The data preprocessing effect on radiometric survey for petroleum and gas

Based on seven kinds of radiometric survey in an area covering 150 km2 in the middle part of Sichuan, nine parameters at each measuring point with a total number of 1902 data were obtained and a study on statistical distribution characteristics was conducted. The results show that there are three types of statistical distribution pattern among the raw data obtained from seven kinds of radiometric measurement, namely, approximately arithmetic normal distribution, logarithmic normal distribution, and the above two-mixed distribution. By generalized power transformation, the above-mentioned three types of statistical distribution were normalized into a kind of statistical distribution, i.e. normal distribution, thus obtaining objective background and anomaly threshold values, and providing a more proper mathematical model for further statistical analysis