Infrared and Visible Image Object Detection Method based on Generated Adversarial Network

Infrared image and visible image have many complementary features. Combining infrared and visible images for object detection can make full use of the complementary features from multi-band images and effectively improve the detection accuracy. In order to extract complementary features, a parallel convolutional neural network is designed. In parallel convolutional neural network, according to the characteristics of multi-band images, a deep network is designed to extract infrared features, while a shallow neural network is used to extract visible features. In order to improve the detection accuracy and effectively detect the difficult objects, an object detection method based on the generative adversarial network for infrared and visible images is proposed. The generative adversarial network is composed of generator and discriminator. The generator is set as the parallel convolutional object detection network, and the discriminator is set as the classification network. After continuous iterative adversarial training, the generative adversarial network can accurately detect objects in multi-band images.