BOP Pressure-Chart Analysis Approach Uses Computer Vision Technology

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 222116, “BOP Pressure-Chart Analysis Using Computer Vision Technology,” by Ali Mohamed S. Al Hosani, SPE, and Geetha Selvamoorthy, ADNOC, and Koushik Kumar, Trust Technical Services. The paper has not been peer-reviewed. _ Blowout-preventer (BOP) systems require inspection at least every 14 days, with a pressure test conducted to ensure their condition. This test currently involves manually decoding pressure-chart data from a mechanical pressure-chart recorder, a time-consuming process. The solution described in this paper automates this process by capturing the image of the pressure chart and decoding it using artificial-intelligence (AI) image analysis. Computer Vision for Pressure Readings In BOP pressure-chart analysis, computer vision transforms the traditionally manual and error-prone task of decoding pressure-chart data into an automated and precise operation. The system integrates various computer vision techniques to ensure accurate and efficient data extraction from pressure charts, streamlining the entire process. BOP Pressure Chart. The BOP pressure chart is a critical tool used in drilling operations to monitor pressure changes over time. The chart recorder, an essential component of this tool, features two or three pens driven by sensors that plot traces on a rotating paper disk (Fig. 1). The disk, typically predrawn with a green template, includes the following components: - Radial arcs representing time intervals - Concentric circles indicating quantity levels - Names of quantities (axes), which are labels for the data being measured - Quantity values, which are markings for ranges on concentric lines - Time marks denoting hours and weekday names Steps Required To Read the Pressure Chart Using Computer Vision Cropping Out the Pressure Chart From the User-Uploaded Image. The process of cropping the pressure chart involves isolating the relevant contour from the image and then cropping the image based on this contour. This process involves, in order, detecting contours, selecting the largest contour, finding the largest circle inside the contour, creating and applying the mask, and cropping the image. By following these steps, the process efficiently isolates and extracts the pressure chart from the user-uploaded image, setting the stage for subsequent analysis and processing. Changing Orientation of the Image To Align the Starting Point. Aligning the pressure chart’s orientation is crucial for accurate data extraction. This step involves correcting the image’s orientation by aligning the chart’s starting point based on the detected inner circle and using the histogram-projection method (HPM) to detect and align text within the circle. This process involves the following series of steps: 1. Detecting the inner circle 2. Extracting text from the circle 3. Applying HPM, used to align the image based on the detected text 4. Horizontal projection 5. Vertical projection 6. Aligning the starting point