| Aimed at the necessary of gas detection and control in the environment, a gas detection system based porphyrin chemical sensors was developed. The system could be used to identify organic gases by detecting the color change information of porphyrins sensitive chemical sensor before and after reaction with the target gas; images of the porphyrin chemistry sensor array before and after reaction with the analyte were processed with image analysis to get its characteristics, combined with suitable pattern recognition algorithms to identiry the gas. Nine VOCs and NH3 were detected by the proposed system and experiments showed the validity of the system for the qualitative and quantitative identification of VOCs and NH3. The software and pattern recognition were studied here, including:①the signal collection functions: image signal acquisition, including manual acquisition and automatic collection; real-time acquisition and display of temperature, humidity, flow in reaction chamber.②signal processing and analysis: two image processing modes, single image processing and the image sequence processing mode. The single image processing mode consisted of open, simultaneously or separately rotating, cropping, automatic processing, the result printing and so on.③automatic image processing algorithm, according to the features of the porphyrin chemical sensor array chip image, the image filtering, gray, automatic mesh generation based on the projection, segmentation and recognition, feature extraction were realized and the results were displayed visually.④9 VOCs were detected, the image processing results show the validity of the system for the qualitative and quantitative identification of VOCs.⑤4 different concentrations of NH3 were detected, combined with PCA features extraction, two different vectors were used as input neurons to BP neural network and RBF neural networks fot recognition, when using a single time 6 points as the input, the two networks have gained good recognition accuracy. |
PDF Full Text Request