SBIR Phase I: A Novel Method for Atmospheric Correction of Earth Observation Satellite Data

这个小企业创新研究（SBIR）项目的更广泛的影响/商业潜力是提供软件服务，以纠正地球观测卫星（EOS）数据的地面反射率。EOS必须透过地球看？大气中的散射和光的衰减会导致测量地面目标实际反射率的系统误差。大气引起的误差影响数据的效用，因为大气气溶胶含量，即，湿度、灰尘、花粉、烟雾颗粒等，波动很大，影响了全球监测、国防和农业的应用。如果成功满足规格，数据的价值可能使其成为一个重要且不断增长的市场机会。 这SBIR第一阶段项目提出纠正气溶胶引起的错误EOS数据的逆转的效果，发现经验结构，独立的气溶胶类型和潜在的可预测的通过测量暗目标反射率？水体没有水生植被、夹带的沉积物和风浪的镜面反射。研究方法是对陆地卫星8号数据进行提取和统计分析，陆地卫星8号数据是所有其他EOS平台数据校准和验证的标准参考。通过一系列的启发式研究来解决这个问题：（1）重建蓝、绿色和红色波段与近红外（NIR）的关系，这些关系最初是用Landsat 5和7数据拟合的（较长的波长可能不被处理，因为它们抵抗大气影响），（2）使用这些关系来逆转误差，（3）开发选择的方法，证明并应用校准校正的暗目标，以及（4）通过将后算法反射率与由便携式光谱测定法测量的地面反射率进行比较来测量残差。残余误差可能是由于与暗目标相关联的不确定性。一旦开发和验证，该算法将被带到Landsat 8 Cal/瓦尔团队进行验证。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的知识价值和更广泛的影响审查标准进行评估来支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) project is to provide software service to correct Earth observation satellite (EOS) data to at-ground reflectance. EOS must look through the Earth?s atmosphere that induces systematic error in measuring the actual reflectance of ground targets through scatter and attenuation of light. Atmospherically-induced error affects data utility because the atmospheric aerosol content, i.e., humidity, dust, pollen, smoke particles, etc., fluctuates greatly, impacting applications for global monitoring, defense and agriculture. The value of data could make this a significant and growing market opportunity if the specifications are met successfully. This SBIR Phase I project proposes to correct aerosol-induced error in EOS data by reversing the effect, found empirically to be structured, independent of aerosol type and potentially predictable through measurement of dark target-reflectance ? water bodies clear of aquatic vegetation, entrained sediment and specular reflectance from windblown waves. The method of study is extraction and statistical analysis of Landsat 8 data, the standard reference for calibration and validation of data from all other EOS platforms. This problem is approached through a series of heuristic investigations to (1) reconstruct relationships of blue, green and red bands to near infrared (NIR) originally fitted using Landsat 5 and 7 data (longer wavelengths may not be addressed because they are resistant to atmospheric affects), (2) use these relationships to reverse the error, (3) develop methods to select, proof and apply dark targets that calibrate the correction, and (4) measure residual error by comparing post-algorithm reflectance to at-ground reflectance measured by portable spectrometry. The residual error is likely due to uncertainty associated with dark targets. Once developed and proofed, the algorithm will be brought to the Landsat 8 Cal/Val team for validation.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.