比辐射率反演的一个新方法

A New Method for Emissivity Inversion

物体的比辐射率是反映物体热辐射性质的一个重要参数,比辐射率反演不仅提供了一种获得物体比辐射率的便捷途径,并且涉及许多有趣的理论问题.目前该反问题只有一种基于Hermite函数展开的普适函数方法.考虑到这类反问题固有的不适定性问题,发展另外一种独立的方法非常有必要.本文基于Laguerre函数展开为比辐射率反问题发展了一种新的方法.对一些理论谱的数值计算表明,新方法不仅拥有老方法的所有优点,并且跟老方法相比拥有更高的计算精度,能够计算结构更加复杂的理论谱.本文最后通过比较和分析揭示了新方法优越性的根本原因.

How to obtain the emissivity of objects is an important research content of the anti telemetry technology. The inverse emissivity problem provides us a convenient way to obtain the emissivity. The purpose of this inverse problem is to get the emissivity according to measured total radiant power, by solving an integral equation. Although an exact solution formula of this inverse problem was derived, it can never be used directly to obtain emissivity in practical inversion. The integral equation requires that the total radiant power must satisfy a so-called asymptotic behavior control condition. The experimental measured total radiant power is impossible to satisfy this ABC condition because of the ill-posed problem. In order to realize the inversion of real system, some concrete methods are necessary. At present, there is only one existing method, that is the universal function set method based on the Hermite functions. In this paper, starting from the famous Laguerre functions, an independent new method is developed for inverse emissivity problem. The new method has all the advantages of the old method. Moreover, numerical calculation shows that the new method has higher accuracy than the old one. It can calculate the theoretical spectrum with more complex structure. Each expansion term in the new method satisfies the asymptotic behavior control condition, so it can automatically filter out the effective information in the experimental total radiant power data. At the same time, the noise of information can be eliminated. This new method is suitable for the inversion of real system. If there are more sophisticated and perfect radiation power measurement data, this new method can be applied to the emissivity inversion for the first real system.