Abstract: Borehole imaging technology is a crucial method for detailed rock mass structure detection.The unavoidable probe eccentricity issue during testing significantly decreases the quality of borehole images,thereby limiting the precise automatic identification of rock mass structure information.To address this challenge,a borehole image restoration method based on grayscale feature analysis is proposed to enhance the identification of rock mass structure and the accuracy of geometric parameter measurements.Initially,the response characteristics of probe eccentricity in borehole images are analyzed based on the operating principle of borehole imaging.An image grayscale feature model is formulated,and a method for three-dimensional positioning analysis of the panoramic imaging probe is established.Subsequently,by calculating the grayscale error matrix based on spatial trajectory of the probe,the correction and restoration of image grayscale errors caused by probe eccentricity are achieved.A quantitative analysis of the azimuth error in borehole images induced by probe eccentricity is conducted,and a method for image perspective error restoration based on probe spatial positioning correction is developed.Application analysis of typical borehole images demonstrates that,by acquiring spatial position information of the probe based on grayscale feature analysis and conducting grayscale and perspective error corrections,the effectiveness and measurement accuracy of borehole images are significantly enhanced.The binarization processing results also validate that the restored images can more effectively reveal crucial rock mass structure information and minimize interference from external factors,thus providing image data foundation for automatic identification of rock mass structure.