Summary of the invention
The present invention is to solve the deficiency of existing fire monitoring early warning system, provides a kind of fire based on ZigBee technology
Monitoring and warning system can greatly improve the timeliness and accuracy of fire monitoring early warning.
To achieve the goals above, the technical solution adopted by the present invention are as follows:
A kind of fire monitoring early warning system based on ZigBee technology, including central processing unit, PAN coordinator, multiple ZigBee
Monitoring network and display module;The ZigBee monitoring network includes FFD node and isomery RFD node;The isomery RFD node
Including smoke sensor device, temperature sensor and imaging sensor;
The FFD node, it is described for the environmental information for being arranged in the isomery RFD node acquisition in monitored place to be sent to
PAN coordinator, the environmental information include smokescope value, temperature value and image rgb value;
The PAN coordinator, the network topology structure for coordinating to be formed by the multiple ZigBee monitoring network, will be each
The environmental information at ZigBee monitoring network is sent to the central processing unit;
The central processing unit, it is described monitored for being determined using multi-sensor Fusion Algorithm according to the environmental information
The fire condition in place;
The display module, for being monitored the environmental information in place described in real-time display.
Further, the central processing unit, is specifically used for:
The smokescope value, temperature value and the image rgb value in the preset time period before current time are carried out respectively special
Value indicative divides, and obtains real time correlation feature vector;
The smoke sensor device, temperature sensor and imaging sensor are obtained in different fire ranks using Apriori algorithm training
Standard association feature vector under section;
By each element of the real time correlation feature vector respectively from the standard association feature under the different fire stages to
The corresponding element of amount is compared, and determines the fire condition in the monitored place.
Further, the central processing unit, is specifically used for:
The smokescope value is compared with preset first concentration threshold, the second concentration threshold and third concentration threshold,
Determine the characteristic value of the smokescope value;
The temperature value is compared with preset first temperature threshold, second temperature threshold value and third temperature threshold, is determined
The characteristic value of the temperature value;
Described image rgb value is compared with preset first RGB threshold value, the 2nd RGB threshold value and the 3rd RGB threshold value, is determined
The characteristic value of described image rgb value.
Further, the different fire stages include: that Initial Stage of Fire stage, fire controllable stage and fire are uncontrollable
Stage processed.
Further, the central processing unit, is specifically used for:
The real time correlation feature vector is merged from the standard association feature vector under different fire stages respectively, is obtained
One assemblage characteristic vector;
Current sample element is chosen in the assemblage characteristic vector, calculate the current sample element and the assemblage characteristic to
The KNN distance of other elements in amount, using the KNN apart from the standard association feature vector where the smallest element as most connecing
Proximad measure;
Using described closest to fire stage corresponding to vector as current fire condition.
Further, the network topology structure are as follows: the PAN coordinator is located at network center, the FFD node and institute
Isomery RFD Node distribution is stated around the PAN coordinator.
Further, the display module includes wirelessly controlled LED display screen.
Beneficial effects of the present invention:
A kind of fire monitoring early warning system based on ZigBee technology provided by the invention is assisted by setting central processing unit, PAN
It adjusts device, multiple ZigBee monitoring networks and display module to carry out data information collection and focuses on, monitoring information can be improved
Accuracy and reliability;Wherein PAN coordinator, tunable whole network and the communication with central processing unit have information
Transmission speed is fast, the strong feature of timeliness;Also, the tree-network topology knot formed by using FFD, RFD and PAN node
Structure further decreases the power consumption of entire fire monitoring early warning system, avoids because of part using the lower feature of RFD node power consumption
RFD node failure and the paralysis for leading to entire fire monitoring early warning system, improve the reliability and stability of whole system.
Specific embodiment
The present invention is further described with reference to the accompanying drawings and detailed description:
As shown in Figure 1, the present invention provides a kind of fire monitoring early warning system based on ZigBee technology, the fire monitoring early warning
System includes: central processing unit 11, PAN coordinator 12, multiple ZigBee monitoring networks 13 and display module 14;The ZigBee
Monitoring network 13 includes FFD node and isomery RFD node;The isomery RFD node include smoke sensor device, temperature sensor and
Imaging sensor;
The FFD node is used to the environmental information for being arranged in the isomery RFD node acquisition in monitored place being sent to the PAN
Coordinator 12, the environmental information include smokescope value, temperature value and image rgb value;
The PAN coordinator 12 is used for the network topology structure for coordinating to be formed by the multiple ZigBee monitoring network, will be each
The environmental information at ZigBee monitoring network 13 is sent to the central processing unit 11;
The central processing unit 11 is used to determine described supervised according to the environmental information using multi-sensor Fusion Algorithm
The fire condition of geodetic point;
The display module 14 is for being monitored the environmental information in place described in real-time display.The display module includes wireless controlled
LED display processed.
Specifically, central processing unit 11 uses the treatment process of multi-sensor Fusion Algorithm specifically: step 1 is divided
The smokescope value, temperature value and image rgb value in the other preset time period to before current time carry out characteristic value and draw
Point, obtain real time correlation feature vector;
The specific implementation procedure of the step are as follows: by the smokescope value and preset first concentration threshold, the second concentration threshold and
Third concentration threshold is compared, and determines the characteristic value of the smokescope value;By the temperature value and preset first temperature
Threshold value, second temperature threshold value and third temperature threshold are compared, and determine the characteristic value of the temperature value;By described image RGB
Value is compared with preset first RGB threshold value, the 2nd RGB threshold value and the 3rd RGB threshold value, determines the spy of described image rgb value
Value indicative.
