Computer Vision Patent Strategy for Autonomous Underwater Inspection

Overview of Technical Issues:

In autonomous underwater inspection systems, the water environment harmfully absorbs and scatters light, causing severe image degradation and limited visibility range that directly reduces inspection accuracy and operational coverage; additionally, illumination creates excessive backscatter that introduces noise and reduces contrast; the goal is to achieve reliable computer vision-based defect detection and measurement in turbid underwater conditions across extended inspection areas.

Solution directions generated for this problem

Problem Direction 1 :

ImproveImage signal contrast
VS
ConstraintBackscatter noise level

Inspiration 1 : Cross-domain reference

Application Principle: #2 Taking out (Extraction)
Cross-domain applicability Assess applicability
Super-resolution using natural handheld-motion applied to a user device
Innovative Solution Refine solution

Cross-polarized dual-filter imaging system for backscatter rejection

Separate target reflection from backscatter using polarization state difference
How to solve :
  • Mount circular polarization filter (right-handed) on LED array
  • illuminate at 480-520nm wavelength matching seawater transmission window
  • target surface reflection reverses to left-handed polarization while suspended particle backscatter remains right-handed
  • Install left-handed circular polarizer on camera lens (extinction ratio ≥1000:1)
  • camera captures only polarization-reversed target signal, rejecting 85-92% of co-polarized backscatter noise from water column particles
  • Operate at 200-300W LED power with beam divergence 15-20°
  • achieve >3:1 signal-to-noise ratio at 3-5m range
  • validate contrast using calibrated gray card targets in turbidity 5-15 NTU conditions
Expected Effect : Contrast ratio >3:1 at 5m; backscatter rejection 85-92%; defect detection accuracy +60%
Risk Control :
  • polarizer alignment tolerance ±2°
  • depolarization from multiple scattering in high turbidity >20 NTU
  • filter transmission loss 15-25%

Problem Direction 2 :

ImproveInspection visibility range
VS
ConstraintIllumination energy consumption

Inspiration 1 : Cross-domain reference

Application Principle: #19 Periodic action
Cross-domain applicability Assess applicability
Portable wireless access point
Innovative Solution Refine solution

Pulsed laser illumination with duty-cycle gating for extended underwater visibility

Pulsed laser replaces continuous lighting
How to solve :
  • Replace continuous LED with pulsed blue-green laser (520nm, 10Hz repetition, 1ms pulse width) — 1% duty cycle achieves 5m peak range with average power equal to current 2m continuous system
  • Synchronize high-speed camera gating (exposure window 0.8-1.2ms aligned to pulse arrival) to capture target reflection while rejecting inter-pulse darkness, maintaining >3:1 SNR
  • Implement adaptive pulse energy control — range sensor feeds distance data to modulate per-pulse energy (50-200mJ) inversely with target proximity, optimizing energy distribution across inspection path
Expected Effect : 5m range at baseline power; 60min operation duration maintained; SNR >3.5:1
Risk Control :
  • laser-camera synchronization jitter >50μs
  • pulse energy stability <±8%
  • thermal management of peak power

Problem Direction 3 :

ImproveDefect measurement precision
VS
ConstraintIllumination energy consumption

Inspiration 1 : Cross-domain reference

Application Principle: #28 Mechanics substitution (Replace mechanical system)
Cross-domain applicability Assess applicability
Wireless communication device and method for controlling transmission power
Innovative Solution Refine solution

Adaptive computational triangulation for energy-efficient precision measurement

Replace intensity-based measurement with geometric encoding
How to solve :
  • Project single-line laser pattern (520nm, 15mW continuous) onto defect surface
  • capture pattern deformation with monochrome camera (exposure 10ms, ISO 400)
  • apply triangulation algorithm to compute 3D coordinates from line displacement, achieving ±1mm precision with baseline-to-distance ratio ≥0.15
  • Mount laser projector 200mm offset from camera axis at 25° angle
  • calibrate system using checkerboard target (10×10mm squares) at 1m, 3m, 5m distances
  • store calibration matrix (intrinsic + extrinsic parameters) for real-time coordinate transformation
  • Implement multi-frame averaging (5 frames at 2Hz) to suppress noise from water turbulence
  • apply Gaussian filter (σ=1.5 pixels) to extracted laser line
  • use sub-pixel edge detection (Steger algorithm) to locate line centroid within 0.3-pixel accuracy, ensuring measurement repeatability <0.8mm
Expected Effect : Precision ±1mm at 5m range; energy consumption reduced 82% (15mW vs 85mW baseline); operation duration extended 5.6×
Risk Control :
  • calibration drift from pressure/temperature changes
  • laser line extraction failure in high-turbidity zones
  • triangulation accuracy degradation beyond 5.5m

Problem Direction 4 :

ImproveIllumination intensity
VS
ConstraintMust not deteriorate

Inspiration 1 : Cross-domain reference

Application Principle: #1 Segmentation
Cross-domain applicability Assess applicability
Adaptive light source
Innovative Solution Refine solution

Segmented dual-zone adaptive illumination system for underwater inspection

Divide illumination into near-field and far-field zones with independent intensity control
How to solve :
  • Deploy segmented LED array with two spatial zones: Zone-1 (0–1.5m) uses low-intensity wide-angle LEDs (20–30 lux) to minimize near-field backscatter
  • Zone-2 (1.5–5m) uses high-intensity collimated laser diodes (150–200 lux at target) to maximize target illumination
  • Implement distance-sensing trigger circuit using ultrasonic rangefinder (±2cm accuracy) to dynamically activate appropriate zone based on real-time target distance, switching threshold at 1.5m with 50ms response time
  • Mount Zone-1 LEDs in 120° diffuser ring around camera lens, Zone-2 lasers in focused 15° beam array offset 25cm from optical axis to separate backscatter cone from imaging path
Expected Effect : Contrast ratio >3.5:1 at 5m range; backscatter noise reduced 65%; energy consumption +40% vs single-zone but −55% vs uniform high-intensity
Risk Control :
  • ultrasonic rangefinder accuracy degradation in high-turbidity water
  • LED-laser switching delay causing frame dropout
  • thermal management for high-density laser array in sealed housing
Patsnap Eureka Solution