Research For Extraction Of Smaze Using Multi-spectral Data Of HJ-1

Smaze is a kind of atmospheric pollution phenomenon that contains toxic gases and particulate matter from material burning, which can cause enormous harm to the environment and human’s health. This study focuses on Nanjing and its surrounding areas in China, which has suffered from adverse effects of smaze because of outdoor straw burning, and summarizes the formation causes, harm, chemical composition, physical properties and optical properties, etc of smaze. By using the multispectral data of Environmental Satellite-1,the spectral characteristics of smaze, fog-haze, dust-haze is extracted, the spectral contrast of different kinds of smaze is compared, the influence of underlying surface features to smoke haze spectrum is analysed, and particularity of spectrum of smaze is explored. After that, a new index named Smaze Index (SI) was established for Environmental Satellite-1data. It is suggested that SI can retrieve the smaze information from the satellite images quickly and accurately, by using the data of environmental and meteorological department and multi-temporal haze images.The paper’s main research results are as follows:(1) The spectral differences between smaze, fog-haze, dust-haze are analyzed.In blue band(0.43～0.52μm)、green band(0.52～0.60μm) and red band(0.63-0.69μm) of CCD, the reflectance of smaze is higher than fog-haze, dust-haze,while in NIR band(0.76～0.90μm), the reflectance of dust-haze is the highest in the fourth band, and the reflectance of samze is the second highest.The result suggested that the scattering ability of smaze is stronger than the other two in the visible band, and the absorption ability of dust-haze is stronger than others in the NIR band. In a word, there are some differences between these spectral characters of different haze.(2)The Smaze Index(SI) is establishedBecause of the spectrum specificity of smaze, certain proper band combination is explored to discover that smaze features and other features are separable by the sum of reflectivity between blue band(bl) and green band(b2) above the water, smaze features and other features are separable by the blue band(bl) above the vegetation. Smaze features and dust-haze features are separable by the ratio or difference of reflectivity between green band(b2) and NIR band(b4) above the building, and smaze features and fog-haze features are also separable by the sum of reflectivity between blue band(bl) and green band(b2). Through the analysis above, a new index named Smaze Index (SI=B1*(B1+B2)*(B2/B4)) was established using HJ-1CCD data.