A typical object detection method and system in complex background based on deep learning
A technology with typical targets and complex backgrounds, applied in neural learning methods, biological neural network models, instruments, etc., it can solve the problem that the image background affects the target detection effect, and achieve the effect of improving the detection effect, reducing interference, and suppressing the impact.
Active Publication Date: 2022-03-22
武汉东沃慧达科技有限公司
View PDF22 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] In view of this, the present invention proposes a typical target detection method, system, device and storage medium based on deep learning in a complex background, which is used to solve the problem that the complex image background seriously affects the target detection effect in the existing image target detection
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
[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the implementation manners in the present invention, all other implementation manners obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.
[0052] see figure 1 , the present invention proposes a typical target detection method in a complex background based on deep learning, the method comprising:
[0053] S1. Collect image data containing typical targets and perform preprocessing to make a data set;
[0054] Specifically, a large amount of image data containing typical targets is collected, and the typical targets include but not limited to airplanes, ships, storage tanks, bridges, sm...
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 typical target detection method and system based on deep learning in complex backgrounds. The method includes: collecting image data containing typical targets and performing preprocessing to make a data set; through a backbone network and a bidirectional multi-scale feature fusion network A multi-feature selection module based on the attention mechanism, a classification sub-network and a regression sub-network construct a target detection network model; design a loss function based on the classification task and regression task, and train and test the target detection network model through the data set; after passing the test The object detection network model of the test image performs typical object detection. The invention can accurately detect multi-scale typical targets and perform angle prediction.
Description
technical field [0001] The invention belongs to the technical field of target detection, and in particular relates to a typical target detection method, system, device and storage medium based on deep learning in a complex background. Background technique [0002] Object detection technology is an important research direction of computer vision, which is widely used in environmental management, regional planning, and military affairs. It is of great significance to make high-precision predictions for typical targets such as airplanes, ships, storage tanks, bridges, small vehicles, and large vehicles. [0003] Target detection technology can generally be divided into two types, the detection algorithm combining traditional image processing and machine learning and the algorithm based on deep convolutional neural network (CNN). The former is mainly to manually design features, which is not effective in detecting complex image objects. In recent years, deep learning has devel...
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): G06V10/764G06V10/80G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V2201/07G06N3/045G06F18/24
Inventor 韩金波袁理韩鲁诚李增非
Owner 武汉东沃慧达科技有限公司