Online Exemplar-Based Fully Convolutional Network for Aircraft Detection in Remote Sensing Images

Convolutional neural network obtains remarkable achievements on target detection, due to its prominent capability on feature extraction. However, it still needs further study for aircraft detection task, since intraclass variation still restricts the accuracy of aircraft detection in remote sensing images. In this letter, we adopt regularity of aircraft circle response to design our end-to-end fully convolutional network (FCN), and embed online exemplar mining into our network to handle intraclass variation. The mined exemplars are employed to capture different intraclass characteristics, which effectively reduces the burden of network training. Specifically, we first select basic exemplars based on labeled information and initialize the relationships between exemplars and aircraft examples. Then, these relationships will be updated by the similarity of these examples in high-level features space. Finally, aircraft examples will be used to train different exemplar detectors according to updated relationships. Motivated by the geometric shape of aircraft, a circle response map is developed to construct our FCN to achieve more efficient aircraft detection. The comparative experiments indicate that superior performance of our network in accurate and efficient aircraft detection.

I. Introduction

Aircraft detection is a representative task in remote sensing images (RSIs) and has attracted increasing attention [1]–[9]. For a cross-shaped geometric structure of aircrafts, a great amount of methods design handcrafted features to accomplish aircraft detection. An et al. [2] utilized circle frequency filter to locate the region of interest (RoI) and the extracted histogram of gradients (HoGs) features to classify the regions containing an aircraft. To regulate the dominant orientation of a region and capture more detailed information, Zhang et al. [3] designed a new rotation-invariant appearance feature called histogram of oriented gradients normalized by polar angle. Zhao et al. [7] adopted aggregate channel features (ACFs) to describe aircrafts in RSIs, which offered richer representations and speeds up computations. However, the handcrafted features always need to adjust the parameters carefully and are not able to accurately handle target detection task within various scales and rotations.