Remote target temperature accurate measurement method based on monocular infrared stereoscopic vision correction

Abstract

The invention relates to a remote target temperature accurate measurement method based on monocular infrared stereoscopic vision correction, which belongs to the technical field of infrared image recognition and detection. The method comprises the following steps of acquiring an infrared image containing temperature field information through a monocular infrared camera, inputting the infrared image into a target 3D information extraction network, and performing feature extraction and infrared face target detection so as to obtain target three-dimensional information and temperature field information, and constructing a face temperature three-dimensional correction and compensation network, inputting the obtained three-dimensional information and temperature field information into the face temperature three-dimensional correction and compensation network, carrying out three-dimensional vision temperature correction, and outputting an infrared image with temperature information, a 3D frame and angle information. The detection range can be expanded, the temperature measurement error is reduced, and the cost is reduced.

Description

technical field [0001] The invention belongs to the technical field of infrared image recognition and detection, and relates to a method for accurately measuring the temperature of a long-distance target based on monocular infrared stereo vision correction. Background technique [0002] During the epidemic of infectious diseases, human body temperature is one of the important information for judging whether there is infection, and its data is very important. In order to quickly screen the temperature of the human body, there is an urgent need for a device that can quickly, efficiently and accurately measure the temperature of the human body in large public places, and this device must be able to isolate the risk of cross-infection. [0003] In order to solve the problems of long measurement time and poor accuracy in the infrared temperature measurement system, in 2017, Tian Qijie, a researcher at the Chinese Academy of Sciences, proposed a method of using a low-temperature b...

AI Technical Summary

Problems solved by technology

However, this method only considers the temperature compensation in a single scene, and does not consider the depth, angle and other information that have a great impact on temperature measurement in multi-scenes, and these information are quite critical for temperature correction, so it caused The problem of poor generalization ability

[0006] In order to solve the problem of poor generalization ability, consider the nonlinear influence of angle, depth and other information on the measurement results in temperature correction, and give full play to the advantages of deep learning, this invention proposes a long-distance target temperature correction based on monocular infrared stereo vision accurate measurement method

Images