Youfang Huang

Two-Stage Classification Approach for Human Detection in Camera Video in Bulk Ports

Abstract With the development of automation in ports, the video surveillance systems with automated human detection begun to be applied in open-air handling operation areas for safety and security. The accuracy of traditional human detection based on the video camera is not high enough to meet the requirements of operation surveillance. One of the key reasons is that Histograms of Oriented Gradients (HOG) features of the human body will show great different between front & back standing (F&B) and side standing (Side) human body. Therefore, the final training for classifier will only gain a few useful specific features which have contribution to classification and are insufficient to support effective classification, while using the HOG features directly extracted by the samples from different human postures. This paper proposes a two-stage classification method to improve the accuracy of human detection. In the first stage, during preprocessing classification, images is mainly divided into possible F&B human body and not F&B human body, and then they were put into the second-stage classification among side human and non-human recognition. The experimental results in Tianjin port show that the two-stage classifier can improve the classification accuracy of human detection obviously.

Research on a Fast Human-Detection Algorithm for Unmanned Surveillance Area in Bulk Ports

With the development of port automation, most operational fields utilizing heavy equipment have gradually become unmanned. It is therefore imperative to monitor these fields in an effective and real-time manner. In this paper, a fast human-detection algorithm is proposed based on image processing. To speed up the detection process, the optimized histograms of oriented gradients (HOG) algorithm that can avoid the large number of double calculations of the original HOG and ignore insignificant features is used to describe the contour of the human body in real time. Based on the HOG features, using a training sample set consisting of scene images of a bulk port, a support vector machine (SVM) classifier combined with the AdaBoost classifier is trained to detect human. Finally, the results of the human detection experiments on Tianjin Port show that the accuracy of the proposed optimized algorithm has roughly the same accuracy as a traditional algorithm, while the proposed algorithm only takes 1/7 the amount of time. The accuracy and computing time of the proposed fast human-detection algorithm were verified to meet the security requirements of unmanned port areas.

A FAST AUTOMATED VISION SYSTEM FOR CONTAINER CORNER CASTING RECOGNITION

With the development of container port automation, the automated vision systems for containers have been widely used in automated ports. This paper presents a rapid automated vision system for container corner casting recognition. The histograms of oriented gradients (HOG) descriptors are used to preprocess the image of the container and the vectors of HOG are then built. A trained support vector machine (SVM) classifier is applied to recognize the right corner casting of the container. At last, through symmetry, a flipping mirror algorithm is used for quick left corner casting recognition. The experimental results show this algorithm scans and detect the two corner castings of the container almost twice as fast as the traditional algorithms.

Ship Identification Algorithm Based on 3D Point Cloud for Automated Ship Loaders

ABSTRACT Mi, C.; Shen, Y.; Mi, W., and Huang, Y., 2015. Ship identification algorithm based on 3D point cloud for automated ship loader. With the development of bulk port automation, the ship loader as the main quayside machine of bulk terminal is required for transformation from manual operation to automation. The ship identification method is a key inspection technique for automated ship loaders. In this paper, a fast ship identification algorithm was formulated based on the 3D point cloud of the ship, as generated by the Laser Measurement Systems (LMS) mounted on the ship loader. To meet the requirement of real-time computing for the automated ship loader, the 3D point cloud was first processed to reduce its dimensions from 3D point cloud into a 2D image. A projection method was then applied to locate and identify all bulk cargo holds in the ship. Finally, a group of experiments on ship identification was conducted using this algorithm in the Coal Terminal of Tianjin Port. The results showed that the computing time for a whole ship was lower than 200 ms and the error of the algorithm was lower than 10%, meeting the requirement of automated ship loaders.