Multiple target detection algorithm based on SSD networks
CAI Han ming1, ZHAO Zhen xing1, HAN Lu1, ZENG Xiang yong2
(1.College of Mechanical and Electrical Engineering, Qingdao University of Science and Technology, Qingdao 
266061, China;2.Bejing Interjoy Technology Limited, Beijing 100089, China)

Abstract: Aiming at the problems of heavy computational tasks and slow speed of operation during visual robots or intelligent terminal handling multiple target detection algorithm in modern factory, network communications technology was investigated. In order to establish the relationship between the intelligent terminal and the cloud, TCP/IP protocol were researched, and this method was proposed that the image data collected from the terminals should be preprocessed and sent to the cloud by Socket multithread communication based on the TCP, then the image data should be processed in the cloud with multi target detection of SSD network model, and the results were returned to the terminal at last. This algorithm was tested on the processing time by computer LAN or intelligent terminal online on the processing time. The results indicate that online measuring is slower, but it meets the practical needs. Intelligent terminal online detection reduces the requirements for the robot terminal hardware, and the data recycled can be reused, besides, the method can also make the algorithm dynamically be upgraded.

Key words: target detection; convolutional neural network; single shot multibox detector; intelligent robot; Socket network communication