Nonlinear method for area-wide near surface air temperature precision retrieval
US10831949B2 · kind B2 · utility
Assignees
Inventors
Key dates
| Filing date | Mar 6, 2017 |
| Grant date | Nov 10, 2020 |
| Priority date | — |
| Expiry date | Nov 23, 2037 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F2111/10
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A novel nonlinear method for area-wide near surface air temperature precision retrieval is described. The steps include: First, construct the 1st sub-model modelVEC1 to the f-th sub-model modelVECf. Establish and normalize raw data vectors of each gridded pixel sBlkVEC in the targeted area. Calculate the retrieved full maps (surfTf) of near surface air temperatures using each sub-model. Then, identify abnormal samples and define their near-range regions in surfTf Apply a selective arithmetic mean (SAM) approach to achieve precision temperature map surfT. And finally apply further modification to the pixels of surfT where pixlf∈badsurfTδf is true to all f=1, 2, 3, . . . .Using the super nonlinear algorithm, this invention provides a solution of retrieving near surface air temperature based on combinations of various factors (information fusion) to achieve satisfied prediction errors, which are independent of cloud levels and topographic characteristics. Specifically, the information fusion between space and ground surface enables reliable prediction of near surface air temperature maps overcoming the inference of cloud.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.