Abstract
Compared to traditional object detection of horizontal bounding box, detecting rotated objects with arbitrary orientations and various scales is a critical yet challenging task especially in remote sensing images. Although considerable progresses have been made through the utilization of deep CNN, there still exists space for exploration of oriented object detection. In this article, we propose a refined single-stage detector for oriented objects, which is equipped with the enhanced feature extraction network and an adaptive feature alignment module for finer detection. For feature enhancement, a bidirectional and inner residual feature pyramid network and a multiscale feature aggregation module are devised for getting more representative features. Then, to address the problem of feature misalignment, we propose the adaptive feature alignment module to reallocate the sampling locations and weight the positive and negative feature points during the refined detection process. The reallocation and weighting will be adjusted adaptively according to the results from the coarse detection. Experiments conducted on public remote sensing databases show the effectiveness of our method.
Highlights
O BJECT detection is one of the fundamental tasks in the area of computer vision, a number of high-performance general-purpose object detectors have been proposed with significant advances in deep convolutional networks (ConvNets)
We propose an Adaptive Feature Alignment Module (AFAM) to make the convolution features of the predictor be more appropriately coordinated with the adjusted default boxes
Creating feature pyramids in different ways, we observe that Bidirectional and Inner Residual Feature Pyramid Network (BIRFPN) achieves the best promotion (69.27% → 69.35% → 69.54%→ 70.29%)
Summary
O BJECT detection is one of the fundamental tasks in the area of computer vision, a number of high-performance general-purpose object detectors have been proposed with significant advances in deep convolutional networks (ConvNets). The challenges in rotational object detection in aerial remote sensing images are analyzed with respect to the following:. Objects in natural scenes are generally oriented upward, while objects in aerial remote sensing images are often oriented arbitrarily To address these issues, we mainly focus on how to enhance feature extraction capability of backbone network and how to make refined detector fit for oriented targets better in this paper. With the regression result from the coarse stage, the convolution region will be shifted, scaled and rotated appropriately, which makes the detector “seeing” more accurate and reliable features during the refined stage.
Sign-in/Register to view highlights & summary 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 callMore From: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
