Target Tracking in Water Areas Using Neural Network Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing boat safety systems face challenges in accurately and efficiently tracking objects in water areas due to limited computing capacity on board, leading to potential collisions with close or dangerous objects.

Innovation Solution

A tracking system comprising a pan-tilt unit, a camera module, and a control module implementing a neural network, which continuously generates a video stream, detects objects, and calculates pan and tilt angles to center the camera on selected objects, thereby enhancing tracking accuracy and frequency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If distance and trajectory calculations are performed for all detected objects, then tracking accuracy is improved, but computing load increases

Engineering Contradiction:
Improvetracking accuracyVSAvoidcomputing load
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

The patent extracts and focuses computational resources on a specific target object rather than processing all detected objects. The system identifies a particular object of interest and concentrates tracking calculations on this selected object, thereby reducing overall computing load while maintaining high tracking accuracy for the critical target.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies different processing qualities to different objects: high-precision tracking calculations are applied locally to the selected target object, while other objects receive standard detection processing. This localized quality approach optimizes computational resource allocation by intensifying processing only where necessary.

Inventive Principle:
Principle #3Local quality

2Reliability

If tracking calculations are performed at high frequency, then collision risk is reduced, but computing capacity requirements increase

Engineering Contradiction:
Improvecollision risk reductionVSAvoidcomputing capacity
Core Design Contradiction:
ReliabilityVSPower

Solution Approach 1:

The patent extracts high-frequency tracking calculations from the general object detection process and applies them specifically to the selected target object. This allows the system to perform frequent distance and trajectory updates for the critical object without requiring high-frequency processing for all detected objects, thus reducing overall computing capacity requirements.

Inventive Principle:
Principle #2Taking out (Extraction)

3Loss of information

If all detected objects are tracked, then detection completeness is improved, but system complexity increases

Engineering Contradiction:
Improvedetection completenessVSAvoidsystem complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent extracts the complex tracking functionality and applies it selectively to one selected object rather than implementing universal tracking for all detected objects. This maintains detection completeness by still detecting all objects, but reduces system complexity by concentrating sophisticated tracking resources on a single target.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP4539490A1Target tracking in a water area
Publication Date: 2025.04.16 SEA AI GMBH
  • EP4539490A1 patent drawingFigure 1
  • EP4539490A1 patent drawingFigure 2~3
  • EP4539490A1 patent drawingFigure 4

AI summary

The invention relates to a method for tracking an object (5) in a water area (2) using a tracking system (30), comprising a tracking phase including controlling a pan-tilt unit to orientate a camera module around the pan axis and the tilt axis to centre the field of view of said camera module on a selected object (5) and thus track said selected object (5).