SingTown Technology Empowers the Development of Underwater Robots Capable of "Seeing, Avoiding, and Navigating

SingTown Technology嘅AI智能相機系統，基於OpenMV，為水下機械人提供實時避障同路徑規劃功能。利用機器視覺同圖像識別技術，該系統能夠喺複雜嘅水下環境中識別障礙物形狀、距離同運動趨勢，實現自主導航、目標定位同安全操作。廣泛應用於科研探索、水產養殖監測同水下救援等場景。

The underwater environment is complex and variable, with light, reflections, and floating debris often posing challenges for robot navigation. SingTown Technology's AI smart camera solution, developed based on the OpenMV platform, provides underwater robots with precise visual obstacle avoidance capabilities through computer vision and image recognition algorithms.

The system can capture underwater images in real-time through the camera, and the edge-side AI analyses the shape, colour contrast, and depth features in the images to identify different types of targets such as rocks, aquatic plants, obstacles, and fish schools, while calculating relative distances. When obstacles are detected, the system automatically generates path correction instructions to guide the robot in adjusting its heading or speed, enabling autonomous obstacle avoidance and safe navigation.

The OpenMV intelligent camera boasts high sensitivity and low power consumption, allowing it to adapt to environments with insufficient underwater lighting and complex backgrounds. Users can train custom models to enable the system to recognise specific targets, such as cables, pipelines, or underwater structures, thereby supporting refined operational tasks.

This solution can be deployed at the edge on the device side, independently performing recognition and decision-making tasks. It also supports collaboration with sonar, gyroscopes, and propulsion control systems to build a multi-sensor fusion AI navigation system. Therefore, it is highly suitable for providing targeted embedded vision technology for underwater equipment with limited conditions and complex requirements.

SingTown Technology，以機器視覺、人工智能及電腦視覺為核心，令水下機械人具備「清晰視野、避障及穩定導航」之智能能力，為科研、工業及安防領域嘅水下任務提供可靠技術支援。