent remote monitoring and early warning management system according to claim 3, wherein the IO output interface is further used for connecting driving devices, and the driving devices include but are not limited to a water pump, an exhaust fan, light and an oxygenation pump.

7. The intelligent remote monitoring and early warning management system according to claim 1, wherein the cloud server screens the received data before storing the received data according to a time axis, and the screening specifically comprises the following steps:

setting weight coefficients for combustible gas data, water flow data, current/voltage data, temperature data, humidity data, equipment switch state data, image data and position data, and setting positive sample plate data and negative sample plate data;

respectively matching the received data with positive template data, if the data with the weight coefficient higher than the set value are all the positive template data, randomly selecting N data from the received data to be compared with the last group of received data, and if the N data are not consistent, judging the received data to be legal data and performing legal data storage operation; n is an integer greater than 1;

respectively matching the received data with negative template data, if the data with the weight coefficient higher than the set value are all negative template data, randomly selecting N data from the received data to compare with the last group of received data, and if the N data are not consistent, judging the received data as illegal data and performing illegal data storage operation; n is an integer greater than 1;

and if the legal data storage operation amount and the illegal data storage operation reach the set threshold, updating the weight coefficient, the positive sample plate data and the negative sample plate data.

8. The intelligent remote monitoring and early warning management system of claim 1, wherein the cloud server is further used for intelligent personnel statistics, the terminal collector identifies personnel information in image data based on an AIYOLO model algorithm and counts the quantity of personnel, sends the personnel information to the cloud server, and the cloud server records the quantity of personnel, generates false alarm prevention analysis and transmits the analysis to a client for display.

Background

Closed indoor spaces such as kitchens and warehouses have certain potential safety hazards, however, the real-time monitoring can not be carried out by manpower for 24 hours, and the safety problem is difficult to avoid. Taking a kitchen as an example, the kitchen environment in the catering industry is extremely complex, including the use of water, electricity, wind, fire, oil and gas. In order to solve the potential safety hazard problem and optimize related safety management of a kitchen, the traditional methods of manual observation and off-line equipment detection at present are far from reaching an ideal target state, and in order to more effectively find the potential safety hazard problem and avoid accidents, a novel intelligent unattended safety system capable of assisting workers is very necessary to be provided.

Disclosure of Invention

Aiming at the defects in the prior art, the invention aims to provide an intelligent remote monitoring and early warning management system which can self-define defense deployment time according to customer requirements, collect and analyze potential safety hazard parameters of an indoor environment, realize automatic early warning, and solve the technical defects that in the prior art, the labor cost is increased only by detecting gas leakage through artificial olfaction, only threshold value warning is relied on through sensor detection, but no trend analysis exists.

In order to achieve the purpose, the invention adopts the following specific technical scheme:

an intelligent remote monitoring and early warning management system comprises a cloud server, a client, a terminal collector and a sensing assembly, wherein the sensing assembly collects relevant parameters of an indoor environment, the parameters at least comprise combustible gas data, water flow data, current/voltage data, temperature data, humidity data, equipment on-off state data, image data and position data, the parameters are transmitted to the terminal collector in a wireless or wired mode, the terminal collector carries out local filtering pretreatment on the received relevant parameters of the indoor environment and transmits the parameters to the cloud server at regular time or when an uploading instruction is received, the cloud server is used for storing the received data according to a time axis, and report data are obtained through accumulative statistics and time-sharing statistics and are transmitted to the client for display; the cloud server is further used for training an intelligent recognition model aiming at a target position according to customer requirements and transmitting the intelligent recognition model to the terminal collector, the terminal collector conducts target recognition preprocessing on image data based on the intelligent recognition model, if a target is recognized, the corresponding image data is transmitted to the cloud server, the cloud server further processes the received image data, generates a target track according to the customer requirements and transmits the target track to the client side for displaying; the cloud server is also used for carrying out trend pre-judging analysis on big data on the received related parameters of the indoor environment, judging whether the related parameters meet early warning conditions or not, and if so, sending early warning information to the client for displaying and controlling the sensing assembly to carry out linkage alarm and early warning processing; the client is also used for the user to input the defense setting parameters and provide the defense setting scheme.

