Convolutional neural network-based tire outer wall intact degree identification method

Abstract

The invention discloses a tire outer wall intact degree identification method based on a convolutional neural network. The method comprises the following steps: firstly, obtaining and carrying out image preprocessing on a tire outer wall image of which the damage is classified; then, an asymmetric convolution structure is constructed directly based on a DenseNet neural network model; constructing a feature fusion module to fuse target features of different levels, and inputting and fusing the tire outer wall image into a DenseNet neural network model; training, verifying and testing the DenseNet neural network model by the tire outer wall image in the DenseNet neural network model through a training strategy of 10-fold cross check; and finally, optimizing a cross entropy cost function by using an Adam algorithm, and optimizing the DenseNet neural network model. Testing is completed in the driving process, the testing efficiency is improved, the integrity of the outer wall of the tire can be quantified and judged, and the outer wall of the tire is recognized through an algorithm to screen out manual errors.

Description

technical field [0001] The present invention relates to the technical field of application of convolutional neural network, in particular to a method for identifying the integrity of tire outer wall based on convolutional neural network. Background technique [0002] At present, the detection of the integrity of the outer wall of the tire has not been covered in this field, and there is no quantifiable detection equipment for the integrity of the outer wall of the tire. It has always relied on manual visual inspection and manual judgment, and manual detection, The need is that the vehicle stays still on the parking space, and there are a series of practical problems such as low efficiency, easy omission and error. [0003] With the industrialization of society, the manufacture of special testing equipment to prevent tire damage is a defect that needs to be faced squarely. Contents of the invention [0004] Aiming at the above-mentioned technical problems in the related ar...

AI Technical Summary

Problems solved by technology

[0002] At present, the detection of the integrity of the outer wall of the tire has not been covered in this field, and there is no quantifiable detection equipment for the integrity of the outer wall of the tire. It has always relied on manual visual inspection and manual judgment, and manual detection, The need is that the vehicle stays still in the parking space, and there are a series of practical problems such as low efficiency, easy to miss and error-prone

Images