Still images captured using mobile phone cameras have been shown to be very useful for bit forensics purposes. However, since still images can only capture views at certain viewing angles, they often provide insufficient information on each cutter's diamond table and substrate. Videos solve the problem of lack of view angles. This paper demonstrates how a video captured on a phone can provide additional bit information benefiting forensics, and provides recommendations for capturing videos to maximize information content.
Bit forensics as well as the future IADC dull grading system require meticulous data collection for various drill bit regions, including critical information such as blade numbers and cutter locations. To automate this process, this paper introduces a multi-stage approach using bit videos. Initially, a detection model is employed to accurately locate the drill bit within the video. Next, computer vision algorithms are utilized to segment the different blades. Spatial geometry algorithms are then applied to reconstruct the camera trajectory, which aids in estimating the blade numbers. Finally, each cutter within the segmented blades is further segmented into different regions.
This study explores the utilization of videos for the automated location segmentation of drill bits, a crucial aspect of the revised IADC dull grading system currently being proposed. The videos capture the entire drill bit from multiple angles, encompassing top-down views and a full 360-degree rotation. The IADC dull grading system necessitates the precise recording of position information including blade numbers, pocket numbers, and bit zones. By employing videos instead of still images, this study demonstrates that spatial geometric information of the drill bit can be obtained more completely and efficiently. Given a video that conforms to established shooting standards, the proposed automatic position calculation algorithm efficiently completes the segmentation of different parts of the drill bit and labels them in accordance with the relevant standards. Notably, video capturing offers several advantages over still photography; it obviates the need for complex training for operators, only necessitating adherence to basic camera trajectory guidelines, and substantially reduces the time needed to collect such data. Importantly, the location segmentation algorithm employed in this study is capable of running in real-time, thereby streamlining and accelerating the IADC dull grading and bit forensics processes.
This paper introduces a novel video-based algorithm for drill bit segmentation. The algorithm automatically segments and labels various components of the drill bit as per established criteria, generating comprehensive data vital for damage analysis. By employing this algorithm, videos of PDC drill bits can be processed with remarkable speed and accuracy. This represents a substantial advancement in data collection methods, with implications for improving the quality of bit damage assessment.
