Trust-based multi-frame heterogeneous data fusion identification method

Abstract

The invention discloses a trust-based multi-frame heterogeneous data fusion identification method. The method comprises the following steps of training a corresponding classifier in a feature space ofeach identification framework; classifying the plurality of known training samples through each classifier to obtain a classification result of each known training sample, and respectively calculating a conversion relation between each identification framework and a target identification framework according to a corresponding relation between the classification result of each known training sample and a real classification result of the known training sample; converting the classification results of the target sample in different identification frameworks into a target identification framework through the conversion relationship, and fusing the classification results to obtain a final classification result of the target sample. Classification results obtained by different classifiers (located in two different identification frameworks) are converted into the same identification framework. The conversion rule is estimated through fusion by utilizing an evidence theory. Finally, fusiondecision of different framework data is realized, and the accuracy of heterogeneous data identification can be improved.

Description

[0001] 【Technical field】 [0002] The invention belongs to the technical field of target recognition, in particular to a trust-based multi-frame heterogeneous data fusion recognition method. [0003] 【Background technique】 [0004] The recognition of complex pattern systems based on classifier fusion tasks is an important and challenging field of current research, and one of the key issues is how to obtain more available knowledge and improve classification accuracy, especially in knowledge-unknown and complex patterns. in the classification system. The idea of ​​classifier fusion is that different classifiers can provide (more or less) complementary information to achieve higher classification accuracy. In the classifier fusion technology, the identification framework must be unified first, and two classification information under the same identification framework can be fused to increase their context information. However, due to the different object recognition frameworks ...

AI Technical Summary

Problems solved by technology

In reality, due to differences in the identification frameworks of different sensors, the relationship between different identification frameworks is unknown and non-linear. Therefore, it is difficult to achieve multi-frame heterogeneous data fusion recognition.

Images