三维定量荧光等值线谱量化相似度对比方法研究及应用

Similarity comparison for quantifying three-dimensional quantitative fluorescence contour spectra

定量荧光常规谱图对比法局限于出峰位置和数量的比较,完全靠手工操作和肉眼观察,不能进行谱图峰形特征的精确对比,难以区分特征峰相近的谱图。结合常规谱图对比法(谱图重叠观察、特征参数比较等方法)的优点,以特征峰位相近、特征等值线环相似为原则,建立了谱图相似度对比法,全面体现了特征峰峰位、数量和形状等多种信息。该方法将样品谱与标准谱中所有配对的特征峰的高点到其等值线环的每组特定测线距离的比值进行线性回归,其相关系数作为配对成功峰的相似度,除以样品谱和标准谱中非重复特征峰总数作为谱图总的对比相似度,根据总相似度大小,量化判断两谱图相似程度。在此基础上,应用软件技术实现了三维定量荧光录井谱图的自动对比和荧光物质鉴别。应用分析表明,该方法能够有效发现与识别储集层真假油气显示、确定储集层原油性质、鉴别各类荧光物质成分,对于提高油气显示发现率、有效评价储集层具有重要作用。

The conventional spectra contrast method of quantitative fluorescence was limited to a comparison of the peak positions and the peak numbers and entirely relied on manual operation and visual inspection, which can′t carry out precise comparison for the spectra peak characteristics and is difficult to distinguish the characteristic peaks with similar spectra. Combining with the advantages of the conventional spectrum contrast method (spectra overlap observation, the comparison of characteristic parameters etc.), the authors established the spectra similarity comparison method by using the characteristic peaks with the similar positions and the similar characteristic contour rings, which fully reflected the characteristic peaks′positions, number, shapes and the other information. The method made linear regression to the ratio of high points of all paired characteristic peaks in sample spectra and standard spectra to the distance of a specific group of sruvey lines of their contour rings, and their correlated coefficient was used as the similarity of the successfuly paired peaks, and regression value divided by total non-repeated peaks in the sample spectra and standard spectra as the total contrast similarity of the spectrogram. The similar degree of two spectra was quantitatively determined according to the total similarity. On this basis, three-dimensional quantitative fluorescence logging spectra achieve the automatic contrast and identification of fluorescent substances by the use of software technology. The application and analysis showed that this method can effectively discover and identify oil and gas shows, determine oil properties of the reservoirs, identify various types of fluorescent material compositions and is an important role in improving the discovery rate of oil and gas shows and effectively evaluating reservoirs.