To reduce the computation required in determining the proper scale of salient object, a fast visual saliency based on multi‐scale difference of Gaussians fusion in frequency domain (MDF) is proposed. First, based on the phenomenon that the foreground energy is highlighted and densely distributes on certain band of spectrum, the scale coefficients of foreground in an image can be literately approximated on the amplitude spectrum. Next, relying on the linear integration property of Fourier transform, the feature spectrum is obtained through the weighted infinite integral of difference of Gaussian feature maps with respect to the scale of object. Then, the saliency of each channel is obtained from feature spectrum by the inverse Fourier transform and scale filtering. Finally, through the channel integration, the MDF saliency map is obtained. Experiments on Li‐Jian data set demonstrate that combined with most appropriate colour space and scale filter, MDF achieves obvious acceleration (5.4 times faster than frequency domain analysis and spatial information) while getting desired accuracy (area under the curve, 0.8814 at Li‐Jian data set), which achieves the best accuracy efficiency trade‐off.
Automated lane detection is a vital part of driver assistance systems in intelligent vehicles. In this study, a multi‐lane detection method based on omnidirectional images is presented to conquer the difficulties stemming from the limited view field of the rectilinear cameras. The contributions of this study are twofold. First, to extract the features of the lane markings under various illumination and road‐surface scenarios, a feature extractor based on anisotropic steerable filter is proposed. Second, a parabola lane model is used to fit the straight as well as curved lanes. According to the parabola lane model, the straight lines and curves of feature maps can be represented as straight lines in a linear space coordinate system. Then lane modelling can be treated as an optimisation question in linear space and the parameters of lanes can be estimated by minimising the objection function. The method has been tested on publicly available data sets and the real car experiments. Experimental results show that the proposed method outperforms state‐of‐the‐arts approaches and obtains a detection accuracy of 99% in real world scenes.
Range-Doppler (RD) imaging has eclipsed synthetic aperture radar (SAR) imaging as the latest hotspot in the field of radar image object detection, owing to its low cost, high speed, and broad application scope. However, RD images are often of low quality due to the loss of effective features. Aiming at the problem of insufficient accuracy of the existing deep-learning-based sea surface RD image object detection, this article presents an improved YOLOv8 object detection algorithm for RD images based on feature enhancement and attention mechanism (YOLO-RDFEA). First, we have designed a feature extraction network DarknetSD with fewer parameters, to provide fine-grained information and compensate for the lack of abstract information. In addition, by introducing the coordinate attention (CA) mechanism in the feature-fusion stage, the model’s attention to spatial and channel features is improved. Moreover, the classification loss is improved using the slide loss function, which enhances the algorithm’s focus on the features of hard examples. Finally, comprehensive tests and evaluations are conducted using a self-built RD image dataset. Compared with the YOLOv8 baseline, the YOLO-RDFEA algorithm significantly reduced the misdetection of ships, its R was elevated by 17.9%. The all-category F1 score increased by 5.1% and mAP0.5 improved by 3.4%, which proved that the algorithm improves the detection and identification performance of all object categories. At the same time, the number of model parameters was reduced by 65.1%, which provides some basis for the algorithm deployment on the hardware platform.
Detecting maritime ship targets in complex aerial images is significantly challenging due to the relatively small size of the ship compared to the image’s predominant background. Current methodologies used to identify maritime targets in aerial imagery are susceptible to disruption from intricate backgrounds, leading to misidentification, overlooked detection, and inaccurate target localization. To address these challenges, we proposed the maritime background suppression network (MBSDet), which facilitates the reliable detection of multi-scale targets in intricate backgrounds. Firstly, we developed a background suppression (BS) module that integrates the Transformer’s context-aware capabilities with local attention features to enhance the model’s ability to identify multi-scale objects. Secondly, we present the multidimensional feature enrichment (MFE) module, which employs feature pyramid-based rearrangement modules and dynamic upsampling operators instead of conventional upsampling methods to enhance feature richness while attenuating background features. The mean average precision (mAP) values can reach 90.54% and 88.89% on HRSC2016 and DOTA v1.0 datasets, respectively. The experimental results indicate that MBSDet exhibits outstanding detection and localization accuracy for objects in difficult background situations.
A sparse antenna array of subarrays can effectively reduce the digital channels of array antennas, system complexities, and hardware cost while simultaneously increasing the antenna aperture. In this study, a new optimal design is proposed for a sparse antenna array of subarrays in the full-phased multiple input multiple output (FPMIMO) operation mode based on genetic algorithm–particle swarm optimization (GA–PSO) and ambiguity functions. The proposed algorithm can adaptively adjust the number of optimization iterations for yielding the optimization results of the PSO algorithm and GA to ensure the global optimization performance of algorithms and combine ambiguity functions to determine the final optimized sparse antenna array of subarrays. The effectiveness of the proposed algorithm is verified via simulation tests.