Information source adaptive dynamic selection-based efficient fusion identification method
A fusion recognition and adaptive technology, applied in the field of target recognition, can solve problems such as poor robustness, high cost, optimization of non-local optimum, etc., to achieve the effects of avoiding limitations, saving costs, and reducing quantities
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Publication Date
- 2017-10-20
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of target recognition, and specifically relates to an efficient fusion recognition method based on adaptive dynamic selection of information sources.
Background technique
[0002] With the rapid development of modern science and technology and its increasingly widespread application in the military field, fundamental changes have taken place in traditional combat thinking and combat methods. Strategic early warning has become an important guarantee for national security and strategic military operations, and an indispensable foundation for national strategic defense and deterrence. Target recognition technology is an important technical support method for radar intelligence and informatization. In modern warfare, target recognition technology has broad application prospects in military fields such as early warning and detection, precision guidance, battlefield command and reconnaissance, and identification o...
Examples
Embodiment Construction
[0041] The present invention will be described in detail below with reference to the drawings and specific embodiments.
[0042] The invention discloses a method for self-adaptive and dynamic selection of source information for efficient fusion identification, which combines figure 1 , figure 2 As shown, it specifically includes the following steps:
[0043] Step 1. Collect data from the training sample set through multiple sensors, and perform preprocessing and feature extraction on the collected data. For each training sample in the training sample set, its attributes are divided into N attributes according to the same rules Set, namely {a 1 ,a 2 ,...,A N }, where N is an integer greater than 0, a N Represents the Nth attribute set.
[0044] Step 2: Cross-validate the training sample set through each attribute set, and obtain the attribute set that makes the classification accuracy of the training sample set the highest.
[0045] Step 3: Classify the observation target according t...