文件列表:

satelite-data-classification-tensorflow-cnn.ipynb

# Satellite Image Classification with TensorFlow This repository contains the code and documentation for a satellite image classification project using TensorFlow. The goal of the project is to accurately predict the category of satellite images by building a convolutional neural network (CNN). ## Project Overview The model is designed to classify satellite images into four categories: cloudy, desert, green_area, and water. The dataset consists of labeled images collected from various sensors and Google Maps snapshots. ## Key Features - Data preparation using TensorFlow's ImageDataGenerator for real-time data augmentation. - Understanding the functionality of `flow_from_directory` for efficient data input pipeline. - Visualization of the model's predictions on satellite images. - Implementation of a CNN with several convolutional layers, max-pooling, dropout, and dense layers. - Usage of early stopping to prevent overfitting. - Evaluation of the model's performance using accuracy, precision, recall, and F1-score metrics. ## Dataset The dataset used for training and validation is the RSI-CB256 dataset, which includes satellite images across four classes with different environmental features. ### 4 class Images:     ## Model Architecture The model is a CNN with the following layers: - Conv2D with ReLU activation, followed by BatchNormalization and MaxPooling. - Dropout layers to reduce overfitting. - A Flatten layer followed by a Dense layer with ReLU activation. - The output layer is a Dense layer with a softmax activation function to output class probabilities. ## Results The trained model achieved an accuracy of approximately 94.23% on the test set. The detailed classification report is available in the repository, showing precision, recall, and F1-score for each class. ## Requirements The project requires the following libraries: - TensorFlow - Keras - NumPy - Matplotlib - scikit-learn - Pandas

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：