Preferably, the perception subassembly includes that the combustible gas sensor that carries out data acquisition analysis to the combustible explosive gas leakage, carries out the rivers sensor of rivers monitoring to the inlet tube, the electric current/voltage sensor of collection electric current/voltage data, the temperature sensor of collection temperature data, the humidity transducer of collection humidity data, whether indoor consumer is in appointed operating condition's equipment state sensor, gather the network camera that realizes the required image data of intelligent visual analysis, gather door and window switch signal's position switch sensor.

Preferably, the terminal collector comprises a core processing module realized by ARM + NPU + SOC, a wireless communication module connected with the cloud server through a TCP/UDP interface, a wired communication module connected with the cloud server through an RJ45 network interface/optical fiber interface, an expansion interface data collection control module connected through a UART interface, a state display module and a data exchange module, wherein the expansion interface data collection control module is connected with the sensing assembly through a wire or a wireless connection.

Preferably, the expansion interface data acquisition control module comprises an expansion interface and an MCU control module connected with the expansion interface, and the expansion interface comprises an IO input interface, an IO output interface, an ADC input interface, a radio frequency communication circuit, a UART serial port/RS 485 interface, an I2C data bus and a PWM pulse circuit.

Preferably, the wireless communication module comprises a 4G/5G module and a WiFi module.

Preferably, the IO output interface is further configured to connect to a driving device, where the driving device includes, but is not limited to, a water pump, an exhaust fan, a light, and an oxygenation pump.

Preferably, the cloud server may further perform screening on the received data before storing the received data according to a time axis, where the screening specifically includes the following steps:

setting weight coefficients for combustible gas data, water flow data, current/voltage data, temperature data, humidity data, equipment switch state data, image data and position data, and setting positive sample plate data and negative sample plate data;

respectively matching the received data with positive template data, if the data with the weight coefficient higher than the set value are all the positive template data, randomly selecting N data from the received data to be compared with the last group of received data, and if the N data are not consistent, judging the received data to be legal data and performing legal data storage operation; n is an integer greater than 1;

respectively matching the received data with negative template data, if the data with the weight coefficient higher than the set value are all negative template data, randomly selecting N data from the received data to compare with the last group of received data, and if the N data are not consistent, judging the received data as illegal data and performing illegal data storage operation; n is an integer greater than 1;

and if the legal data storage operation amount and the illegal data storage operation reach the set threshold, updating the weight coefficient, the positive sample plate data and the negative sample plate data.

Preferably, the cloud server is also used for intelligent personnel statistics, the terminal collector identifies personnel information in the image data based on an AIYOLO model algorithm, counts the personnel number, sends the personnel number to the cloud server, and the cloud server records the personnel number, generates an anti-false-alarm analysis and transmits the anti-false-alarm analysis to the client for display.

The invention has the beneficial effects that: the invention can self-define defense deploying time according to customer requirements, collect and analyze potential safety hazard parameters of indoor environment, realize automatic early warning, solve the technical defects that gas leakage can only be detected through artificial olfaction to improve labor cost, and only threshold alarm is relied on through sensor detection but no trend analysis exists in the prior art.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a schematic block diagram of an intelligent remote monitoring and early warning management system according to an embodiment of the present invention;

fig. 2 is a schematic diagram of an embodiment of an intelligent remote monitoring and early warning management system according to an embodiment of the present invention;

fig. 3 is a circuit diagram of a terminal collector provided in an embodiment of the present invention;

fig. 4 is a flow chart of data screening.

In the figure, 1-cloud server, 2-client, 3-terminal collector and 4-perception component.

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. Other embodiments, which can be derived by one of ordinary skill in the art from the embodiments given herein without any creative effort, shall fall within the protection scope of the present invention.

In the description of the present invention, it should be noted that the terms "vertical", "upper", "lower", "horizontal", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience of describing the present invention and simplifying the description, but do not indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and thus, should not be construed as limiting the present invention.

In the description of the present invention, it should also be noted that, unless otherwise explicitly specified or limited, the terms "disposed," "mounted," "connected," and "connected" are to be construed broadly and may, for example, be fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.

