Abstract
The content-based remote sensing image retrieval (CBRSIR) has recently become a hot topic due to its wide applications in analysis of remote sensing data. However, since conventional CBRSIR is unsuitable in harsh environments, this article focuses on the cross-modality CBRSIR (CM-CBRSIR) between synthetic aperture radar (SAR) and optical images. Besides the large interclass and small intraclass in CBRSIR, CM-CBRSIR is limited by prominent modality discrepancy caused by different imaging mechanisms. To address this limitation, this study proposes a deep cross-modality hashing network. First, we transform optical images with three channels into four different types of single-channel images to increase diversity of the training modalities. This helps the network to mainly focus on extracting the contour and texture shared features and makes it less sensitive to color information for images across modalities. Second, we combine any type of randomly selected transformed images and its corresponding SAR or optical images to form image pairs that are fed into the networks. The training strategy, with paired image data, eliminates the large cross-modality variations caused by different modalities. Finally, the triplet loss, in combination with the hash function, helps the modal to extract the discriminative features of images and upgrade the retrieval efficiency. To further evaluate the proposed modality, we construct a SAR-optical dual-modality remote sensing image dataset containing 12 categories. Experimental results demonstrate the superiority of the proposed method with regards to efficiency and generality.
Highlights
U NPRECEDENTED advances in earth observation technologies, over the past few decades, have caused a significant increase in both quality and quantity of remote sensing image archives [1], [2]
Content-based remote sensing image retrieval (CBRSIR), which is defined as the search for remote sensing images of similar information content within a large archive with a given query image serving as a Manuscript received June 13, 2020; revised August 14, 2020; accepted August 31, 2020
To solve the aforementioned CM-content-based remote sensing image retrieval (CBRSIR) challenges, we propose a deep cross-modality hashing network (DCMHN)
Summary
U NPRECEDENTED advances in earth observation technologies, over the past few decades, have caused a significant increase in both quality and quantity of remote sensing image archives [1], [2]. Optical remote sensing images have many advantages over SAR images They are intuitionistic and easy to understand, have rich color and texture information, present obvious target structure characteristics, high resolution, and a large field angle. Rapid development of feature learning has accelerated exploration of cross-modality retrieval tasks in the field of natural image analysis These include retrieval between image and text [8], [9], image and audio [10], [11], as well as RGB and infrared images [12], [13]. SAR images lack the specific imaging principle and presence of speckle noise, as well as the rich color information contained in optical images (see Fig. 1) Based on these factors, the existing works cannot. XIONG et al.: DEEP CROSS-MODALITY HASHING NETWORK FOR SAR AND OPTICAL REMOTE SENSING IMAGES RETRIEVAL
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
