The importance of transportation to national development is growing. It is a significant factor in determining patterns of production and commerce, and as a result, economic integration. It can also assist some nations in generating cash by offering transportation services. ¹ There is a need to detect and monitor the most current condition of transportation, especially the mode of transportation, the container types, as well as the related infrastructures. The information acquired can be used for taking immediate actions that will allow maintenance, mitigations, and sustainable development.

However, evaluating, monitoring, and detecting the mode of transportation is time-consuming expensive, and labour-intensive. The traditional way can only cover small scale and use in situ measurement techniques. One of the most famous methods that allow us to collect information on a large scale, time-efficient, and without direct interaction with the object is using remote sensing technologies. Especially Synthetic Aperture Radar (SAR) remote sensing, which can be operated day or night, penetrate clouds, and without weather issues because it uses microwave electromagnetic energy.

The length of microwave electromagnetic wave used in SAR remote sensing is between 1 centimetre to 1 meter. Similar to optical remote sensing, radar sensors also operate with one or more bands. It is identified by letters, such as P, L, S, C, X, K, Q, V, and W. The longest band is P with 100 cm of wavelength, while the shortest band is W with 0.3 cm of wavelength. On the contrary, the highest frequency is the W-band, with 100 GHz, and the lowest frequency is the P-band with 0.3 GHz. ²

L band has higher penetration than C- or X-bands, therefore for mapping the volume of vegetation L- and C-band is better since they can be penetrated deeper into the vegetation canopy. While the X-band will scatter close to the surface of the vegetation canopy. ³ When the frequency becomes lower, it can propagate with low attenuation and penetrate deeper, like the P-band that can penetrate several centimetres of the forest, dry ground, snow, ice, and the soil surface. ⁴

Surface roughness affects the dark level of the produced SAR remote sensing image. Smooth surfaces like water bodies, roads, and other paved surfaces will produce a dark image since there is no return of backscatter signal due to it having a specular reflection. Rough surfaces like buildings, towers, tree trunks, and other vertical structures will produce a bright image since there is a strong return of backscatter signal due to it having diffuse scattering. ³

SAR remote sensing is categorized as an active sensor and can be classified as an imaging sensor. The image produced using SAR remote sensing depends on the polarization of an electromagnetic wave. Different images will produce different visualizations if using different types of polarization. Polarization is the orientation of the plane of oscillation of a propagating signal. ⁵ Concerning the Earth's surface, the perpendicular polarization planes are commonly referred to arbitrarily as horizontal and vertical. Polarization interacts differently with objects on the Earth’s surface, which leads to the different brightness levels recorded in a specific polarization channel. ³

In SAR remote sensing the polarization uses abbreviations, such as horizontal transmission and horizontal reception (HH), vertical transmission and vertical reception (VV), horizontal transmission and vertical reception (HV), and vertical transmission and horizontal reception (VH).

One of the earliest generations of SAR remote sensing sensors is Canada’s RADARSAT program includes RADARSAT-1, RADARSAT-2, and the RADARSAT Constellation Mission (RCM). RADARSAT-1 (launched in 1995) and RADARSAT-2 (2007) operated in the C-band, offering high-resolution imaging for disaster monitoring, agriculture, and forestry. The RCM, launched in 2019, comprises three satellites that ensure daily global coverage, improving monitoring capabilities for Arctic regions, sea ice, and maritime safety. RADARSAT-2, with its ability to capture fine-resolution imagery, is also used for oil spill detection and urban mapping ( https://www.asc-csa.gc.ca/eng/satellites/radarsat/).

Currently, there are many types of SAR remote sensing sensors available. One of the well-known is the Sentinel-1 SAR belonging European Space Agency (ESA). Sentinel-1 carries a C-band with dual polarization (HH+HV, VV+VH) and a temporal resolution of 12 days ( sentinels.copernicus.eu). PALSAR-3 belongs to Japan Aerospace Exploration Agency (JAXA), revisit time is 46 days, and carries an L-band with multiple polarization (eorc.jaxa.jp). PALSAR with HH signal polarization can penetrate the forest canopy deeper and is returned from the bottom of the forest stronger than the VV signal polarization. ⁶ TerraSAR-X is a commercial SAR remote sensing sensor belonging to the German Aerospace Center (DLR) with a 1-meter spatial resolution and temporal resolution of 11 days. TerraSAR-X carries an X-band, with a range of different modes of operation, allowing it to record images with different swath widths, resolutions, and polarizations ( dlr.de).

