AI-Powered App for Remote Monitoring of Oil Storage Tanks

Summary

An international oil and gas company needed to monitor the liquid level in its oil tanks remotely, to optimize inventory and catch leaks in time. INNERLUXES designed a desktop application that uses machine vision on images from onsite cameras and drones to detect oil tanks, read their numbers, and measure the oil level in each one in real time — through a seven-stage image-recognition pipeline.

About the Client

The Client is an international company operating in the oil and gas industry.

Remote Liquid-Level Monitoring to Optimize Inventory and Catch Leaks

The Client needed to monitor liquid levels in its oil tanks remotely, so it could optimize oil inventory management and detect leaks promptly.

A Computer-Vision App With a Seven-Stage Image Pipeline

After analyzing the Client's business needs, INNERLUXES designed a desktop application that works with images delivered by pre-installed onsite cameras and drones. Using machine vision, the application detects oil tanks, identifies their number, and determines the liquid level in each tank remotely and in real time, through a seven-stage image-recognition pipeline:

• Stage 1 — image preprocessing.
• Stage 2 — oil tank detection.
• Stage 3 — oil tank number detection.
• Stage 4 — oil tank number recognition.
• Stage 5 — detection of the area with liquid.
• Stage 6 — detection of the size of the area with liquid.
• Stage 7 — liquid level detection.

First, the application uses artificial neural networks trained on a sufficient volume of data to distinguish oil tanks from the background in the delivered images. It then applies neural networks reinforced with a filtering algorithm to scan and recognize the numbers shown on the detected tanks. To determine the oil level, computer-vision algorithms capture and process images of each tank's internal visible part, external visible part, and roof, after which the tanks are segmented and classified by their liquid level.

Real-Time Tank Monitoring for Better Inventory and Leak Detection

With an application that enables real-time remote detection of the liquid level in oil tanks, the Client can optimize its oil inventory management and detect oil leaks in good time.

Technologies and Tools

Linux, Python, PyQt, OpenCV, Keras, TensorFlow, NumPy.