經濟發展造成許多環境問題，引貣社會大眾的重視與關注，其中空氣品質優劣不僅影響人類的自身健康，亦造成其他動植物的危害，甚至是塑膠、金屬的鏽蝕，這一切的變化與空氣中的污染物有一定程度的影響，因此行政院環保署空氣品質監測站，進行空氣污染物的監測，以利了解各地區之污染物變化，民眾如欲了解居家附近品質可透過行政院環保署空氣品質監測站之統計資料，但居住於距離測站較遠的地區則需透過模式推算該位置污染物濃度，因此如果能直接在民眾居家環境使用簡易快速的測量方法進行採樣及提供空氣污染物即時濃度給民眾參考，將使得民眾更了解居家附近的空氣品質。 本研究將針對空氣污染物中的懸浮微粒(PM10)、二氧化硫(SO2)、二氧化氮(NO2)、臭氧(O3)及細懸浮微粒(PM2.5)進行污染物濃度變動趨勢預測，以台中市大甲區為研究樣區，進行微型光譜儀光譜強度變化的採樣，將採樣結果與空氣污染物使用類神經網路進行變動趨勢的模擬，研究結果顯示以光譜強度預測污染物之結果，可以發現A 組僅探討光譜與污染物，在無氣象因子輔助下，訓練及預測變化趨勢均不明顯，但在氣象因子輔助後B 組之預測結果較佳，訓練及預測趨勢大致相同，而C 組中僅以氣象因子與污染物進行探討，該組之訓練結果與B 組相似。整體而言，微型光譜儀運用於空氣品質污染物趨勢變動預測，在NO2 及O3 變動趨勢較能夠掌握，但在PM10、PM2.5、SO2 預測時，變動趨勢不明顯及誤差較大，故日後可進一步探討應用在NO2 及O3 濃度與變動趨勢方面。 The economic activities bring the environmental issues, and people aware that air quality is not just affected the health of human being but also animals and plants. The metals and plastics could be corroded by the chemical substance in the air. All thesephenomena are strongly related to the air pollution. However, the Environmental Protection Administration of the Executive Yuan has set up the monitoring stations for air pollution level tracing. People can access these data from website to know the air quality around them. This technology also can estimate the air quality of remote area by simulation model. Therefore, if there is a simple method which can provide the real-time data of air quality to people that will be much batter.This research provides a method to predict the trend of air pollutants, it focus on suspended particles (PM10), sulfur dioxide (SO2), nitrogen dioxide (NO2), ozone (O3) and fine suspended particles (PM2.5). Sampling spectrum intensity changes byMicro-spectrometer in Dajia then analyze data and predict the trend by artificial neural network (ANN). There are three categories of input data. Group A explore the relationship between spectral intensities and air pollutants. Group B explore therelationship between spectral intensities, meteorological factors and air pollutants. Group C explore the relationship between meteorological factors and air pollutants. Three sets of results in group B and group C is better than group A.The result of this research, the Micro-spectrometer is used to forecast air quality by detecting NO2 and O3 and the accuracy are better than using PM10, PM2.5 and SO2.
