基于深度学习的岩芯完整性智能预测方法OA

Core integrity serves as a critical evaluation metric in geological exploration,mineral resource development,and geotechnical engineering design,with its core quantitative parameter being the rock quality designation(RQD).Traditional RQD calculations rely on manual measurement of core block lengths,resulting in inefficiency.This paper proposes an intelligent core integrity prediction method based on YOLOv11.The method utilizes the YOLOv11 object detection algorithm to locate and identify intact core blocks,enabling automatic calculation of core RQD based on the detection results.Results demonstrate that the proposed algorithm achieves an average precision mean(mAP)of 92.4%and an F1 score of 90.7%,with a frame rate of 52.3 frames per second(FPS).These metrics surpass those of four comparison models:YOLOv8,YOLOv5,Faster-RCNN,and EfficientNet.Furthermore,the average error between the predicted RQD and manual measurements was only 3.84%.This method provides efficient and accurate support for evaluating core integrity,advancing intelligent development in geotechnical engineering.