Marine Species Recognition Using Image Analysis

• Author(s):

  P.Dinesh Kumar | G.Ananthi | N.Chaithanya | K.Keerthana | R.Nikitha

• Keywords:

  C Marine Species Classification, Underwater Image Processing, Deep Learning, CNNs, Object Detection, Biodiversity Monitoring.

• Abstract:

  The Identification And Classification Of Marine Species Play A Vital Role In Ecological Studies, Biodiversity Protection, And Sustainable Fisheries Management. Traditional Approaches That Rely On Manual Observation Are Not Only Time-intensive But Also Susceptible To Errors Due To Human Limitations And Challenging Underwater Conditions. These Limitations Underscore The Need For Automated, Intelligent Solutions. The Core Of This Approach Combines CNNs With High-performance Object Detection Algorithms Such As YOLOv8. These Models Are Trained To Recognize Distinct Marine Species From Underwater Images. Since Such Images Often Suffer From Poor Lighting And Color Distortion, The System Integrates A Preprocessing Unit That Applies Image Sharpening, Contrast Boosting, And Noise Removal To Enhance Visual Quality.A Domain-specific Dataset Comprising Labeled Images Of Various Marine Species Was Curated To Train And Validate The System. Multiple Deep Learning Architectures Were Tested Namely ResNet, YOLOv8, EfficientNet, And Faster R-CNN. Evaluation Metrics Such As Accuracy, F1-score, Precision, And Mean Average Precision (mAP) Were Used To Assess Performance.Evaluation Metrics Such As Recall, Precision, Accuracy, And MAP Validate The Effectiveness Of The Method. This Project Proves That AI And Deep Learning Can Significantly Improve Marine Data Collection, Reduce Human Error, And Enable Scalable Biodiversity Monitoring SolutionsThis Work Demonstrates That Integrating AI With Marine Biology Not Only Automates Species Monitoring But Also Supports Large-scale Conservation Efforts And Ecological Analysis.

---

Other Details

• Paper id:

  IJSARTV11I6103775

• Published in:

  Volume: 11 Issue: 6 June 2025

• Publication Date:

  2025-06-13

Download Article