Infrared image target detection method for unmanned aerial vehicle inspection

The invention discloses an infrared image target detection method for unmanned aerial vehicle inspection, which comprises the following steps of: firstly, constructing a Yov5 improved model and an infrared image data set for unmanned aerial vehicle inspection, and introducing a CBAM convolution atte...

Full description

Saved in:
Bibliographic Details
Main Authors LIU RONGWANG, WANG JIANRONG, TAN ZHISONG, FANG GUANREN, HU GUANYU
Format Patent
LanguageChinese
English
Published 30.07.2024
Subjects
CALCULATING
COMPUTING
COUNTING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention discloses an infrared image target detection method for unmanned aerial vehicle inspection, which comprises the following steps of: firstly, constructing a Yov5 improved model and an infrared image data set for unmanned aerial vehicle inspection, and introducing a CBAM convolution attention mechanism into a Yov5 reference model by the improved model so as to enhance the channel relevance of a feature map and the features of a region of interest; the influence of noise characteristics is effectively suppressed; secondly, a K-Means algorithm is adopted to calculate the anchor frame size of the infrared image data set so as to reduce the omission ratio and the false detection ratio of the model to the infrared target; secondly, training the Yolov5 improved model by introducing a complete intersection-parallel ratio positioning loss function and utilizing an infrared image data set oriented to unmanned aerial vehicle inspection; and finally, performing target detection on the image by using the infr
Bibliography:Application Number: CN202410307803