Binocular image rapid target detection method based on double-flow convolutional neural network

Abstract

The invention discloses a binocular image rapid target detection method based on a double-flow convolutional neural network, and the method comprises the steps: carrying out the calibration of a binocular camera, and obtaining calibration parameters; correcting a training image according to the calibration parameters, and training an implicit deep semantic mining network for implicitly learning deep semantic information on the binocular image and training a multi-modal feature hybrid detection network; combining the features output by the implicit deep semantic mining network with the featuresof the multi-modal feature hybrid detection network in a channel series connection manner to form a double-flow convolutional neural network, and training the double-flow convolutional neural networkby using the training image; and obtaining a test image through the binocular camera, correcting the test image, and inputting the corrected image into the double-flow convolutional neural network for target detection to obtain a target detection result. The complementarity of RGB and deep semantic information can be comprehensively utilized, and the method has the advantages of being high in efficiency and more accurate in target detection result.

Description

technical field [0001] The invention relates to the research field of target detection in video surveillance, in particular to a method for fast target detection of binocular images based on a dual-stream convolutional neural network. Background technique [0002] The task of target detection is to determine the position of the target object of interest in the image and identify its category. Object detection is the foundation of many advanced computer vision tasks such as face recognition and object tracking. For example, in the face recognition scene, the face needs to be detected first, and then the features can be extracted in a specific area to verify the identity; similarly, the target tracking also needs to detect the target position first, and then the feature similarity matching can be performed to track. object. At present, object detection has received a lot of attention from academia and industry, and is widely used in public security, smart cities, and autonom...

AI Technical Summary

Problems solved by technology

This is undoubtedly not an end-to-end method. The intermediate calculation process of the disparity map will make the application process cumbersome and affect the efficiency of practical applications.

Images