Image analysis speeds the search for oil
In a specific temperature range and under the right conditions, organic material transforms to oil and gas. At higher temperatures, oil and gas break down to worthless carbon residue. Because of this, prospectors need a temperature indicator to detect if oil and gas are present. Derived from woody material, vitrinite, common in the sedimentary rocks that can hold hydrocarbons, is an ideal temperature indicator. Analysis of vitrinite shows whether rocks have reached the correct temperature range for oil and gas to be present. Although analyzing vitrinite normally takes three to five days, Roy Enrico and Robert Borger of Mobil Oil Corp. (Dallas, TX) are now using computer-based video and image analysis to obtain accurate results in only eight hours.
When vitrinite is heated, carbon atoms reorganize, and the vitrinite reflects more light. By measuring the amount of light reflected, the highest temperature ever reached in vitrinite-bearing rocks can be determined. In turn, this indicates whether oil or gas is present.
Automating the process
In the past, researchers analyzed vitrinite by manually searching for 100 vitrinite grains in each rock sample and measuring re flectance on each grain. Now, this process has been automated using off-the-shelf cameras, frame grabbers, and image-processing software. In the system developed by Enrico and Borger, a 3-CCD RGB video camera from Sony (San Jose, CA) is attached to a microscope. Captured image data are then digitized into a workstation using the XVideo S-bus digitizer board from Parallax Graphics (Santa Clara, CA).
In operation, researchers automatically capture 24-bit color images with up to 10 or 20 grains of vitrinite in each. This process is then repeated up to 10 times per sample to analyze 100 individual grains. After capturing the digital images, images are analyzed with WIT image processing software from Logical Vision (Burnaby, QC, Canada). Pixel values are converted to a measurement of light intensity, which can be related to known reflectance values. "Because each vitrinite grain is digitized as hundreds of pixels, it produces hundreds of measurements instead of one, increasing accuracy over manual methods," says Borger.