As shown in fig. 1-4, the present invention provides an intelligent remote monitoring and early warning management system, which comprises a cloud server 1, a client 2, a terminal collector 3, and a sensing component 4, the sensing component 4 collects relevant parameters of indoor environment, the parameters at least comprise combustible gas data, water flow data, current/voltage data, temperature data, humidity data, equipment switch state data, image data and position data, the information is transmitted to the terminal collector 3 in a wireless or wired manner, the terminal collector 3 performs local filtering pretreatment on the received relevant parameters of the indoor environment, transmits the parameters to the cloud server 1 at regular time or when an uploading instruction is received, the cloud server 1 is used for storing the received data according to a time axis, performing cumulative statistics and time-sharing statistics to obtain report data, and sending the report data to the client 2 for display; the cloud server 1 is further configured to train an intelligent recognition model for a target position according to customer requirements and transmit the intelligent recognition model to the terminal collector 3, the terminal collector 3 performs target recognition preprocessing on image data based on the intelligent recognition model, if a target is recognized, corresponding image data is transmitted to the cloud server 1, the cloud server 1 further processes the received image data, generates a target track according to the customer requirements, and transmits the target track to the client 2 for display; the cloud server 1 is further configured to perform trend pre-judgment analysis on the received relevant parameters of the indoor environment for big data, judge whether the relevant parameters meet an early warning condition, and if the relevant parameters meet the early warning condition, send early warning information to the client 2 for display and control the sensing component 4 to perform linkage alarm and early warning processing; the client 2 is also used for the user to input defense setting parameters and provide a defense setting scheme.

The cloud server 1 screens the received data before storing the received data according to a time axis, the data collected by the sensor becomes more accurate as the use time is longer, useless and repeated data are screened out, and a server storage mechanism is optimized to the maximum extent, wherein the data comprise image data and irregular data acquired by the sensor. The screening method comprises the following specific steps:

setting weight coefficients for combustible gas data, water flow data, current/voltage data, temperature data, humidity data, equipment switch state data, image data and position data, and setting positive sample plate data and negative sample plate data;

respectively matching the received data with positive template data, if the data with the weight coefficient higher than the set value are all the positive template data, randomly selecting N data from the received data to be compared with the last group of received data, and if the N data are not consistent, judging the received data to be legal data and performing legal data storage operation; n is an integer greater than 1;

respectively matching the received data with negative template data, if the data with the weight coefficient higher than the set value are all negative template data, randomly selecting N data from the received data to compare with the last group of received data, and if the N data are not consistent, judging the received data as illegal data and performing illegal data storage operation; n is an integer greater than 1;

and if the legal data storage operation amount and the illegal data storage operation reach the set threshold, updating the weight coefficient, the positive sample plate data and the negative sample plate data.

Preferably, the terminal collector 3 includes a core processing module implemented by ARM + NPU + SOC, a wireless communication module connected to the cloud server 1 through a TCP/UDP interface, a wired communication module connected to the cloud server 1 through an RJ45 network interface/optical fiber interface, an expansion interface data collection control module connected through a UART interface, a status display module, and a data exchange module, and the expansion interface data collection control module is connected to the sensing component 4 through a wired or wireless connection. The expansion interface data acquisition control module comprises an expansion interface and an MCU control module connected with the expansion interface, wherein the expansion interface comprises an IO input interface, an IO output interface, an ADC input interface, a radio frequency communication circuit, a UART serial port/RS 485 interface, an I2C data bus and a PWM pulse circuit.

Preferably, the wireless communication module comprises a 4G/5G module and a WiFi module. The IO output interface is also used for connecting driving equipment, and the driving equipment comprises but is not limited to a water pump, an exhaust fan, light and an oxygenation pump.

Preferably, perception subassembly 4 includes that the combustible gas sensor that carries out data acquisition analysis to the combustible explosive gas leakage, carries out the rivers sensor of rivers monitoring to the inlet tube, gathers the electric current/voltage sensor of electric current/voltage data, gathers temperature data's temperature sensor, gathers humidity data's humidity transducer, whether indoor consumer is in appointed operating condition's equipment state sensor, gathers the network camera that realizes the required image data of intelligent visual analysis, gathers door and window switch signal's position switch sensor.

