| With the pollution problem getting more and more serious, the intension and the need of protecting the environment by the public are rising. Especially recently, the continuously heavy fog and haze weather make the public to raise the standard of their pollution policy and monitoring, many people even add the checking the latest PM2.5 to their routine and mixed with their lives. At the meantime, we have more concerns about the detection and monitoring of the cooking-oil pollution by the restaurants. So how to make the monitoring more efficient has become a very important study issue. This paper is designed under this situation, and designed a real-time cooking-fume monitoring system based on the data-fusion theory, compared to the traditional monitoring system which has problems of low-precision, high-rate of false alarm and low-robustness, this system has improved a lot.This paper firstly introduced the background of this issue, and analysis the progress and the problem of cooking-fume detection area both domestic and abroad, thus introduced the necessity of designing this system, which is the meaning of choice this topic. Then considering the practicability of this system, based on the microchip of STC51, the hardware of the system series connection of four relative sensors to monitor at the same time; considering the real-time monitoring, we use the GPRS module to transmit the information to the server, and progress the main data-fusion work. And then, this paper use data-fusion theory to process the data collected, and compared the results between different data-fusion theories. The order of describing the system is as follows:system architecture-system hardware design-system software design-main algorithm description-test result comparison.This system compared to the traditional real-time monitoring system, the biggest difference is using multi-sensor to measuring at the same time, thus adding the amount of the information, then using data-fusion to make the judgment, thus improving the precision and the robustness. The data-fusion way of the paper mainly include the innovative-introduced of data-map-fast-location way, BP/SVM neutral network and so on; meanwhile using D-S evidence theory to process the validity, thus making the base of the theory more strong. After that, we cooperated with Environment Department and collect the data from the restaurants, using multi-circle data to test, thus making the system more trustworthy. Through comparison, the precision after data-fusion is much higher than that of single sensor; besides, because of multi-senor, the robustness is much higher too.To measure the performance of the system, we cooperated with the Environment Department to take the data from the restaurant, and using multi-circle data to do the fusing working to compare. The results show that, the deviation of the results is smaller after data-fusion, and the precision is much higher; at the meantime, because of using multi-sensors, the robustness is higher than that of single sensor. Above all, this paper designed a real-time, effective, and high precision real-time cooking fume detection system, and the system has some feasibility and practicality in protecting the environment. |
