Deep learning-based multi-camera image acquisition article recognition method

Abstract

The invention discloses a deep learning-based article recognition method for collecting images by multiple cameras. The method relates to the technical field of computer vision and intelligent recognition, and comprises the following steps: firstly, collecting images of a to-be-identified object through a plurality of cameras, and dividing the collected images into a training set image, a verification set image and a test set image; then establishing a multi-input algorithm model based on deep learning; training the multi-input algorithm model by using the training set image, and verifying a model effect through the verification set image; and finally, testing the multi-input algorithm model obtained through the training set image on the test set image to obtain a final article recognitionalgorithm model. According to the invention, multiple cameras are arranged to acquire images at multiple angles, complete features of the to-be-identified article are obtained, the multi-branch convolutional neural network model is established, and multiple images are input into the model at the same time, so that the effect of article classification and recognition is significantly improved.

Description

technical field [0001] The present invention relates to the technical fields of computer vision and intelligent recognition, and in particular to a deep learning-based object recognition method for collecting images by multiple cameras. Background technique [0002] Object recognition is an important application of computer vision technology. With the continuous development of artificial intelligence, object recognition has been widely used in daily life. Item recognition is closely related to deep learning theory, from manual feature extraction before the emergence of deep learning theory, to feature extraction and classification of items using convolutional neural network structures, and then to end-to-end models using a network to complete all tasks, Finally, the effect of real-time object detection is achieved with faster speed and lower resource consumption. At the same time, convolutional neural network, as one of the representative algorithms of deep learning, is a k...

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is to overcome the problem in the prior art that a single camera is easy to lose other acquisition angles of the item to be identified, resulting in lower accuracy of item identification

Images