Motivated by the 2013 International UAV Innovation Grand Prix, we design and implement a real-time vision system for an unmanned helicopter autonomously transferring cargoes between two platforms. In the competition, four cargoes are initially placed inside four circles on one platform, respectively. They are required to be transferred one by one into the four circles on the other platform. This paper presents the core algorithms of the proposed vision system on ellipse detection, ellipse tracking, and single-circle-based position estimation. Experiments and the great success of our team in the competition have verified the efficiency, accuracy, and robustness of the algorithms. Our team was ranked first in the final round competition.

Multispectral object detection is a crucial technology in remote sensing image processing, particularly in low-light environments. Most current methods extract features at a single scale, resulting in the fusion of invalid features and the failure to detect small objects. To address these issues, we propose a multispectral object detection network based on multilevel feature fusion and dual feature modulation (GMD-YOLO). Firstly, a novel dual-channel CSPDarknet53 network is used to extract deep features from visible-infrared images. This network incorporates a Ghost module, which generates additional feature maps through a series of linear operations, achieving a balance between accuracy and speed. Secondly, the multilevel feature fusion (MLF) module is designed to utilize cross-modal information through the construction of hierarchical residual connections. This approach strengthens the complementarity between different modalities, allowing the network to improve multiscale representation capabilities at a more refined granularity level. Finally, a dual feature modulation (DFM) decoupling head is introduced to enhance small object detection. This decoupled head effectively meets the distinct requirements of classification and localization tasks. GMD-YOLO is validated on three public visible-infrared datasets: DroneVehicle, KAIST, and LLVIP. DroneVehicle and LLVIP achieved mAP@0.5 of 78.0% and 98.0%, outperforming baseline methods by 3.6% and 4.4%, respectively. KAIST exhibited an MR of 7.73% with an FPS of 61.7. Experimental results demonstrated that our method surpasses existing advanced methods and exhibits strong robustness.

Recent advanced trackers, composed of discriminative classification and dedicated bounding box estimation, have achieved remarkable advancements in performance of visual object tracking. However, existing methods cannot satisfy the demands of tracking tasks in complex scenes, such as occlusion, scale variations, and etc. To this end, we propose a novel online multi-scale classification and global feature modulation for robust visual tracking, which is developed over accurate tracking by overlap maximization, named ATOM+. First, coordinate attention (CA) is applied to enhance the target features in the channel dimension and spatial dimension, which can effectively optimize the feature representation ability of the backbone network. Second, an online multi-scale classification (OMC) module is designed. During the online tracking phase, more reliable matching responses are comprehensively generated by aggregating information from different scales related to the target. This new operation enables stable perception of the target by the tracker, particularly when severe changes in the appearance and posture of the target are encountered. Third, a global feature modulation (GFM) mechanism is constructed, which requires only a small amount of computational resources, to fuse the spatial contextual information of the template image into the search region. This integration refines the bounding box to obtain an accurate estimate of the target state. Finally, comprehensive experiments on conventional tracking benchmarks of OTB100, LaSOT, and VOT2018 show that our tracker can sufficiently address different challenging scenarios, and achieves state-of-the-art performance. For the average running speed, our tracker can achieve 37 FPS in real time.

Lane detection is a vital component of intelligent driving systems, offering indispensable functionality to keep the vehicle within its designated lane, thereby reducing the risk of lane departure. However, the complexity of the traffic environment, coupled with the rapid movement of vehicles, creates many challenges for detection tasks. Current lane detection methods suffer from issues such as low feature extraction capability, poor real-time detection, and inadequate robustness. Addressing these issues, this paper proposes a lane detection algorithm that combines an online re-parameterization ResNet with a hybrid attention mechanism. Firstly, we replaced standard convolution with online re-parameterization convolution, simplifying the convolutional operations during the inference phase and subsequently reducing the detection time. In an effort to enhance the performance of the model, a hybrid attention module is incorporated to enhance the ability to focus on elongated targets. Finally, a row anchor lane detection method is introduced to analyze the existence and location of lane lines row by row in the image and output the predicted lane positions. The experimental outcomes illustrate that the model achieves F1 scores of 96.84% and 75.60% on the publicly available TuSimple and CULane lane datasets, respectively. Moreover, the inference speed reaches a notable 304 frames per second (FPS). The overall performance outperforms other detection models and fulfills the requirements of real-time responsiveness and robustness for lane detection tasks.

The recent development of advanced trackers, which use nighttime image enhancement technology, has led to marked advances in the performance of visual tracking at night. However, the images recovered by currently available enhancement methods still have some weaknesses, such as blurred target details and obvious image noise. To this end, we propose a novel method for learning multidimensional spatial attention for robust nighttime visual tracking, which is developed over a spatial channel transformer based low light enhancer (SCT), named MSA-SCT. First, a novel multidimensional spatial attention (MSA) is designed. Additional reliable feature responses are generated by aggregating channel and multi-scale spatial information, thus making the model more adaptable to illumination conditions and noise levels in different regions of the image. Second, with optimized skip connections, the effects of redundant information and noise can be limited, which is more useful for the propagation of fine detail features in nighttime images from low to high level features and improves the enhancement effect. Finally, the tracker with enhancers was tested on multiple tracking benchmarks to fully demonstrate the effectiveness and superiority of MSA-SCT.