Philip L. Kurkoski

Zero latency image compression for real time logging while drilling applications

LWD (logging while drilling) has very limited bandwidth, typically 4 bit/sec. Many applications such as quantifying dip and azimuth, fault detection, borehole breakout and stress analysis require the image be available in real time. Due to the relatively large volume of data within an array measurement, such as a density, gamma, or resistively image, transmission of images in real time over limited bandwidth is a very challenging problem. Standard compression techniques require continuously accumulating a large block of data, then compressing that data. Continuous transmission of compressed data is not only a very complex problem, but also introduces a latency equivalent to the time required to accumulate the whole block of data. Two approaches discuss herein. The first approach is to compress small block of data, then apply a filter on the reconstructed data/image. The filter eliminates burst noise. The second approach is to carry out feature detection and extraction, followed by image compression for only the interval(s) of interest. This detection approach processes data in real time. It triggers on the start and end of a feature without processing any further features. This provides feature detection with zero latency. Compressing only the interval of interest provides a data compression with both higher resolution and higher compression ratio (Gamal and Phil).