Wherein, combustible gas sensor is leaked to inflammable and explosive gas and carries out data acquisition analysis one and install a plurality of combustible gas sensor nodes (be no less than 3) and carry out 7 x 24 hours incessant data acquisition for fire prevention and explosion-proof can adopt multiple power supply mode if: battery, low voltage 5-24v power, etc.

The rivers sensor carries out rivers monitoring prevention to the water pipe that comes in leads the occurence of failure at the ageing fracture of water pipe during the vacation, and quantity is installed according to the water pipe that comes in, can adopt multiple power supply mode if: battery, low voltage 5-24v power, etc.

The equipment state sensor is used for judging whether the managed equipment works in a specified state, such as whether a refrigerator works normally after an employee goes off duty, whether a heater works, whether an oxygen increasing device of a fishpond works and the like. The intelligent socket with the state monitoring function is characterized in that the intelligent socket with the state monitoring function is used, equipment needing to be monitored and managed is powered through the intelligent socket, and the equipment needing to be monitored is connected to a state acquisition socket or a terminal.

The network camera acquires image data or video data and sends the image data or the video data to the terminal collector 3, whether mice or cockroaches, flames and smog exist in a current picture is identified, triggering and alarming are conducted to send the current picture to the client side 2, the terminal collector 3 is connected with driving equipment through IO output, and the driving equipment drives the mice or the cockroaches.

The position switch sensor is arranged at the door and window to prevent the door and window from being not well related.

Preferably, the cloud server 1 is further used for intelligent personnel statistics, the terminal collector 3 identifies personnel information in the image data based on an AIYOLO model algorithm, counts the quantity of the personnel, sends the personnel information to the cloud server 1, the cloud server 1 inputs the quantity of the personnel, generates false alarm prevention analysis, and transmits the false alarm prevention analysis to the client 2 for display. The system sends data acquired by IPC to a core processing module to perform AIYOLO pre-trained model algorithm to identify current personnel, obtains the number of the current personnel through statistics, finally sends the number of the current personnel to a cloud server 1 to perform data entry and deep false alarm prevention analysis, and pushes the number of the current personnel to a mobile phone or a computer of a client 2 to display.

If the cloud server trains the intelligent recognition model aiming at the targets such as mice, cats, dogs and the like according to the customer demands, the network camera is a thermal infrared camera which integrates a camera and a pyroelectric infrared sensor to acquire images or videos and thermal infrared images or thermal infrared videos.

Firstly, acquiring a data set of an intelligent recognition model, and acquiring multi-angle images and thermal infrared images of mice, cats and dogs by adopting a thermal infrared camera; the data set was divided into training and test sets in an 8:2 ratio.

Optimizing a network structure of YOLOv3, and performing clustering analysis on target frames in a data set of YOLOv3 by adopting a k-means clustering algorithm to finally determine the optimal number k of candidate frames. The k-means clustering algorithm uses Euclidean distance, the YOLOv3 algorithm adopts overlapping degree to eliminate errors, the distance function is d (box, centroid) ═ 1-IOU (box, centroid), the objective function of clustering isbox is a candidate box, truth is a target real box, and k represents the number of anchors.

The size of the grid is modified according to the size difference of the mouse, the cat and the dog, and the detection effect of the YOLOv3 algorithm is improved from multiple angles.

Training a mouse, cat and dog thermal infrared shape detector by taking an open source deep learning frame Darknet as a basis, taking an improved YOLOv3 network structure as a model and combining methods of dimension cluster analysis, network pre-training and multi-scale training models;

training a mouse, cat and dog feature detector by taking an open source deep learning frame Darknet as a basis, taking an improved YOLOv3 network structure as a model and combining methods of dimension cluster analysis, network pre-training and multi-scale training models;

and inputting the test set into a mouse, cat and dog thermal infrared shape detector and a mouse, cat and dog characteristic detector, verifying and adjusting to finally obtain the intelligent identification model.