Italy's COSMO-SkyMed constellation, developed by the Italian Space Agency (ASI), consists of four satellites operating in the X-band. These satellites provide rapid revisit times and high-resolution imaging for military and civilian applications, including disaster management, land cover mapping, and infrastructure monitoring. A second-generation COSMO-SkyMed constellation has been launched to enhance data quality and revisit frequency ( https://portal.cosmo-skymed.it/CDMFE/home).

Other Asian countries also have launched SAR remote sensing sensors. It is India’s Radar Imaging Satellite (RISAT) series, including RISAT-1 (C-band) and RISAT-2 (X-band). These satellites are used for agricultural monitoring, disaster management, and surveillance. Additionally, the upcoming NASA-ISRO Synthetic Aperture Radar (NISAR) mission is a collaboration between NASA and the Indian Space Research Organisation (ISRO). Scheduled for launch in 2025, NISAR will operate in both L-band and S-band, making it the first dual-band SAR system. It is expected to provide unprecedented data for applications ranging from ice sheet dynamics to natural hazards ( https://www.isro.gov.in/NISARSatellite.html).

South American countries also have SAR remote sensing sensors. The Argentine Space Agency (CONAE) developed the Satélite Argentino de Observación COn Microondas (SAOCOM) constellation, consisting of SAOCOM-1A and SAOCOM-1B. These satellites operate in the L-band and are primarily designed for agriculture and soil moisture studies. Their ability to penetrate soil makes them excellent for monitoring water content in agricultural fields, and improving crop management ( https://saocom.invap.com.ar/).

With better spatial and temporal resolution, SAR remote sensing is suitable for transportation detection and monitoring activities. Many kinds of research used SAR remote sensing for the mode of transportation studies. Studies for monitoring the airport were done in 2001 ⁷ and 2004. ⁸ Studies for ship detection were introduced in 2003 ⁹ and car detection in 2007 ¹⁰ and 2016. ¹¹

As discussed, the remarkable capabilities of SAR remote sensing provide unprecedented opportunities to employ these technologies in a broad variety of modes of transportation studies. There are currently some SAR remote sensing literature review studies conducted. Gend and Genderen ¹² was the first comprehensive SAR remote sensing review paper conducted related to the Interferometric Synthetic Aperture Radar (InSAR). The author comprehensively discussed the issues, techniques, and application of InSAR. Ouchi ¹³ has summarized the recent trends and advances in SAR remote sensing on various topics. A topic such as fields of applications, specifications of airborne and spaceborne SAR, and information content of interpretation, InSAR, and Polarimetric SAR (PolSAR). However, the review is too general and does not focus on transportation issues. Other references [ 14] also briefly discussed the remote sensing techniques for road evaluation. They concluded that remote sensing techniques offer new potential for pavement stakeholders to evaluate large areas in an efficient time. The authors provided comprehensive information about remote sensing technologies, not only SAR but also Ground Penetrating Radar (GPR), infrared thermography, LiDAR and terrestrial laser scanning, hyperspectral, and emerging technology such as mobile smartphones. Further, other references [ 15] also introduced the trend in commercial SAR remote sensing. They introduced high-resolution wide-swath capabilities, multi-polarimetry, and the development of increased bandwidth of SAR sensors. They also emphasized that the timely availability and delivery of SAR remote sensing data is important. Therefore, expanding the number of satellite constellations and ground station networks is needed to fulfil the requirement.

There is still a need for a more comprehensive and focused review to discuss various aspects of the application of SAR remote sensing for transportation studies. Thus, the main objective of this study is to 1) identify the trend and gaps in the use of SAR remote sensing technologies for them to be easily adopted by transportation stakeholders; and 2) to classify the use of various SAR remote sensing technologies regarding the information they can provide in various mode of transportation detection, evaluation, and monitoring.

