Airborne target recognition model construction platform and airborne target recognition method and equipment

Abstract

The invention relates to an airborne target recognition model construction platform and an airborne target recognition method and equipment. The airborne target recognition model construction platformcomprises a storage server which is used for storing the actually measured data and the characteristic data to the corresponding classified training data sets; a training server which is used for inputting each sample of the classified training data sets to a deep learning network to perform characteristic training and generating the airborne target recognition model for the result of characteristic training when the generation instruction is received and reversely inputting the result of characteristic training to the deep learning network to perform characteristic training when receiving the back propagation instruction; and a recognition server which is used for performing recognition test on the result of characteristic training and performing statistics of the recognition rate for the recognition test result and transmitting the model generation instruction when the recognition rate is not less than the preset recognition reference value or transmitting the back propagation instruction. According to the scheme, the accuracy of airborne target recognition can be effectively enhanced.

Description

technical field [0001] The invention relates to the technical field of image target recognition, in particular to an airborne target recognition model building platform, an airborne target recognition method and equipment. Background technique [0002] At present, in the down-looking target recognition of platforms with limited airborne resources, such as unmanned aerial vehicles, template matching methods are basically used to assist discrimination manually. The template matching method mainly uses the characteristics of the target to construct a characteristic template, such as: constructing a car characteristic template based on the four wheels of a car, constructing a tree characteristic template based on the leaves and branches of a tree, etc., by combining the constructed characteristic template with Targets are matched to identify targets. The discrimination accuracy of this traditional method depends heavily on the accuracy of feature selection and feature modeling,...

AI Technical Summary

Problems solved by technology

The discrimination accuracy of this traditional method depends heavily on the accuracy of feature selection and feature modeling, which is limited by human experience.

Images