This project aims to address environmental and safety hazards such as unauthorized personnel/vehicle intrusion, oil leakage, wellsite water accumulation, and equipment anomalies in the oilfield production process. It involves the development of an integrated solution that combines specialized oilfield algorithms, edge cloud servers, and an alarm processing platform in both software and hardware components.

Please refer to the official website of Objecteye for more details.

Research contents

• Research on the Construction of a High-Risk Operation Video and Image Database and an Intelligent Augmentation Method.
• Research on Efficient Image Processing Methods for Complex Scenes.
• Research on Precise Recognition Technology for Violations in High-Risk Oilfield Operation Scenarios.
• Research on the Analysis of Patterns of Typical Violations and Control Measures.

Responsibilities

• Using ONNX and TensorRT for model conversion and high-speed inference, and addressing the bugs in the model converted with TensorRT 7.
• Deploying the model for testing in the Hikvision NVR.