Self- Adaptive Glow Warm Swarm Optimization Technique in Optimal Feature Selection in Grading of Fruit Mango

An image classification hybridized approach based on CNN and Fuzzy Classifier was developed for enhancing the classification accuracy of the mango fruit. In the feature extraction, texture, geometric and color-based characteristics are applied. In feature selection Self-adaptive Glow worm Swarm Optimization is proposed (SA-GSO). In fruit maturity prediction KNN classifier is used with inclusion of SA-GSO. Finally in classification process fruit segmented image features are subjected to Fuzzy classifier and fruit segmented image is subjected to CNN. Finally generated classification results depend on results generated by fuzzy classifier and CNN. Classification of fruit mango is done in three classes: good, average and poor that is based on fruit maturity, texture and geometric features. Finally, the suggested model's assessed findings investigate a variety of defective and healthy mango images, demonstrating that when compared to other approaches in terms of diverse metrics approaches, the suggested strategy offers the highest degree of accuracy.