Unlock instant, AI-driven research and patent intelligence for your innovation.

Ship detection method based on local saliency features and cnn-svm

A ship detection and remarkable technology, applied in the field of image recognition, can solve the problem of low accuracy rate of ship detection, achieve the effect of improving the recall rate and improving the accuracy rate

Active Publication Date: 2020-11-10
TOPOTEK BEIJING TECH CO LTD
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a ship detection method based on local saliency features and CNN-SVM in order to solve the problem of ship detection accuracy in the complex background existing in the prior art. lower technical issues

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Ship detection method based on local saliency features and cnn-svm
  • Ship detection method based on local saliency features and cnn-svm
  • Ship detection method based on local saliency features and cnn-svm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] Below in conjunction with accompanying drawing and specific embodiment, the present invention is described in detail:

[0035] refer to figure 1 , a ship detection method based on local salient features and CNN-SVM, including the following steps:

[0036] Step 1) Build a training sample set:

[0037] Select M pieces of optical remote sensing ship images from the database, and use the image copied from each image with the center of the bow and the size of K×K as a positive sample image, and copy the image copied from each image with The positive sample images have the same size and do not contain the image of the bow as the negative sample image, and all the positive sample images and negative sample images form the training sample set, where M≥100, 32≤K≤48.

[0038] The size of remote sensing images is generally between 100×100 and 2000×2000, so the remote sensing images in most databases meet the above requirements. In the embodiment of the present invention, 300 op...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a ship detection method based on local saliency features and CNN-SVM, which is used to solve the technical problem of low ship detection accuracy under complex background existing in the prior art. The implementation steps are: 1. Construct a training sample set; 2. Construct a CNN-SVM ship head point classification model; 3. Construct a salient feature extraction operator; 4. Preprocess the image of the ship to be detected; 5. The ship to be detected Detect the salient feature points of the bow of the suspected ship in the image; 6. Extract the salient feature points of the bow of the ship to be detected; 7. Confirm the direction of the ship; 8. Obtain the ship detection result. The design and construction of the salient feature extraction operator of the present invention can effectively improve the recall rate of the bow point, and at the same time, use the CNN-SVM bow point classification model to improve the classification accuracy of the bow point, thereby improving the ship detection accuracy in complex scenes.

Description

technical field [0001] The invention belongs to the technical field of image recognition, and relates to an optical image ship detection method, in particular to a ship detection method based on local salient features and CNN-SVM, which can be used to detect ship targets under optical remote sensing images with complex backgrounds for identification and detection. Background technique [0002] In optical remote sensing images, object detection is a key research field in image object recognition and image understanding. By detecting and monitoring the ship target in the port area of ​​interest and estimating its position, size, type and other parameter information, it has broad application prospects in maritime rescue, port traffic management, sea area security, etc. The scale of the ship itself is quite different, and the direction is arbitrary, and it is too dense in the port and other areas, making ship detection a difficult point in remote sensing image target detection....

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/46G06K9/34G06K9/62G06N3/04
CPCG06V20/13G06V10/267G06V10/462G06N3/045G06F18/2411
Inventor 曾操唐小虎刘洋苏海龙
Owner TOPOTEK BEIJING TECH CO LTD