ExxonMobil Research Qatar and Providence Photonics, LLC, a U.S. based firm, are undertaking research to develop a Remote Gas Detection (RGD) system that integrates computer vision algorithms and infrared (IR) optical gas imaging technology to achieve autonomous remote detection of hydrocarbon plumes. The RGD system is designed to provide continuous surveillance and early warning to operations personnel in case of gas releases and to detect fugitive gas emissions. The RGD system utilizes a custom built IR imager and integrated cooler assembly, and a computer vision algorithm that analyzes the video output from the IR imager to determine the presence of hydrocarbon plumes. Most hydrocarbons have strong absorption peaks in a narrow mid-wave IR (MWIR) region. The algorithm takes advantage of the difference in contrast between a hydrocarbon plume and the background in an IR image and the temporal changes due to plume behavior for the analysis. The algorithm compares sequentially collected IR images and uses a multi-stage confirmation process to confirm the detection. It has multiple filters that mitigate interferences like steam and other movement of objects in the scene such as humans, vehicles, and trees. Early field tests indicate that a 4 lb/hr propane leak could be autonomously detected from a distance of up to 800 feet. The RGD camera assembly enclosure is designed to obtained explosion proof certification using the ATEX standard for deployment at classified/hazardous areas in oil and gas processing facilities. Instrument air provides cooling and is used to purge the system. Multiple deployment opportunities at process facilities are currently underway. Results from field testing at these process facilities will help researchers investigate the effect of temperate and harsh weather conditions, the effect of varying temperatures and gain a better understanding of equipment wear and tear, maintenance requirements and life expectancies. These data sets will produce an accurate assessment of the performance of the RGD system under actual working conditions and will be used to qualify the technology for widespread adoption within the industry. Work has also been undertaken to compare the performance of the RGD system versus existing detection technologies. The most common leak detection technology is point sensors and path infrared sensors. This technology requires dispersed gas to physically contact the point sensors or move between two path detectors. Field tests are used to compare the performance of these mature technologies to the capabilities of RGD.
Accurately object searching plays an important role in computer vision. Retrieving and locating target objects in images are object searching's two sub-tasks. Aiming to promote the precision and recall of object searching, selecting appropriate image representation methods is the core issues. The representation method needs to provide enough discriminative features. Our approach adopts locality sensitive hashing method to extract enough sift features. The extracted features contain inliers and outliers. In order to distinguish them, random context confidence scores of features are computed. Our algorithm offers 3 benefits:1) A novel partition method is adopted to divide images. It is easy to be parallelized during computing contexts.2) A novel random points selecting method is adopted to avoids ill-defined boundary for target objects; 3) Multiple target objects in one image can be located by clustering all the features of each image with their coordinates. The experiment on a challenging Belgalogo dataset highlights the performance of our approach.
Hydrocarbon leaks in large scale LNG and gas processing and handling facilities present a risk to health and safety, as well as a negative impact on the environment. Existing autonomous leak detection systems rely primarily on point and path detection. This technology requires dispersed gas to physically contact the point sensors or move between two path detectors. Manual leak checks are also performed periodically by checking one component at a time using a flame ionization detector (FID) or optical gas imager, however the size and complexity of these facilities means that smaller leaks may go undetected for extended periods of time and unintended releases may occur when plant personnel are not present. ExxonMobil Research Qatar Ltd. and Providence Photonics LLC have developed the IntelliRed™ Remote Gas Detection system that integrates computer vision algorithms and infrared (IR) optical gas imaging technology to autonomously scan for and identify leaks. The IntelliRed™ system utilizes a unique mid-wave IR (MWIR) imager and a computer vision algorithm that analyzes the video output from the IR imager to determine the presence of hydrocarbon plumes. Most hydrocarbons have strong absorbance peaks in a narrow MWIR region. The algorithm takes advantage of the difference in contrast between a hydrocarbon plume and the background in an IR imager. The algorithm compares sequentially collected IR images and uses a multi-stage confirmation process to model the behavior of the plume and confirm the detection. It has multiple filters that mitigate interferences such as humans, vehicles, and trees. After extended pilot deployments, the IntelliRed™ technology has been qualified and is commercially available for leak detection and environmental applications. An innovative differential infrared (DIR) camera design now offers the possibility of lower detection limits and higher immunity to false alarms for the IntelliRed™ system. The design employs two cooled MWIR sensors with a common optical path. The infrared energy from the scene is divided using a beam splitter, focusing a spatially registered image on each sensor. The spectral filtering for the two sensors is chosen so that one sensor can visualize a hydrocarbon plume while the second sensor cannot. A synchronized system clock for the two sensors ensures that the frames are temporally aligned. The result is a camera which produces both spatially and temporally aligned frames with the plume present in one frame but absent in the other. A differential image is produced by comparing the two sensors frame by frame, providing a robust filter for common interferences such as steam and dust plumes. The DIR camera design requires new computer vision techniques to exploit the information provided by the reference sensor. Results are compared to other autonomous hydrocarbon detection technologies, including single sensor IntelliRed™ technology. New applications enabled by the DIR camera design are discussed, including aerial pipeline surveys. The technology is currently being qualified with DIR based IntelliRed™ pilot deployments underway in Qatar and the United States.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.