A Method for Object Recognition in Remote Sensing Image
A technology of target recognition and remote sensing images, applied in the field of remote sensing image target recognition, can solve problems affecting target detection, target recognition performance, split feature description and feature matching internal relations, etc., to achieve the effect of improving accuracy and reducing ambiguity
Active Publication Date: 2019-05-03
INST OF AUTOMATION CHINESE ACAD OF SCI
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, these feature description methods ignore the semantic differences between the target and the background, and between different types of targets, and split the internal relationship between feature description and feature matching, which greatly affects the performance of target detection and target recognition.
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0020] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with specific embodiments and with reference to the accompanying drawings.
[0021] figure 1 It is a flowchart of a remote sensing image target recognition method according to an embodiment of the present invention, such as figure 1 As shown, the method includes the following steps:
[0022] Step S1, extracting appearance features at each pixel of each training image in the training image set;
[0023] In the present invention, the DAISY feature is used to characterize the appearance features of each pixel and its neighborhood.
[0024] Said step S1 further comprises the following steps:
[0025] Step S11, preprocessing the training images;
[0026] If the training image is a grayscale image, you can skip this step and directly extract the appearance features on the training image, that is to ...
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
Login to View More Abstract
The invention discloses a remote sensing image target recognition method, comprising: extracting appearance features from each pixel of each training image in a training image set; constructing a target-background ordered semantic pair by using the appearance features and type annotations of the training images, And learn the target-background semantic model; use the appearance features and type annotations of the training image to construct an ordered semantic pair of target types, and learn to obtain the target type semantic model; extract appearance features at each pixel in the test image; use the test image The appearance feature of the image constructs a target-background ordered semantic pair, and utilizes the target-background semantic model to extract the target area in the test image; constructs the target type ordered semantic pair for the appearance feature of the target area of the test image, and uses the The target type semantic model is described to determine the target type. The invention improves the separability between targets and backgrounds and different types of targets, and can be widely used in disaster monitoring, target reconnaissance and many other fields.
Description
technical field [0001] The invention relates to the technical fields of remote sensing image processing, feature matching, target detection and recognition, and in particular to a remote sensing image target recognition method. Background technique [0002] Remote sensing image target recognition is a technology that uses computers to detect targets from input images and judge the target type. Remote sensing image target recognition has a wide range of applications in disaster detection, target search, target location and so on. However, due to the complexity of remote sensing images and the semantic uncertainty of targets, it still faces many challenges to use computers to automatically detect and recognize targets from remote sensing images. [0003] The key technologies of remote sensing image target recognition include feature description and feature matching. Due to differences in sensor types, shooting conditions, and the variability of the appearance of target featu...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/46
CPCG06V20/13G06V10/462
Inventor 霍春雷潘春洪周志鑫
Owner INST OF AUTOMATION CHINESE ACAD OF SCI