Modular sensor fusion platform for hydrocarbon storage equipment monitoring
Various embodiments of the present technology relate to solutions for hydrocarbon equipment monitoring. In some examples, a detection system comprises a compute engine, a gimbal, a sensor suite, and an imaging system. The compute engine generates signaling that directs the gimbal to orient the imaging system, signaling that directs the imaging system to image the equipment, and signaling that directs the sensor suite to sense the equipment. The gimbal orients the imaging system based on the signaling. The imaging system images the equipment and transfers images depicting the storage equipment to the compute engine. The sensor suite senses the equipment and transfers sensor data that characterizes the equipment to the compute engine. The compute engine processes the data with a machine learning engine trained to determine the status of the equipment. The compute engine transfers a machine learning output that indicates the status to a downstream system.
1 . A detection system to monitor a hydrocarbon storage environment, the detection system comprising:
a compute engine;
a gimbal communicatively coupled to the compute engine;
a sensor suite communicatively coupled to the compute engine;
an imaging system communicatively coupled to the compute engine and operatively coupled to the gimbal;
a housing that mounts the compute engine, the gimbal, the sensor suite, and the imaging system, wherein the sensor suite is internally mounted to the housing and detachably coupled to the housing via a quick connect socket;
the compute engine configured to:
receive a command to view a piece of hydrocarbon storage equipment;
generate control signaling that directs the gimbal to orient the imaging system and transfer the control signaling to the gimbal to orient the imaging system to view the piece of hydrocarbon storage equipment;
generate image signaling that directs the imaging system to image the hydrocarbon storage equipment and transfer the image signaling to the imaging system; and
generate sensor signaling that directs the sensor suite to sense the hydrocarbon storage equipment and transfer the sensor signaling to the sensor suite;
the gimbal configured to receive the control signaling and orient the imaging system to view the piece of hydrocarbon storage equipment based on the control signaling;
the imaging system configured to receive the image signaling, image the hydrocarbon storage equipment, and transfer image data that depicts the hydrocarbon storage equipment to the compute engine;
the sensor suite configured to receive the sensor signaling, sense the hydrocarbon storage equipment, and transfer sensor data that characterizes the hydrocarbon storage equipment to the compute engine; and
the compute engine configured to receive the image data and the sensor data, process the image data and the sensor data with a machine learning algorithm trained to determine a status of the hydrocarbon storage equipment, and transfer a machine learning output that indicates the status of the hydrocarbon storage equipment to a downstream system.
2 . The detection system of claim 1 wherein the housing comprises at least one of an aluminum frame, a steel frame, or a plastic frame.
3 . The detection system of claim 1 wherein the compute engine and the gimbal are mounted internally on the housing and the imaging system is mounted externally on the housing.
4 . The detection system of claim 3 wherein the imaging system and the compute engine are detachably coupled to the housing.
5 . The detection system of claim 4 , wherein the imaging system comprises a quick connect attachment to detachably couple to the housing.
6 . The detection system of claim 5 wherein the housing is detachably coupled to a mounting structure that mounts the housing at an elevation.
7 . The detection system of claim 1 wherein the imaging system comprises an infrared video camera.
8 . The detection system of claim 1 wherein the sensor suite comprises a visible spectrum video camera and a rangefinder.
9 . The detection system of claim 1 wherein the gimbal comprises one or more actuators and rotors to adjust one or more of roll, yaw, or tilt of the imaging system.
10 . The detection system of claim 1 wherein:
the machine learning is trained to detect leaks in the hydrocarbon storage equipment based on the image data and sensor data; and
the machine learning output comprises data that indicates when the hydrocarbon storage equipment comprises a leak.
11 . The detection system of claim 1 wherein:
the machine learning is trained to measure fill levels in the hydrocarbon storage equipment based on the image data and sensor data; and
the machine learning output comprises data that indicates the fill level of the hydrocarbon storage equipment.
12 . The detection system of claim 1 further comprising a cooling system to regulate the temperature of the compute engine.
13 . The detection system of claim 12 wherein the cooling system comprises an air jet, an air channel, an inlet filter, and an outlet filter.
14 . The detection system of claim 13 wherein the inlet filter and the outlet filter are permeable to air and impermeable to liquid and particulate matter.
15 . The detection system of claim 13 wherein the inlet filter and the outlet filter are permeable to air and semipermeable to liquid and particulate matter.
16 . The detection system of claim 1 wherein the image data comprises infrared video footage that depicts the hydrocarbon storage equipment.
17 . The detection system of claim 1 wherein the sensor data comprises visible spectrum video footage and a distance between the hydrocarbon storage equipment and the detection system.
18 . A method of operating a detection system to monitor a hydrocarbon storage environment, the method comprising:
mounting a thermal imaging system and a sensor suite, wherein the thermal imaging system is externally coupled to the detection system via a gimbal and the sensor suite is internally mounted to the detection system via a quick connect socket;
orienting a thermal imaging system to view hydrocarbon storage equipment;
generating thermal image data depicting the hydrocarbon storage equipment;
generating sensor data that characterizes the hydrocarbon storage equipment;
processing the thermal imaging data and the sensor data using a machine learning algorithm trained to identify a status of the hydrocarbon storage equipment and responsively generating a machine learning output; and
transferring the machine learning output to downstream systems.
19 . The method of claim 18 further comprising:
vectorizing the thermal imaging data to generate numeric representations of the thermal imaging data; and wherein:
processing the thermal imaging data using the machine learning algorithm comprises processing the numeric representations of the thermal imaging data; and
responsively generating the machine learning output comprises generating the machine learning output that indicates when the hydrocarbon storage equipment comprises a leak.
20 . A non-transitory computer-readable medium stored thereon instructions to monitor a hydrocarbon storage environment using a detection system, that, in response to execution, cause a system comprising a processor to perform operations, the operations comprising:
orienting a thermal imaging system externally mounted to the detection system via a gimbal to view a piece of hydrocarbon storage equipment;
directing the thermal imaging system to generate thermal image data depicting the hydrocarbon storage equipment;
directing a sensor suite to generate sensor data that characterizes the hydrocarbon storage equipment, wherein the sensor suite is internally mounted to the detection system via a quick connect socket;
obtaining the thermal image data depicting the hydrocarbon storage equipment generated by the thermal imaging system and obtaining the sensor data generated by the sensor suite;
processing the thermal imaging data and the sensor data using a machine learning algorithm trained to identify the status of the hydrocarbon storage equipment and responsively generating a machine learning output; and
transferring the machine learning output to downstream systems.