QUANTITATIVE-EVALUATION OF CROSSHOLE SEISMIC-REFLECTION IMAGES USING PHYSICAL MODEL DATA

Physical models give us a known geometry with which to compare our pro cessed reflection images and therefore our imaging techniques. We show how this comparison may be quantified in order to evaluate processed images properly. A crosshole data set was acquired through a model int errogated at ultrasonic frequencies using Durham University's physical modelling system. Various reflectivity images were obtained using pro cessing sequences which include deconvolution, wavefield separation an d migration. An error-energy scheme was used to assess the quality of these images, by comparing them against a best-fit depth model. A synt hetic data set was also used to evaluate the imaging capability of the crosshole geometry and the effectiveness of the different processing schemes.