Recognition counting detection algorithm for circle-like objects based on machine vision and deep learning

A deep learning and machine vision technology, applied in computing, character and pattern recognition, instruments, etc., can solve problems such as low work efficiency, error-prone, large errors, etc., and achieve high accuracy

Active Publication Date: 2022-03-15
NANTONG UNIVERSITY
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Purpose of the invention: for the deficiencies in technology, the purpose of the present invention is to provide a circle-like object recognition and counting detection algorithm based on machine vision and deep learning based on machine vision and deep learning, and optimize the counting and detection method of CN 106529551 A, which is more Efficient and intelligent, solving the problems of high-intensity manual counting, low work efficiency, large errors and other error-prone problems in the existing technology

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
  • Recognition counting detection algorithm for circle-like objects based on machine vision and deep learning
  • Recognition counting detection algorithm for circle-like objects based on machine vision and deep learning
  • Recognition counting detection algorithm for circle-like objects based on machine vision and deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0099] A circle-like object recognition counting detection algorithm based on machine vision and deep learning, such as figure 1 shown, including the following steps:

[0100] Step 1: Use an industrial camera to collect a vertical cross-sectional image of the filter rod object, that is, the real object captured by the industrial camera.

[0101] Step 2: Obtain the region of interest of the filter rod object from the obtained vertical section image by using adaptive binarization and FindContours() function.

[0102] (1) Adaptive threshold binarization method to obtain binary image

[0103] Such as figure 2 For the vertical section image without any processing, the binary image obtained by the adaptive binarization algorithm is as follows image 3 As shown, the algorithm is as follows:

[0104] Use P(n) to represent the gray value of the nth point, and T(n) to represent the value after binarization. Use f s (n) to represent the sum of the gray values ​​of the s points befo...

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 circle-like object recognition and counting detection algorithm based on machine vision and deep learning, which includes adopting an industrial camera to collect the vertical section image of the filter rod object to be counted; The FindContours() function obtains the region of interest of the filter stick object; cuts the obtained region of interest into a A*A pixel submap; puts the obtained a A*A pixel submap into the improved SAA‑unet model Carry out training; The a A*A pixel subgraphs that have been trained are restored, and the region of interest is regained; the region of interest that is regained is carried out to the counting of the filter stick object, and the present invention adds SAA-unet mathematical theory, The counting detection method based on the principle of structural construction to improve the efficiency of detection is more efficient and intelligent, and solves the problems of high-intensity manual counting, low work efficiency, large errors and other error-prone problems. The accuracy of the algorithm is as high as 98.7%.

Description

technical field [0001] The invention relates to a recognition, counting and detection algorithm, in particular to a recognition, counting and detection algorithm for circle-like objects based on machine vision and deep learning. Background technique [0002] Filter rod objects are ubiquitous in life, such as rods, medicines, cigarette packing, etc. This filter rod object inevitably needs to be counted in daily life; at present, for the boxed filter rod object technology, manual direct counting or no detection is often adopted, and the counting method using manual direct detection method is labor-intensive, easy to fatigue the eyes, and has low counting accuracy , and the error varies from person to person. Especially when testing a large number of filter rods, severe vision deterioration and inattention discomfort may occur to inspectors, which will definitely damage their counting accuracy and worker health; due to stability, accuracy and cheap The advantages of machine v...

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): G06V10/25G06V10/26G06V10/28G06V10/44G06V10/56G06V10/774G06K9/62G06T7/00G06T7/66G06T7/13G06T7/11G06T3/40
CPCG06T7/0004G06T7/13G06T7/11G06T7/66G06T3/4038G06T2207/10024G06T2207/20081G06T2207/20104G06T2207/20221G06T2207/30242G06V10/28G06V10/267G06V10/25G06V10/56G06V10/44G06V2201/06G06F18/214
Inventor 张堃吴建国张培建姜朋朋李子杰
Owner NANTONG UNIVERSITY
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