An approximate system for evaluating real-time port operations based on remote sensing images

Yajun Ping,Jiashen Zhao,Hongxin Shi,Qiang Zhang,Bo Xiang

DOI: https://doi.org/10.1080/2150704x.2020.1820612

IF: 3.531

2020-12-10

International Journal of Remote Sensing

Abstract:As an important transportation infrastructure and a strong support for foreign trades, ports play an indispensable role in economic development. However, the acquisition of operating data has a certain hysteresis, making it difficult to analyse and evaluate in real-time. To solve this problem, in this paper, we proposed a new perspective where the number of vessels was used to replace the traditional operation data, which are not easily available, to reflect the health of the economy. We designed an approximate system for evaluating real-time port operations based on remote sensing images. First, the port’s remote sensing images were acquired and pre-processed in real-time. Then, the vessels in the remote sensing images were identified using object detection technology. Finally, the correlation between the vessels and the port’s operating indicators were established to analyse and predict the port operating conditions. The result demonstrated that the number of ships reflected the change of the port throughput to a certain extent and the predicted accuracy was within 10%. This system effectively solved the problems of operating data hysteresis and the difficulty in acquiring, and it opens large prospects for practical applications.

imaging science & photographic technology,remote sensing

What problem does this paper attempt to address?

The paper attempts to address the issue of lagging and difficult-to-obtain port operation data. Traditionally, port economic evaluation indicators (such as throughput, cost, and profit) are obtained through reports released by port authorities on a monthly or quarterly basis, which have significant lag and are difficult to analyze and evaluate in real-time. To solve this problem, the authors propose a real-time port operation evaluation system based on remote sensing images.

Specifically, the paper proposes the following points:
1. **New Perspective**: Using the number of ships to replace traditional operational data to reflect the port operation status.
2. **System Design**: Designed a real-time port operation evaluation system based on remote sensing images, including an image preprocessing module, a ship detection module, and an operation evaluation module.
3. **Technical Implementation**: Used the Faster R-CNN algorithm in deep learning for ship detection, and processed remote sensing images through techniques such as geometric correction and image stitching.
4. **Data Analysis**: Conducted correlation analysis between the detected number of ships and traditional economic indicators (such as cargo throughput, container volume, and financial reports), and established a prediction model.

Through these methods, the system can effectively solve the problems of lagging and difficult-to-obtain traditional operational data, and has demonstrated high accuracy and feasibility in practical applications.