The Differential Interferometric Synthetic Aperture Radar (DInSAR) method has been widely used in studies aimed at monitoring transportation infrastructures, as highlighted by numerous works in the literature. ^{7, 8, 27, 30, 43, 45, 46, 48, 50, 52, 80– 82} This method is particularly valuable because it enables researchers to detect and measure the three-dimensional displacement of transportation infrastructure over time, providing a detailed time series of movement or deformation. The capability to track such displacements is crucial for understanding the structural stability and functionality of infrastructure systems, such as roads, railways, bridges, and airports which are integral to modern transportation networks ( Figure 9).

The DInSAR process involves using two radar images of the same area, captured from slightly different vantage points by a satellite or airborne radar system. These images are acquired at different times, and their slight positional difference creates a phase shift in the returned radar signals. This phase difference arises because the radar signals travel slightly different distances to and from the sensor, a phenomenon referred to as the "range difference." By analyzing the phase differences between the two images, researchers can derive critical height and displacement information about the study area.

To create an interferogram, which is the core product of the DInSAR technique, the phase differences between the two radar images are processed. The interferogram visually represents these differences, and with further processing, it can reveal subtle ground movements or deformations in the monitored area. These deformations are often caused by natural processes, such as earthquakes, landslides, or subsidence, or by human activities, such as construction or mining.

The utility of DInSAR lies not only in its ability to detect minute displacements but also in its capacity to do so over large areas with high spatial resolution. This makes it an indispensable tool for transportation infrastructure monitoring, as it helps engineers and policymakers assess potential risks and develop mitigation strategies. For instance, runways of airports experiencing subsidence in a specific section could be identified early using DInSAR, allowing maintenance efforts to be targeted effectively, thereby minimizing risks to public safety and reducing economic losses.

Synthetic Aperture Radar (SAR) remote sensing has been extensively utilized in transportation studies for over three decades, contributing significantly to monitoring and understanding various modes of transportation. The evolution of SAR sensor technology, coupled with advancements in algorithm development, has enabled the application of SAR in this domain. Moving forward, the role of SAR remote sensing is anticipated to expand, particularly in the detection and monitoring of moving transportation objects such as vehicles, ships, and aircraft.

Most transportation systems studied using SAR remote sensing are concentrated in the Northern Hemisphere, where the most developed transportation networks are located. This geographic focus reflects the higher demand for advanced transportation research in industrialized and densely populated regions. Among the different modes of transportation, maritime transportation has been the most extensively studied. The primary reason for this lies in the strong contrast between the ship, as the detected object, and the surrounding sea, which serves as a relatively homogeneous background. This distinct contrast makes ship detection more feasible and reliable using SAR. However, future research aims to address the challenges of detecting ships in more heterogeneous backgrounds, such as coastal areas or inland waterways, where the environment is more complex.

When it comes to transportation infrastructure, SAR studies have primarily focused on airports, followed by railways and harbours. Airports are often studied due to their critical role in global transportation networks and their vulnerability to environmental changes, such as land subsidence. The Differential Interferometric Synthetic Aperture Radar (DInSAR) method has been the most commonly employed technique for monitoring transportation infrastructure. By using time-series SAR data, DInSAR allows researchers to detect and measure land subsidence in three dimensions. This capability is invaluable for assessing the stability and safety of transportation infrastructure, particularly in areas prone to geological or environmental changes.

In terms of algorithms, significant advancements have been made in the last decade. Neural network algorithms began to gain traction after 2010, revolutionizing how SAR data is processed and analysed for transportation studies. More recently, deep learning techniques have emerged as the state-of-the-art approach for detecting and monitoring modes of transportation. These methods are highly effective in identifying complex patterns in SAR datasets, enabling the accurate detection of transportation objects and activities. However, deep learning is computationally intensive, requiring significant processing power and time. Therefore, future research must balance the trade-offs between accuracy and computational efficiency when applying deep learning algorithms to SAR data from diverse sources.

SAR's potential in transportation studies continues to grow, driven by technological advancements and the increasing availability of diverse SAR datasets. Future research is likely to focus on improving detection capabilities in complex environments, enhancing algorithm efficiency, and expanding the scope of applications to include emerging transportation technologies and systems.

Ethical approval and consent were not required