🎯 Problem Statement

The aquaculture industry faces significant challenges in monitoring feed waste and shrimp mortality. Current manual inspection methods are subjective and inefficient, impacting both growth outcomes and economic viability. Real-time data access is crucial for informed decision-making in shrimp farming operations.

💡 Proposed Solution

Our innovative approach combines NIR Spectral Imaging Technology with convolutional neural networks (CNN) to:

• Accurately identify residual feed pellets
• Detect deceased shrimp specimens
• Analyze soil properties in real-time (pH and organic matter content)
• Provide spectral signature-based identification

🚧 Technical Challenges

Primary Challenges

1. Real-time result visualization
2. High TSS (Total Suspended Solids) and organic matter accumulation
3. Spectral signature acquisition in turbid conditions
4. Differentiation between feed pellets and deceased shrimp on clay soils

Previous Approaches Tested

Echo Sounders

• Echo Sounder OVA ES-6000 (200Khz transducer)
• Furuno FCV-628 (600W advanced frequency transducer)
• Result: Effective for live shrimp detection but inadequate for feed and mortality detection

Underwater Cameras

• SeaLife DC2000 Professional Camera
• Custom infrared camera systems
• Result: Limited effectiveness due to high turbidity levels

🔬 Research Focus

Current Investigation

1. Spectral signature validation for:
   • Deceased shrimp detection
   • Saturated feed pellet identification
2. Sensor performance in high-turbidity environments
3. Comparative analysis of multi vs hyperspectral sensors
4. Optimal imaging position (surface vs underwater)

Alternative Approach

We’ve explored underwater RGB cameras with image processing and CNN implementation, including:

• Advanced detection algorithms
• Image enhancement techniques
• Comprehensive dataset development

However, due to turbidity challenges, our primary focus has shifted to spectral imaging solutions.

🛣️ Development Roadmap

1. Hypothesis Validation

   • Scientific community review
   • Hardware/software provider consultation
   • Prototype development

2. Field Testing

   • Ecuadorian shrimp farm trials
   • Farmer feedback collection
   • Solution refinement

3. Future Expansion

   • Comprehensive AI-powered data collection system
   • Adaptation for other agricultural sectors (poultry, pork)

📑 Project Presentation

View Full Presentation

📬 Contact Information