The system of the invention is composed of a server, a client 2, a terminal collector 3 and a sensing component 4, and the working principle is as follows: the sensing component 4 of the system collects various information data and sends the information data to the terminal collector 3 through wired and wireless communication, wherein the data comprises the current concentration of combustible gas, water flow, current, temperature, humidity, equipment switch state, visual information and the like, the terminal collector 3 carries out local filtering pretreatment after receiving the information and sends the information to the server for trend prejudgment analysis of big data, and finally, an analysis result is pushed to the client 2(APP, short message and PC end). For example, when a gas pipe of a certain dining room is slightly leaked due to aging, a combustible gas sensor arranged in a kitchen collects data continuously for 7 × 24 hours as usual and then sends the data to a server through a collector for trend analysis, and if the concentration of the combustible gas is gradually increased and more than one sensor data has an ascending trend and meets an alarm condition, the server sends alarm information to a client in advance and correspondingly opens a ventilation system for gas dilution processing.

The server has the functions of data analysis and data forwarding, generally, one server can bear 2000 acquisition terminal accesses and 2000 client terminals 2 accesses, and the server is distributed on cloud services such as Ali cloud, Tencent cloud, Huacheng cloud and the like.

The client 2 mainly has the functions of receiving the alarm information and browsing the history records and displaying the alarm information and the history records to relevant management personnel, wherein the client 2 comprises a mobile phone APP, a PC (personal computer) terminal and a mobile phone short message.

The terminal collector 3 mainly has the functions of collecting various sensors and IO state information, preliminarily filtering useless data and pushing the useless data to a server.

The sensing module mainly comprises various sensors and has the functions of collecting the current state information of equipment or articles which may cause problems and finally sending the information to the terminal collector 3 by using WIFI or RS485, IO and 433/315 Mhz.

The system of the invention functions as follows:

1. custom defense deployment according to a schedule

And setting system defense deployment according to the actual condition of the client, wherein the setting defense deployment can be set at the APP/PC end of the mobile phone. For example, 20:00 evening before morning 06: 00 cut prevention, full day prevention every six days, and the like.

2. Automatic reporting of information and statistical analysis

And as long as the paired sensor system terminal collector 3 can automatically report to the server at regular time to generate daily report data. The user can look up all information in the current month and even the current day at any time.

3. Can count the water consumption, electricity consumption and safety level in the same month

In the monitoring process of the system, data are not discarded, the data are stored in a time shaft, accumulated statistics and time-sharing statistics are carried out through service, and finally relevant reports such as monthly electricity consumption and water consumption are derived.

4. Seamless access to various sensors

Terminal collector 3 can match various sensor access and drive IO output at will if: when a customer wants to monitor the odor of kitchen odor, the exhaust fan is automatically started, and the customer's requirement can be easily met only by one hydrogen sulfide sensor of the matcher. The sensors comprise a combustible gas sensor, a carbon monoxide sensor, a position state switch, a water flow sensor, a current sensor, a temperature and humidity sensor and the like. The driving equipment comprises a water pump, an exhaust fan, light, an oxygenation pump and the like.

5. AI neural network image analysis and irregular data analysis background 80T computational force support.

The matching library is trained according to the target position according to the requirements of a client, for example, as long as a mouse, cockroach, cat, dog, flame, smoke and the like appear in a kitchen system, the picture can be automatically identified, grabbed, evidence can be obtained and pushed to a client mobile phone or a computer terminal for displaying, and the corresponding output IO of the field terminal collector 3 triggers the driving sound to sound and the light warning until the target disappears. The system also has a target track generation function, if a mouse appears in a monitoring area, the system can continuously monitor the coordinates of a target object and generate a track schematic diagram to display the track schematic diagram to the client 2

6. Really realizing unattended operation

The invention solves the technical defects that the labor cost is increased only by detecting gas leakage through artificial olfaction and no trend analysis exists by only depending on threshold value alarm through sensor detection in the prior art, can carry out time trend analysis according to a time axis and collected data so as to forecast in advance, and sends information to related personnel through a cloud server so as to really realize unattended operation.

In light of the foregoing description of the preferred embodiments of the present invention, those skilled in the art can now make various alterations and modifications without departing from the scope of the invention. The technical scope of the present invention is not limited to the contents of the specification, and must be determined according to the scope of the claims.