Dipper throated optimization with deep convolutional neural network-based crop classification for remote sensing...

Area of Science:

• Remote Sensing
• Agricultural Science
• Computer Science

Background:

• Advancements in remote sensing necessitate efficient crop classification for food security and environmental monitoring.
• Traditional methods struggle with accuracy and scalability for high-resolution remote sensing data.
• Accurate crop classification is crucial for sustainable agriculture and resource management.

Purpose of the Study:

• To develop a novel crop classification technique, Dipper Throated Optimization with Deep Convolutional Neural Networks based Crop Classification (DTODCNN-CC).
• To enhance the accuracy and efficiency of crop classification from remote sensing images.
• To achieve high classification accuracy for diverse food crops.

Main Methods:

• Utilized a GoogleNet architecture (Deep Convolutional Neural Network - DCNN) for feature extraction.
• Employed Dipper Throated Optimization (DTO) for hyperparameter tuning of the GoogleNet model.
• Used Extreme Learning Machine (ELM) for crop classification, with parameters fine-tuned by the Modified Sine Cosine Algorithm (MSCA).

Main Results:

• The DTODCNN-CC approach demonstrated significantly higher crop classification accuracy.
• Experimental analyses confirmed the superior performance compared to existing state-of-the-art deep learning methods.
• The optimized GoogleNet and ELM models achieved robust feature extraction and classification.

Conclusions:

• DTODCNN-CC offers a promising solution for accurate and efficient crop classification using remote sensing data.
• This technique has substantial potential for applications in agriculture, food security, and environmental monitoring.
• The study highlights the effectiveness of integrating optimization algorithms with deep learning for remote sensing applications.