In response to the problems of the slow detection of formation effluent and weak anti-interference ability in drilling site,this study proposed a rapid detection system for shale moisture content based on near-infrared images with significant absorp-tion characteristics in water.This system used the 940 nm band near-infrared camera to capture images,which eliminated visible light interference and improved the sensitivity of detection.It also used the filtering algorithm based on the wavelet transform ef-fectively to remove the scattering noise in the image,and utilized the regional growth algorithm for image segmentation to exclude interference and avoid false segmentation,thus improving the anti-interference ability of the system.In addition,the HDBSCAN clustering algorithm was utilized for rapid and precise extraction of shale areas and the grayscale average of shale images was aquired.Finally,a gray value-water content model was constructed using segmented linear regression algorithm to predict the water content of shale.The experimental results show that this system works effectively in the complex natural gas drilling environment,with maximum absolute error of 1.87%and high accuracy.The average detection time is not more than 6 seconds,which is fast.Compared with the drying method,this system significantly improves the detection speed on the basis of ensuring the accuracy.Compared with humidity sensors,this system not only has higher detection accuracy,but also covers the moisture detection re-quirements of larger areas.
关键词
气体钻井/岩屑含水量/近红外图像/HDBSCAN聚类
Key words
gas drilling/shale water content/near-infrared image/HDBSCAN cluster
维普
万方数据