FEATURE FUSION AND DENSE CONNECTION BASED INFRARED PLANE TARGET DETECTION METHOD

摘要

An infrared plane target detection method based on feature fusion and dense connection, comprising: constructing an infrared image dataset containing a target to be recognized, calibrating the position and kind of said target to be recognized in the infrared image dataset, and obtaining an original known label image; dividing the infrared image dataset into a training set and a verification set; performing image enhancement preprocessing on images in the training set and performing feature extraction and feature fusion, obtaining a classification result and a bounding box through a regression network; performing a loss function calculation on the classification result and the bounding box with the original known label image, and updating parameter values of convolutional neural network (CNN); repeating iteration update on the CNN parameters until the error is small enough or the number of iterations reaches a set upper limit; and processing images in the verification set through the trained CNN parameters to obtain accuracy and required time for target detection, and a final target detection result graph.