Ship Detection Method Based on Block of Interest Extraction in Remote Sensing Image

Abstract

The invention discloses an optical remote sensing image ship detection method based on interest block extraction, which mainly solves the problems of low detection accuracy and many false alarms in the prior art. The implementation plan is as follows: construct a ship detection data set of optical remote sensing image; down-sample wide remote sensing image and dehaze enhancement, use context information and image global features to perform water and land segmentation; use the constructed data set to train the target detection model based on SCRDet ;According to the water and land segmentation results, use partially overlapping sliding windows to scan the original wide-width remote sensing image to extract the region of interest as the region to be detected, and input the image of the region to be detected into the detection model to obtain the region detection result; map the region result to the original wide-width image In terms of scale, the improved non-maximum suppression is used to optimize the preliminary detection results; the detection results are optimized again according to the structural characteristics of the ship. The invention has high detection precision and low false alarm rate, and can be used to acquire interested ship targets and their positions in large-format remote sensing images.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an optical remote sensing image ship detection method, which can be used for target recognition in large format remote sensing images. Background technique [0002] Target detection of optical remote sensing images is one of the important issues in the field of remote sensing image research, and ship target detection has extremely important application value in fishery management, military reconnaissance and strategic deployment due to its particularity and criticality. Ship target detection is to determine whether there is a ship in the water or on the shore from a complex scene, and to locate it. [0003] Traditional ship target detection methods mainly include methods of using sea and land segmentation and using prior geographic information. The adaptability is not strong when the target is detected simultaneously. [0004] In recent years, deep learning has develop...

AI Technical Summary

Problems solved by technology

However, the method based on deep learning does not segment the image into water and land, but directly inputs the wide image slice into the trained model to generate the result. Inputting the complex land area that does not contain water into the detection model not only reduces the detection efficiency, but also It may cause obvious false alarms on land. In addition, the hull incompleteness caused by the segmented image will also cause obvious missed detection, so the detection accuracy is not high.

Images