Research Of Gas Monitoring System Based On Porphyrin-based Chemical Sensor

With the social progress, modern industry development and human living standard improving, the human survival atmospheric environment becomes more and more complex and in which there are a variety of harmful substances seriously endangering human health. In recent years, a large number of scholars at home and abroad have done extensively study in the field of gas detection. Electronic nose has been a mature product abroad and it can achieve rapid and accurate detection in a variety of occasions; internal research is in a relatively backward situation, and the majority is limited to laboratory research.In this paper, a gas detection system based on prophyrin chemical sensors was designed. The system could be used to identify gases by detecting the color change information of porphyrins sensitive chemical sensor before and after reaction with the target gas. Sensor image signal processing algorithms in the gas detection system was studied and the detecting software system of the upper machine was developed. The characteristic study completed are as follows:1. Detection software system was designed and completed. The software system was composed of lower hardware control module, image acquisition module, image signal processing and analysis module, and database management module. And human-computer interaction interface was established, which realized monitoring and controlling of the detection process and the detecting parameters, controlling and displaying of image signal processing and analysis and management of the detected data.2. Image processing technology in the sensor image processing applications was studied, including image pre-processing (including geometric correction, image gray processing, image filtering), sensitive spots identification, feature extraction. A mathematical morphology-based filtering method was proposed, which could effectively filter the bar pollution interference brought in the production process of the image sensor. An automatic identification algorithm based on gridding of sensitive spots was designed; Experiments showed the identification of weak signals could be realized quickly, effectively and stably, which laid the foundation of the automation and intelligence of the entire detecting process. The sensitive spots of the effective area were selected according to their color distribution, eliminating the interference of the uneven color of the sensitive spots effectively.3. Neural networks in the application of characteristic signal recognition was studied. Multi-network hierarchical identification was used to realize the qualitative and quantitative gas identification respectively; Better recognition results were obtained by training with small samples. Experiments showed that the neural network trained by the system had a high accuracy.