Sea-surface ship object detecting and extracting method of optical remote sensing image

Abstract

The invention discloses a sea-surface ship object detecting and extracting method of an optical remote sensing image, and aims at reducing the false alarm rate effectively, extracting ship objects of different sizes rapidly and accurately, obtaining amount and position information of the objects, and being low in computing complexity. Multi-vision significance is detected on the basis of a frequency-domain model, a hyper complex frequency domain transformation model and a quaternion Fourier transform phase spectral module are fused in a weighted manner to overcome disadvantages of the two models and enhance advantages of the two models, and further sea-surface background interface is inhibited, the integral continuity of detected objects and differentiation performance among the objects are enhanced, and the target area of the sea surface is searched effectively. False alarm against possible heavy cloud layers and islands in the images is reduced, an improved histogram in the gradient direction is used to represent the distribution feature of the gradient structure of the object, the detected objects are discriminated according to established rules and conditions, whether a detected object is a ship is determined, the false alarm rate is reduced greatly, and the detecting accuracy is improved.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing and analysis, in particular to a method for detecting and extracting a sea surface ship target in an optical remote sensing image. Background technique [0002] In recent years, with the rapid development of earth observation technology, a large number of optical remote sensing imaging satellites with high spatial resolution have emerged, such as: SPOT-5, IKONOS, GeoEye, Quickbird, WorldView, Pleiades series, Skysat series, etc. Panchromatic images with meter-level resolution; and aerial images such as drones can achieve high-precision target acquisition near the ground. Aerospace remote sensing provides an extremely rich data source for target detection and recognition in sea areas. As ships are important targets for maritime monitoring and wartime strikes, detecting and identifying them can monitor the distribution of ships in key sea areas, grasp the enemy's combat stren...

AI Technical Summary

Problems solved by technology

[0003] However, in the actual optical remote sensing images, due to the lack of prior knowledge of the target and the background, the remote sensing shooting distance is long, and compared with the ship target, the sea surface background is complex, especially in bad weather conditions, clouds, sea fog, and poor lighting. Uniformity and other effects are serious, resulting in image quality degradation, and sea surface ship targets are weak; sea clutter, sea surface reflected light, shadows, ship wakes and other interference factors (such as garbage floating objects, small islands, etc.) will also affect the detection results. It is easy to cause false alarms and missed detections; in addition, ship types and ship materials are diverse, the black and white polarity of ship targets, and the brightness and texture characteristics of different parts of the same ship target may also vary greatly. Ship target detection has brought challenges. How to quickly, accurately and robustly detect and extract ship target areas in the ocean background and win more possible response and processing time has become an urgent problem to be solved

[0004] At present, the methods for extracting target areas of ships on the sea surface are mainly summarized as follows: Traditional optical remote sensing image ship detection is mostly based on grayscale statistical features and edge information segmentation methods, which are suitable for calm seas, uniform textures, and gray water bodies However, for complex situations such as large waves, cloud cover, and black and white polarity of ships, it is easy to cause false alarms; based on fractal models and fuzzy theory methods, The difference in fractal characteristics between natural background and artificial background can be used for detection, but when the cloud and fog interfere, the self-similarity of the background is reduced, and the fitting error of the fractal model is large; the method based on machine learning can deal with various complex environmental backgrounds, but This type of method often involves the extraction and search and matching of multiple complex features, which is time-consuming and difficult to meet the requirements of fast processing; the target detection method based on the visual attention mechanism can quickly perceive the information related to the current scene and task, It is mainly divided into space domain and frequency domain models. The space domain model method is to extract various features of the image and fuse them for target detection. However, when it is applied to optical remote sensing ship target detection, the target is relatively small and is easily affected by sea conditions, weather, and light. and other interferences, and the background suppression ability is relatively weak, and the time-consuming is relatively large, while the saliency detection method based on the frequency domain has obvious advantages in calculation speed and background suppression, but it is difficult for the interior of the target (especially when the ship target is large) Continuity and inter-target distinguishability when targets are close cannot be compromised

In addition, when there are interferences such as heavy clouds and islands, the false alarm rate of the above method will be greatly increased. Although there are some methods for removing cloud interference, they still cannot remove heavy clouds.

Images