Step 2 obtains the smoke sensor device, temperature sensor and imaging sensor and exists using Apriori algorithm training
Standard association feature vector under different fire stages;
In general, causing the fuel of fire when being burnt with oxygen generation chemical reaction, comburant can be generated, is sent out simultaneously
Light, heat and other gases out.For example, Initial Stage of Fire stage, the CO gas generated when burning insufficient;And fire is gradually climing
Delay, the CO that when full combustion of fuel generates2Gas etc..Under normal conditions, the color of flame is partially red, however, with around ring
The raising of border temperature, the color of flame can also change.That is, under different fire stages, temperature, the smokescope of environment
And the rgb value of the brightness of image Flame or image entirety is all different.This step exactly utilizes Apriori algorithm to excavate
When fire occurs, existing association between temperature, smokescope and image rgb value, to obtain under different fire stages
The standard association feature vector being made of three characteristic value elements.
The Apriori model that can be used in SPSS Clementine data mining software in practical applications is associated
Property analysis.By Apriori model select respectively this 3 variables of temperature in different fire stages, smokescope and image rgb value because
The high-frequency characteristic value combination of son.Then, the high-frequency characteristic value of initial stage is combined the standard association feature as initial stage
Vector;Fire be can control into standard association of the high-frequency characteristic value combination in stage (can also claim the fire development stage) as the stage
Feature vector;The high-frequency characteristic value in fire uncontrollable stage (fierce combustion phases can also be claimed) combine the mark as the stage
Quasi- linked character vector.
Step 3, by each element of the real time correlation feature vector respectively from the standard under the different fire stages
The corresponding element of linked character vector is compared, and determines the fire condition in the monitored place.
The specific implementation procedure are as follows: by the real time correlation feature vector respectively from the standard association under different fire stages
Feature vector merges, and obtains an assemblage characteristic vector;Current sample element, meter are chosen in the assemblage characteristic vector
The KNN distance for calculating other elements in the current sample element and the assemblage characteristic vector, by the KNN apart from the smallest
Standard association feature vector where element is used as closest to vector;Using it is described closest to fire stage corresponding to vector as
Current fire condition.
The network topology structure of PAN coordinator and multiple ZigBee monitoring networks in the fire monitoring early warning system are as follows:
The PAN coordinator is located at network center, the FFD node and the isomery RFD Node distribution in the week of the PAN coordinator
It encloses.
As shown in Fig. 2, the gray shade node that the oblique line for being located at network center is filled is PAN coordinator, it is located at PAN and coordinates
The dark shaded nodes of 4 grids filling around device are FFD node, and the isomery RFD node around FFD node
(N1, N2, W, P).Wherein N1 and N2 respectively indicates two distinct types of smoke sensor device, and W indicates temperature sensor and P
Indicate imaging sensor.
Specifically, the ionic formula that the smoke sensor device used in the present embodiment produces for Pu En Science and Technology Ltd. of Shenzhen
The intelligent temperature that the gas fume sensor MQ-2 of smoke sensor device HIS-07, TORO company production, DALLAS company of the U.S. produce
Spend sensor DS18B20 and ccd image sensor.
When a fire, the isomery RFD node of corresponding position sends a signal to neighbouring FFD node, the FFD immediately
Information is simultaneously reached PAN coordinator by node sinks information, and PAN coordinator is by the information and the location information of isomery RFD node
Central processing unit is forwarded to be handled.
From the network topology structure figure it is found that RFD node is only communicated with FFD node, therefore RFD node is searching out
Suitable FFD node and after sending data, can disconnect the connection with FFD node, into battery saving mode, electric quantity consumption is non-immediately
It is often small.The embodiment of the present invention exactly utilizes this point, so that the quantity of RFD node is much larger than the quantity of FFD node, thus very great Cheng
The electric quantity consumption that whole system is reduced on degree achievees the purpose that save electricity on the basis of guaranteeing that information is effectively transmitted.
As shown in the above, a kind of fire monitoring early warning system based on ZigBee technology provided by the invention, has
Following benefit: the first, the network connection data of ZigBee transmission 128kbps takes around 30 milliseconds, and real-time is good;The second, may be used
Quickly to connect, and the mode matched by using RFD, FFD and PAN node, utilize the spy of RFD node low power consumption
Sign, can maintain prolonged battery life;Third passes through using electricity wisely, guarantees effective operating time of each RFD node, keeps away
Exempt from the paralysis for leading to entire fire monitoring early warning system because of part RFD node because of not enough power supply frequent failure, improves entire system
The reliability and stability of system;4th, ZigBee monitoring network can configure tens of thousands of a nodes, the network expansion of whole system
It is good.
The accuracy and reliability of monitoring information can be improved in the present invention, have information transfer rate it is fast, timeliness by force and
The characteristics of long-range monitoring, human cost is greatly reduced, administrator can know fire information at any time, and reaction saves treasured rapidly
The expensive time loses caused by reduction fire as much as possible.
The embodiment of the above, only presently preferred embodiments of the present invention, not limits practical range of the invention,
Therefore all equivalent change or modifications done according to structure, feature and principle described in the invention patent range, it should be included in this hair
In bright claim.