Ship Detection Based on Faster R-CNN Using Range-Compressed Airborne Radar Data

Abstract

Near real-time ship monitoring is crucial for ensuring safety and security at sea. Established ship monitoring systems are the automatic identification system (AIS) and marine radars. However, not all ships are committed to carry an AIS transponder and the marine radars suffer from limited visibility. For these reasons, airborne radars can be used as an additional and supportive sensor for ship monitoring, especially on the open sea. State-of-the-art algorithms for ship detection in radar imagery are based on constant false alarm rate (CFAR). Such algorithms are pixel-based and therefore it can be challenging in practice to achieve near real-time detection. This letter presents two object-oriented ship detectors based on the faster region-based convolutional neural network (R-CNN). The first detector operates in time domain and the second detector operates in Doppler domain of airborne Range-Compressed (RC) radar data patches. The Faster R-CNN models are trained on thousands of real X-band airborne RC radar data patches containing several ship signals. The robustness of the proposed object-oriented ship detectors is tested on multiple scenarios, showing high recall performance of the models even in very dense multitarget scenarios in the complex inshore environment of the North Sea.

The applications of sonar and radar systems in underwater detection

“…Another application of sonar technology is in groundwater exploration using nuclear magnetic resonance (NMR), a promising geophysical technique. According to the principles of UNMR (Ultra-Nuclear Magnetic Resonance), the spin rotation axis of hydrogen protons in groundwater becomes altered when energized, leading to observable NMR phenomena [3].…”

“…Through a method that estimates noise in the first half of the detection coil using data from the second half, it is feasible to determine the transfer function between the reference coil and detection coil. By employing a non-linear fitting method, the UNMR signal is estimated, separated from noise estimates, and preserved within the detecting coil [3].…”

An Incept-TextCNN Model for Ship Target Detection in SAR Range-Compressed Domain

“…To overcome these shortcomings, some target detection methods in the non-image domain have been presented in recently years and one representative example is the rangecompressed domain (RCD) as it does not require timeconsuming azimuth focusing. In [7], a ship detector is proposed based on Faster R-CNN working in the RCD; in [8], a two-step detection method is presented with the first step using complex signal kurtosis in the RCD to screen possible ship areas coarsely, and the second step applying CNN to further detect the potential ship areas; an oriented ship detection strategy is designed in [9], which calculates the constant false alarm rate detection threshold in the range-Doppler domain; a supportive ship tracking concept is introduced in [10] in the range-Doppler domain using an airborne-based radar sensor; in addition, a method for ship detection from raw SAR echo data is proposed in [11]. However, most of them are based on two-dimensional data for detection, where the model size tends to be large and dependent on CPU resources, and detection based on onedimensional data often doesn't take advantage of deep learning that can extract deep features.…”

Ship Detection From Raw SAR Echoes Using Convolutional Neural Networks

“…To by-pass this computational burden, recent efforts have been directed towards processing intermediate SAR data to detect ships. Using range-compressed data as input to neural network models is a promising avenue [35]- [38].…”