Keywords: 
• sedimentation microfacies /
• well logging curve shape /
• direction probability density /
• wavelet descriptors /
• automatical identification

Abstract: For the sedimentation microfacies identification based on the well logging curves of different shapes, three conditions including the invariance of translation and dimension scaling and the modifiability of rotation should be satisfied. In this paper, a new method of sedimentation microfacies identification using the shape of well logging curves which can satisfy the three conditions above was set .Through computing the boundary direction probability density of different shapes of well logging curves, then the wavelet transformation should be made, and few low frequency wavelet descriptors whose direction probability density can be reconstructed can be used as the sedimentation microfacies identification and the wavelet descriptor identification of different sedimentation microfacies can be set at the same time. The wavelet analyze should be made for the direction probability density function of different well logging curve, so the sedimentation microfacies information should be mapped to a reduced dimensionality characteristic vector space consist of few low frequency wavelet descriptors from the high dimensionality characteristic vector space, therefore, the discrepancy information of different sedimentation microfacies can be enlarged. For the complicated problem of sedimentation microfacies identification based on the shapes of well logging curves, this method can degenerate it into a simple problem that judging few low frequency wavelet descriptors and the wavelet descriptors can also be used for the sedimentology analysis.