TCD-Net: A Novel Deep Learning Framework for Fully Polarimetric Change Detection Using Transfer Learning

Abstract

Due to anthropogenic and natural activities, the land surface continuously changes over time. The accurate and timely detection of changes is greatly important for environmental monitoring, resource management and planning activities. In this study, a novel deep learning-based change detection algorithm is proposed for bi-temporal polarimetric synthetic aperture radar (PolSAR) imagery using a transfer learning (TL) method. In particular, this method has been designed to automatically extract changes by applying three main steps as follows: (1) pre-processing, (2) parallel pseudo-label training sample generation based on a pre-trained model and fuzzy c-means (FCM) clustering algorithm, and (3) classification. Moreover, a new end-to-end three-channel deep neural network, called TCD-Net, has been introduced in this study. TCD-Net can learn more strong and abstract representations for the spatial information of a certain pixel. In addition, by adding an adaptive multi-scale shallow block and an adaptive multi-scale residual block to the TCD-Net architecture, this model with much lower parameters is sensitive to objects of various sizes. Experimental results on two Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) bi-temporal datasets demonstrated the effectiveness of the proposed algorithm compared to other well-known methods with an overall accuracy of 96.71% and a kappa coefficient of 0.82.

Dual Attention-Based Global-Local Feature Extraction Network for Unsupervised Change Detection in PolSAR Images

“…2) SE [20] obtains the DI using the information entropy and further detects changes using the ETS based on the chi-square distribution. 3) TCD-Net [32] uses adaptive multiscale convolutional blocks and residual blocks to capture the change-aware features of objects with different sizes. 4) EEMCNN [33] uniformly utilizes multidimensional dilated convolution layers to extract deep features from bitemporal images and change deep features.…”

“…As for the research of PolSAR image CD, a local restricted CNN [31] effectively detects changes from the discriminative DI. Based on the DI generated by transfer learning, a novel endto-end three-channel deep neural network (TCD-Net) [32] uses adaptive multi-scale shallow blocks and residual blocks to detect changes in an unsupervised manner. Seydi et al [33] proposed an end-to-end multi-dimensional CNN (EEM-CNN) for PolSAR image CD, which adopts multi-dimensional dilated convolution to simultaneously extract change-aware features.…”

A Comparative Study of Machine Learning Classifiers for Crop Type Mapping Using Vegetation Indices

“…We used the two most common statistical metrics, including OA and Kappa Coefficient (KC), to demonstrate the performance of the models and the impact of the input variables on crop type mapping (Habibollahi et al, 2022). These statistical criteria can be calculated using Equation (1-3):…”

“…A comprehensive study compared different ML and deep learning algorithms with a dual attention deep neural network for crop mapping in Iran. The outcomes demonstrated the best OA of 98.54% for the Aq-Qala agricultural area (Seydi et al, 2022). Any data correlating with vegetable growth can be used as a proper source for monitoring the seasonal growth cycle of crops.…”

Endoscopic Image Classification Based on Explainable Deep Learning

“…In addition, there are a number of useful methods to overcome overfitting and increase the accuracy of the model, which are denoising [ 6 ], initialization and setting momentum [ 7 ], batch normalization [ 8 , 9 ], dropout [ 10 ], and drop connect [ 11 ]. Transfer learning is also a great technique to enhance the deep learning model with a high positive benefit in medical imaging using pre-trained weight in the last part of the convolutional neural network [ 12 , 13 , 14 ]. Traditional deep learning trains each model on a specific domain, whereas transfer learning uses natural image datasets or similar types of medical image datasets with pre-trained weight to reduce the training time.…”