Fishing_Event_Detection_Presentation-1.pptx

Hjjg

1. Fishing Event Detection and Fish Species Classification Using AI A Presentation on Automated Monitoring of Fisheries using Computer Vision and AI

2. Abstract • Fisheries regulations require accurate and efficient monitoring of fish catches. • Traditional manual monitoring is costly and time-consuming. • This project uses deep learning and computer vision for automated fish detection, tracking, and classification. • The approach enhances fisheries management efficiency and reduces human workload.

3. Methodology 1. Fish Detection: Uses YOLOX deep learning model. 2. Fish Tracking: Uses Kalman filter and Hungarian algorithm. 3. Fish Classification: Uses ConvNeXt classifier. 4. Dataset Preparation: Annotated video frames for training. 5. Performance Evaluation: Uses Average Precision (AP) and Average Recall (AR).

4. Domain of the Project • Artificial Intelligence & Computer Vision: AI-powered image processing and classification. • Fisheries Management: Automated fish catch monitoring. • Deep Learning & Data Science: Real-time object detection and classification.

5. Project Work Plan 1. Data Collection: Obtain and annotate EM video datasets. 2. Model Development: Train and optimize AI models. 3. Implementation: Integrate models into a functional system. 4. Testing & Evaluation: Validate performance with real-world data. 5. Deployment & Improvement: Fine-tune system for fisheries monitoring.

6. Objective • Automate fish detection and species classification. • Reduce dependency on human observers and lower monitoring costs. • Improve accuracy and efficiency of fisheries data collection. • Enhance sustainability of fish stock management with reliable data analysis.

7. Thank You! Questions?