Floor defect automatic detection method based on computer vision and deep learning

A computer vision and deep learning technology, applied in computing, image data processing, instruments, etc., can solve problems such as difficulty in improving accuracy and speed at the same time, detection effect easily affected by the environment, and poor algorithm adaptability.

Active Publication Date: 2019-06-07
北京亦庄大数据科技发展有限公司
View PDF9 Cites 23 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problems in the prior art, the present invention provides a method for automatic detection of floor defects, which solves the problem that the accuracy and speed of automatic detection

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
  • Floor defect automatic detection method based on computer vision and deep learning
  • Floor defect automatic detection method based on computer vision and deep learning
  • Floor defect automatic detection method based on computer vision and deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0102] figure 1 A schematic flow chart of the method for automatic detection of floor image defects based on computer vision and deep learning in this embodiment is shown.

[0103] The embodiment adopts the actual image of the wall of a residential area, the resolution is 7952*5304, and the file size is generally about 30M. Initially, 70 defect images and 169 defect targets were obtained. As shown in the figure, the green frame is a defect.

[0104] First, according to step 1, pre-train the defect detection model InceptionV2-SSD (single shot multibox detector) on the COCO training set. During the training process, the highest average precision index (COCO mAP) of the model is 24, freeze the model of this node, and obtain the model migration The learned initialization parameters.

[0105] According to step 2, the expansion coefficient a is set to 0.3, and the corrected sample of the initial sample set is as follows Figure 2 to Figure 5 shown (both original and enlarged). ...

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 floor defect automatic detection method based on computer vision and deep learning, which comprises the following steps of: before detection, making a sample by utilizing anexisting floor defect image, and training a floor defect detection model according to a transfer learning theory; the detection process comprises the following steps: firstly, segmenting and extracting a floor image wall body area according to traditional image processing; secondly, partitioning a wall body area of the floor image, detecting each sub-block by applying a model, and accelerating byadopting a parallel processing mode; integrating and converting a sub-block detection result, and marking the position of the defect on the original image; and finally, correcting the detection result, and making a new sample for re-training the model to improve the detection performance of the model. The problems that in the prior art, the speed and precision of floor image defect detection are difficult to achieve at the same time, the adaptability of a detection algorithm is not high, and the detection effect is easily affected by the photo shooting environment are solved.

Description

technical field [0001] The invention belongs to defect detection technology, in particular to an automatic detection method for defects on the external surface of a building based on computer vision and deep learning. Background technique [0002] There are major safety hazards in the loosening or cracking of building exterior wall tiles. It is of great significance to regularly detect wall anomalies and monitor and guarantee building safety. Computer vision uses the camera to acquire images as a tool, and relies on image processing, image analysis, pattern recognition, artificial intelligence and other technologies. It can get a lot of information from the acquired images without touching specific objects. Specific information such as object size is especially suitable for the detection of surface defects and geometric dimensions of products such as wall and floor tiles, so as to achieve comprehensive quality evaluation. At present, there are many researches on surface def...

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
IPC IPC(8): G06T7/00G06T7/11G06T7/155G06T7/187
Inventor 孙光民陈佳阳白云鹍关世奎李煜
Owner 北京亦庄大数据科技发展有限公司
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