A Classification Method of Nut Surface Defects Based on EfficientNet

A defect classification and nut technology, applied in image analysis, image enhancement, instruments, etc., can solve problems such as low accuracy, poor generalization ability, and low efficiency, and achieve the effect of avoiding time-consuming, enhancing adaptability, and avoiding influence

Active Publication Date: 2021-12-10
中科海拓(无锡)科技有限公司
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, most factories use manual inspection to detect surface defects of nuts, which is inefficient
However, the nut surface defect detection algorithm based on traditional vision is susceptible to interference from external environment and other factors, with low accuracy and poor generalization ability.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Classification Method of Nut Surface Defects Based on EfficientNet
  • A Classification Method of Nut Surface Defects Based on EfficientNet
  • A Classification Method of Nut Surface Defects Based on EfficientNet

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] Classification based on a defect in the nut EfficientNet surface, comprising the steps of:

[0034] Step a: collecting training samples, production data collection; EfficientNet model training, training the model converge, and then classified by a nut face data set;

[0035] Step two: collecting target image, cropping the image of the target;

[0036] Step three: Modify loss function focal loss, adjusting the positive and negative sample weights to enhance the accuracy of the model;

[0037] Step Four: preprocessing the image to prepare a template image rotation correction on the input image, and calculating the difference obtained difference image, as the three three-channel image fusion model as an input image;

[0038] Step Five: model reasoning using reasoning model EfficientNet input image;

[0039] Step Six: inference step 5 outputs the image to the host computer display.

[0040] Specifically, in a step, collecting training samples, production data collection, produc...

Embodiment 2

[0043] Classification based on a defect in the nut EfficientNet surface, comprising the steps of:

[0044] Step a: collecting training samples, production data collection; EfficientNet model training, training the model converge, and then classified by a nut face data set;

[0045] Step two: collecting target image, cropping the image of the target;

[0046] Step three: Modify loss function focal loss, adjusting the positive and negative sample weights to enhance the accuracy of the model;

[0047] Step Four: preprocessing the image to prepare a template image rotation correction on the input image, and calculating the difference obtained difference image, as the three three-channel image fusion model as an input image;

[0048] Step Five: model reasoning using reasoning model EfficientNet input image;

[0049] Step Six: inference step 5 outputs the image to the host computer display.

[0050] Specifically, in step II and then input to resize the image by a method using the crop m...

Embodiment 3

[0052] Classification based on a defect in the nut EfficientNet surface, comprising the steps of:

[0053] Step a: collecting training samples, production data collection; EfficientNet model training, training the model converge, and then classified by a nut face data set;

[0054]Step 2: Collect the target image, crop the target image;

[0055] Step 3: Modify the loss function is Focal LOSS, adjust the positive and negative sample weight, improve the accuracy of model;

[0056] Step 4: The image preoperation is made, and the template image rotates the input image, and the difference is obtained, and the three is fused into a three-channel image as a model input image;

[0057] Step 5: Model reasoning, use the EfficientNet model to reason to introduce the input image;

[0058] Step Six: Output the image output to the host computer display in step five.

[0059] Specifically, in industrial production, the probability of defective products is small, and the normal samples typically ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a nut surface defect classification method based on EfficientNet, comprising step 1: collecting training samples and making a data set; training the EfficientNet model through the data set, after the model training converges, and then classifying the nut surface; using EfficientNet to classify the model The scale is self-adjusted to avoid time-consuming manual adjustment and it is difficult to obtain the optimal solution; Step 2: Collect the target image and crop the target image; Step 3: Modify the loss function to focal loss, adjust the weight of positive and negative samples, and improve the accuracy of the model; Avoid the impact of positive and negative sample imbalance on the model; step 4: image preprocessing, make a template image to rotate and correct the input image, and make a difference to obtain a difference image, and fuse the three into a three-channel image as the model input image; Enhance the adaptability of the model to the production environment and improve the generalization ability of the model. Step five: model inference, use the EfficientNet model to infer the input image; step six: output the inference image to the host computer for display.

Description

Technical field [0001] Technical Field The present invention relates to the classification of surface defects nut, particularly to classification based on a defect in the surface of the nut of EfficientNet. Background technique [0002] With the development of science and technology, industrial production level of automation is also rising. In nut production, improve production efficiency at the same time, how to improve the detection efficiency has become urgent; currently, most of the factories have adopted the way of manual inspection of quality inspection, such tests and staff experience, working status and other factors effects, lack of objectivity and the workload, the detection efficiency is low; the traditional visual inspection method for defect detection algorithms generally lower nut accuracy, poor generalization is difficult to adapt the plant environment. Therefore, the defect classification technology based nut EfficientNet has important practical significance. [0...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/00G06T3/60G06T7/11G06T7/136G06T7/194G06T7/187G06K9/62
CPCG06T7/0004G06T3/608G06T7/11G06T7/136G06T7/194G06T7/187G06T2207/20081G06T2207/20084G06T2207/30164G06F18/251G06F18/24G06F18/214
Inventor 李子杰程坦刘涛吕剑
Owner 中科海拓(无锡)科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap