This study develops a simple real-time algorithm inferring the relative distance between containers stacked in operational yards and spreaders of Rubber Tire Gantry (RTG) in support of fully automated RTG operations in ports. The proposed scheme is based on SOLOv2 instance segmentation network with image captured by camera mounted on RTG driver’s cab. Specifically, distance calculation is achieved by containers’ surface recognition via instance segmentation and geometric information pertaining to the containers. To meet the dual demands of accuracy and reduced resource consumption, the original backbone network of SOLOv2, namely resnet18, is replaced by GhostNet. This substitution enables the extraction of multi-level feature maps for fusion, significantly reducing network complexity without sacrificing accuracy. The proposed algorithm is validated by comparative experiments on a self-curated real-life dataset, demonstrating a remarkable reduction of 91.6% in FLOPS and 55.7% in Params for the improved SOLOv2 framework.

Rubber tire gantry (RTG) plays a pivotal role in facilitating efficient container handling within port operations. Conventional RTG, highly depending on human operations, is inefficient, labor-intensive, and also poses safety issues in adverse environments. This article introduces a multitarget detection and tracking (MTDT) algorithm specifically tailored for automated port RTG operations. The approach seamlessly integrates enhanced YOLOX for object detection and improved DeepSORT for object tracking to enhance the MTDT performance in the complex port settings. In particular, Light-YOLOX, an upgraded version of YOLOX incorporating separable convolution and attention mechanism, is introduced to improve real-time capability and small target detection. Subsequently, OSNet-DeepSORT, an enhanced version of DeepSORT, is proposed to mitigate ID switching challenges arising from unreliable data communication or occlusion in real port scenarios. The effectiveness of the proposed method is validated in various real-life port operations. Ablation studies and comparative experiments against typical MTDT algorithms demonstrate noteworthy enhancements in key performance metrics, encompassing small target detection, tracking accuracy, ID switching frequency, and real-time performance.