复杂地层情况下全井周电成像图像修复方法

Image inpainting for fullbore electrical imaging logging in complex formations

在电阻率成像测井资料处理中,图像修复的目的是得到全井周电成像图像。基于多点地质统计的图像修复方法考虑了处理窗长内整幅图像的统计信息,用这种方法修复均匀性较好的地层测井图像时,图像修复效果较好,但在修复空白区域较大、非均匀性比较严重的复杂地层测井图像时,修复结果出现异常。提出了基于多点地质统计与插值相结合的电阻率成像测井图像修复方法——一种混合图像修复方法。该方法充分利用了插值方法速度快的特点,初步快速修复电成像测井空白区域的图像,在空白区域利用其初步修复结果,在滤波域进行匹配,寻找模式匹配块,在非均匀性比较严重的空白区域利用初步修复结果,减少了该区域模式匹配的不确定性,比较容易寻找正确的模式,降低了出现异常模式的概率,使修复结果可靠。实测电成像资料处理结果表明,该方法适用于空白宽度较大、非均匀性较严重的复杂地层图像的修复。

In the processing of resistivity imaging logging data,the purpose of image inpainting is to obtain fullbore electrical imaging images.In view of the presence of gaps in an electrical imaging logging image,a multipoint geostatistics method can be used to achieve effective image inpainting results for formations with homogeneity because it accounts for the statistical information of the whole image in the processing window length.However,when applied to an electrical imaging image with large gaps from formations with heavy heterogeneity,the method produces incorrect filling results because of the mistake in finding a matching pattern.In this paper,an image inpainting method based on combining multipoint geostatistics with inverse distance weighted interpolation is proposed.First,this method fully uses the high efficiency of the interpolation method to quickly fill the gaps in an electrical imaging logging image.Next,the initial inpainted fullbore logging image is filtered,which could help in finding a matching pattern.It is easier to find the right matching pattern and reduce the probability of incorrect patterns for formations with heavy heterogeneity,owing to the use of the initial inpainted fullbore logging image after the interpolation in the gaps.The processing results for the measured electrical imaging data showed that this proposed method is effective for inpainting an electrical imaging logging image with large gaps from formations with heavy heterogeneity.