Research Of Raman Spectroscopy System In Detecting Water Pollutants Based On Artificial Neural Network

With the fast developing of our country economy, China faces serious water polluted problem. China's development will be confronted with much heavy water crises in future. Enormous destroy take place to the environment while economy develops. GDP maintains in a very high level but the problem is more and more serious. Governing in water pollution is still complex and one to cost the gigantic job. Therefore, in process of governing in water pollution, it is especially important to supervise the pollution source and prevent trouble before it happens. There are many method to detect water contamination, we have designed one kind of the Raman spectroscopy system to check water pollution. Raman spectroscopy is a kind of scattering spectroscopy, which is characterized by the frequency excursion that caused by interactions of molecule. In 60s LASER were invented and were used in Raman spectroscopy. When technology of checking feeble signal is developed and the computer is applies in mangy field, Raman spectroscopy analysis develop in a lot of application field, the Raman spectroscopy analysis is used quilt broadly as one means of the identifying matter. There are many advantages to identify matter with Raman spectroscopy, such as simple system structure, equipment cost lower, the examination scope is widespread. It may realize to the micro sample examination, real-time really carries on the analysis, and may determine the nature of the quota the analysis.In this thesis, we first briefly introduce China's water resources condition, the water pollution situation the water resources. The situation of the water pollution is bad, so checking and governing the pollution of water is urgent. Then we introduce the main theory Raman spectroscopy, Raman spectroscopy occurring mechanism on the classical theory and the quantum theory. Then we introduce the advantage of Raman spectroscopy on investigate matter so as chemistry and material, biological science and so on, then explained the Raman spectroscope design. We simply introduce the Raman spectroscopy the superiority in many aspect such as chemistry, material, biological science and so on. Then we explained the Raman spectroscope system. After that we introduce the artificial neural theory and its model, the elementary theory of artificial nerve network model and it's development, especially the BP nerve network algorithm. We also introduce the limitation of BP algorithm and the improved algorithm, In the fourth chapter introduced the spectroscopy pretreatment, including the compression, smoothing, chirp. Then introduced how the BP algorithm is used in Raman spectroscopy. After that we use the BP network on recognition the Raman spectroscopy, and we get a recognition experimental result, and to changed the related parameter to discuss to the result the influence. Then we use the momentum factor to improve BP network algorithm in the recognition. The result indicated recognizing Raman spectroscopy is accuracy with the BP network. It is suitable for recognizing Raman spectroscopy. In the five chapter we introduce and compare the main database Language. We think SQL is suitable to establish the database of the matter which will pollute the water. Finally we establish the database, which include the matter name, molecular weight, character, and how to guard against it and so on . And we realize to has constructed the pollutant database.