OLCN: An Optimized Low Coupling Network for Small Objects Detection

Abstract

In remotely sensed images, it is quite common to run into small objects, such as cars and small storage tanks. However, these small objects are quite easy to get ignored because of the positioning difficulty. Thus, small objects detection is very challenging for the remote sensing object detection task. In order to deal with this challenge, the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">optimized low coupling network</i> (OLCN) is proposed. First, a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">low coupling robust regression</i> (LCRR) module improves the positioning accuracy to avoid small objects getting missed. Second, a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">receptive field optimizing layer</i> (RFOL) is proposed to train better classifiers by providing more accurate <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">regions of interest</i> (RoIs). Experimental results on the public dataset HRRSD verify the effectiveness of the proposed OLCN. Small objects detection metric is improved from 5.70% of the baseline to 22.90% of the OLCN. Moreover, the proposed method has reached state-of-the-art performance on the HRRSD dataset.