Object detection and tracking are critical parts of unmanned surface vehicles(USV) to achieve automatic obstacle avoidance. Off-the-shelf object detection methods have achieved impressive accuracy in public datasets, though they still meet bottlenecks in practice, such as high time consumption and low detection quality. In this paper, we propose a novel system for USV, which is able to locate the object more accurately while being fast and stable simultaneously. Firstly, we employ Faster R-CNN to acquire several initial raw bounding boxes. Secondly, the image is segmented to a few superpixels. For each initial box, the superpixels inside will be grouped into a whole according to a combination strategy, and a new box is thereafter generated as the circumscribed bounding box of the final superpixel. Thirdly, we utilize KCF to track these objects after several frames, Faster-RCNN is again used to re-detect objects inside tracked boxes to prevent tracking failure as well as remove empty boxes. Finally, we utilize Faster R-CNN to detect objects in the next image, and refine object boxes by repeating the second module of our system. The experimental results demonstrate that our system is fast, robust and accurate, which can be applied to USV in practice.
We are motived by the need for generic object detection algorithm which achieves high recall for small targets in complex scenes with acceptable computational efficiency. We propose a novel object detection algorithm, which has high localization quality with acceptable computational cost. Firstly, we obtain the objectness map as in BING and use NMS to get the top N points. Then, k-means algorithm is used to cluster them into K classes according to their location. We set the center points of the K classes as seed points. For each seed point, an object potential region is extracted. Finally, a fast salient object detection algorithm is applied to the object potential regions to highlight objectlike pixels, and a series of efficient post-processing operations are proposed to locate the targets. Our method runs at 5 FPS on 1000*1000 images, and significantly outperforms previous methods on small targets in cluttered background.
A digital carrier synchronization module with high working frequency is indispensable for high-speed digital coherent optical receivers to recover the transmitted symbols. We proposed a method to increase the working frequency of the digital carrier synchronization (DCS) module based on the commonly used M’th power algorithms. Parallel architecture can increase the throughput of digital signal processing (DSP) modules for a given working frequency. pipelined architecture (PA) leads to a reduction in the critical path, and thus it can be exploited to increase the throughput of DSP modules by increasing the working frequency. It is demonstrated that in PA the working frequency is not limited by the computation time of the M’th power subfunction with the highest complexity because it is feedforward and thus pipelining registers can be introduced to reduce the critical path inside it. Instead, the phase unwrapping subfunction (PUS) becomes the bottleneck of the working frequency because it requires the immediately preceding result and cannot be implemented in PA, which results in the longest critical path among the DCS module. To solve this problem, we propose a feedforward look-up-table-based PUS design that can greatly reduce the critical path and increase the working frequency. Experimental DCS implementation in a Xilinx Virtex7 field programmable gate array shows that with this method the working frequency of the DCS module for quadrature phase-shift keying (QPSK) signals can be increased by 63.8%. Furthermore, using experimental and simulation data, it is demonstrated that the performance of the DCS module with increased working frequency is close to that of the off-line DCS algorithms for QPSK signals.