報文

ベイズの定理に基づく最大エントロピー法による
エネルギー分散型X線スペクトル分析

Applications of a Bayesian Based Maximum Entropy Method
to Energy Dispersive X-ray Spectra

新川隆朗，中村奈津子，加藤裕志，渋谷和好，中田宗隆
Takao SHINKAWA, Natuko NAKAMURA, Hiroshi KATO, Kazuyoshi SHIBUYA,
and Munetaka NAKATA

　An energy dispersive X-ray detector is widely used to measure intensity spectra of characteristic X-rays of component elements in an specimen. In a long-time measurement to obtain a low-noise spectrum, the spatial resolution of X-rays decreases partly due to the radiation of X-rays from the surrounding area larger than the probe size itself, and partly due to the drift of the electron beam and/or of the specimen stage. In a short-time measurement to avoid the above difficulties, the X-ray spectrum contains a much higher-level noise. It has been well known that usual smoothing techniques downgrade the sharpness of the spectrum. In the present study, a Bayesian Based Maximum Entropy Method (BBMEM) is applied to the spectrum obtained in a few seconds, and is proved to be an effective technique to reduce the noise in the spectrum

Keywords : Bayesian Theory, Maximum Entropy Method, Energy Dispersive X-ray Spectrum.