Deck personnel detection system and method based on panoramic stitching and deep learning

Abstract

The invention discloses a deck personnel detection system based on panoramic stitching and deep learning, and the system comprises: a network camera group which is used for carrying out the real-time omnibearing monitoring of deck personnel at a front crossbeam of a quay crane; and / or an analysis server which is used for carrying out visual analysis and identification on deck personnel under the front girder of the quay crane and carrying out panoramic splicing on monitoring results of the network camera group; and / or a video client which is used for simultaneously displaying the analysis results of the plurality of quay cranes and visually presenting the analysis and identification results. The invention further discloses a deck personnel detection method based on panoramic stitching and deep learning, the detection capability of the deck personnel is improved, the detection efficiency of the deck personnel in quay crane operation is improved, and the safety management and control level of quay crane operation is improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, and more specifically, to a deck personnel detection system and method based on panorama stitching and deep learning. Background technique [0002] Deep learning refers to a collection of algorithms that use various machine learning algorithms on multi-layer convolutional neural networks to solve various problems such as video and images. Deep learning can be classified into neural networks from a broad category. The core is feature learning, which aims to The layered network obtains hierarchical feature information, so as to solve the important problem of manually designing features in the past; the convolutional neural network is a kind of artificial neural network, and is currently a research hotspot in the field of image recognition. Its weight sharing network structure makes it more similar to the biological neural network, which reduces the complexity of the network ...

AI Technical Summary

Problems solved by technology

[0003] Panoramic stitching technology is the technology of stitching several overlapping images into a seamless panorama or high-resolution image. The traditional way to obtain panoramic images is to use a fisheye camera or a panoramic camera, in which the fisheye camera will have images It is distorted, unable to present the real situation of the scene, and has the same defects as panoramic cameras that can see everything but not clearly. Panoramic stitching is based on the use of traditional surveillance cameras, taking into account the problems of seeing everything and seeing clearly. On this basis, it can Use technology based on deep learning to detect and identify the behavior of people and objects in the scene. When the personnel on the lower deck of the front girder of the quay bridge are waiting in the area between the bridges, it is necessary to control the overtime personnel. At present, in the traditional control scheme, the control of the deck personnel is realized by observing the deck. This scheme will cause fatigue due to personnel. Omissions, false reports and other technical problems caused by problems such as responsibility, which lead to the emergence of management and control loopholes

Images