Unmanned ship environment intelligent sensing method based on deep learning

Abstract

The invention relates to an unmanned ship environment intelligent sensing method based on deep learning. The method comprises steps of 1, acquiring a training image data set and a test image data set;2, constructing an unmanned ship environment perception model; 3, training an unmanned ship environment perception model by using the training image data set; 4, testing the precision of the unmannedship environment perception model by using the test image data set, judging whether the unmanned ship environment perception model reaches preset precision or not, if so, executing the step 5, and otherwise, returning to the step 3; and 5, acquiring a real-time image in the sight range of the unmanned surface vehicle, and inputting the real-time image into the unmanned surface vehicle environmentsensing model to perform real-time target identification, positioning and front feasible direction segmentation on the surrounding environment of the unmanned surface vehicle. Compared with the priorart, the method is advantaged in that target recognition and feasible navigation channel segmentation are achieved at the same time, the size of the input image is not restrained, and real-time performance and accuracy are both considered.

Description

technical field [0001] The invention relates to the technical field of unmanned boat environment perception, in particular to a deep learning-based method for intelligent perception of the environment of the unmanned boat. Background technique [0002] As a disruptive technology, the maritime unmanned system will occupy an important position in the future society and is a hot spot in the current industry and academia. At present, the maritime intelligent unmanned system has become an inevitable trend of industrial development, and will redefine the rules of the ship equipment industry. In the traditional shipbuilding industry, manual remote control accounts for more than 70%, requiring manual real-time control based on sensor information. In addition, due to the lag of ship control and the complexity of sea conditions, it is often necessary to have certain sailing experience to be competent for the post of captain. In some dynamic positioning scenarios that require precise ...

AI Technical Summary

Problems solved by technology

[0003] At present, the research on target recognition technology for unmanned boats continues. For example, Chinese patent CN107609601A discloses a ship target recognition method based on multi-layer convolutional neural network. This method is only applicable to multi-layer convolutional neural network to realize the Target recognition, although the target recognition accuracy of this method is high, but this method can only identify ship targets, and cannot achieve feasible channel segmentation in the direction of navigation

Images