Inland river ship freeboard detection method based on deep enhanced neural network

Abstract

The invention discloses an inland river ship freeboard detection method based on a deep enhanced neural network. A laser radar and a linked tripod head are set on a high pole on the bank of a river, the tripod head drives the laser radar to adjust a detection height and an orientation angle to scan one side of the ship, ship contour images under different detection parameters are obtained, a reinforcement learning neural network is used as an approximator of a reinforcement learning value function, and ship contour information is input into the reinforcement learning neural network to determine the actions of the current laser radar and the linked tripod head to correctly identify the freeboard of a current ship. According to the method, based on the recognition capability of a convolutional neural network for image recognition, a reinforcement learning algorithm is combined to construct the deep enhanced neural network, the existing technical deficiencies of a deep learning network inthe field of ship overload recognition are overcome, the detection ability of the laser radar for ship freeboard information is improved, and therefore, a technical support is provided for the automatic identification of inland river ship draught.

Description

technical field [0001] The invention relates to a method for detecting the freeboard of an inland river vessel based on a deep reinforced neural network, which belongs to the field of deep learning and the technical field of reinforcement learning. Background technique [0002] In recent years, ship overloading has brought more and more harm to water transportation. The existence of ship overloading has seriously affected the social reputation of water transport companies, hindered the improvement of the competitiveness of water transport companies, and interfered with the healthy development of the water transport industry. . At this stage, due to the constraints of technical means, it is difficult for maritime law enforcement agencies to quickly and accurately detect overloading of ships, resulting in repeated prohibition of overloading of ships. Aiming at the problems exposed by the present situation of overload detection of inland ships, the present invention adopts las...

AI Technical Summary

Problems solved by technology

However, in the training process of the deep learning network, there are often the following problems: Compared with training the traditional three-layer neural network, the deep learning network has an increased time cost due to the large amount of calculation and more parameters that need to be updated.

Secondly, when the output error of the deep learning network does not change much, the training process will slow down and take too long

Images