Infrared target instance segmentation method based on feature fusion and a dense connection network

Abstract

The invention discloses an infrared target instance segmentation method based on feature fusion and a dense connection network, and the method comprises the steps: collecting and constructing an infrared image data set required for instance segmentation, and obtaining an original known infrared tag image; Performing image enhancement preprocessing on the infrared image data set; Processing the preprocessed training set to obtain a classification result, a frame regression result and an instance segmentation mask result graph; Performing back propagation in the convolutional neural network by using a random gradient descent method according to the prediction loss function, and updating parameter values of the convolutional neural network; Selecting a fixed number of infrared image data training sets each time and sending the infrared image data training sets to the network for processing, and repeatedly carrying out iterative updating on the convolutional network parameters until the convolutional network training is completed by the maximum number of iterations; And processing the test set image data to obtain average precision and required time of instance segmentation and a finalinstance segmentation result graph.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an infrared surface target instance segmentation method based on feature fusion and dense connection network. Background technique [0002] The target instance segmentation is to give each pixel in the image the instance category of the pixel, and predict the category label and pixel-level instance mask to locate different numbers of instances in the image. Instance segmentation of an image can obtain the target information of the image, and better understand the high-level content and representative information of the image. Instance segmentation is one of the most important and challenging tasks. The positioning of specific targets in the image, image It is very useful for search, recognition of road scenes in automatic driving, and video surveillance, and has high practical value in practical applications. The existing basic ideas of instance segmentation...

AI Technical Summary

Problems solved by technology

However, the upsampling results used by the network are still blurry and smooth, and are not sensitive enough to the details in the image.

Images