Ship detection method and system based on adaptive data enhancement

Abstract

The invention discloses a ship detection method and system based on adaptive data enhancement. The method comprises the following steps: acquiring a picture data set through a camera and performing data enhancement processing; based on a neural network model, carrying out picture feature extraction processing on the enhanced picture; carrying out multi-scale dimension attention calculation on the feature picture to obtain a feature picture with multi-scale dimensions; performing loss calculation on the feature picture with multiple scales and dimensions through a loss function to obtain a final loss value; updating a neural network model according to the final loss value, and constructing a ship detection model; and detecting the to-be-detected picture based on the ship detection model to obtain a detection result. On the basis of a visible light video detection technology, the detection perception capability of the ship body is improved by constructing the ship detection model, and the ship detection precision can be improved in a special detection environment. The ship detection method and system based on adaptive data enhancement can be widely applied to the technical field of ship detection.

Description

Technical field [0001] The present invention relates to the field of ship detection technology, in particular to ship detection methods and systems based on adaptive data enhancement. Background [0002] Maritime transport is an important part of international logistics, with the continuous development of China's economy, shipping traffic continues to increase, while various problems also follow, such as maritime violations, illegal transportation and other behaviors, further in order to ensure the continuity and timeliness of major ports and waterway supervision, it is necessary to introduce the technology of intelligent detection of ships, the existing ship target detection is mainly based on three technologies, including synthetic aperture radar, remote sensing satellites and visible light video, Although the existing ship object detection algorithm can achieve good results on the recognized standard data set, it is slightly insufficient if it is placed on a special category o...

AI Technical Summary

Problems solved by technology

[0002] Maritime transportation is an important part of international logistics. With the continuous development of our country's economy, the shipping traffic continues to increase. At the same time, various problems are also coming one after another, such as maritime violations and illegal transportation. In order to further ensure that major ports and waterways The continuity and timeliness of supervision must introduce the technology of ship intelligent detection. The existing ship target detection is mainly based on three technologies, including synthetic aperture radar, remote sensing satellite and visible light video. Although the existing ship target detection algorithm can be recognized Good results can be achieved on the standard data set, but if it is placed on a special category of detection tasks, it is slightly insufficient. For example, in ship detection tasks, some environmental variables, such as camera angles, weather conditions, water surface ripples and ship reflections, etc. Strong interference to test results

Images