CN103471514B - A kind of garlic stage division based on machine vision and simple linear regression analysis - Google Patents
A kind of garlic stage division based on machine vision and simple linear regression analysis Download PDFInfo
- Publication number
- CN103471514B CN103471514B CN201310413428.4A CN201310413428A CN103471514B CN 103471514 B CN103471514 B CN 103471514B CN 201310413428 A CN201310413428 A CN 201310413428A CN 103471514 B CN103471514 B CN 103471514B
- Authority
- CN
- China
- Prior art keywords
- garlic
- image
- linear regression
- regression analysis
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Sorting Of Articles (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The present invention relates to garlic classification technique field, particularly a kind of garlic stage division based on machine vision and simple linear regression analysis.Be somebody's turn to do the garlic stage division based on machine vision and simple linear regression analysis, it is characterized in that: image acquisition is carried out to garlic, then to the acute binary conversion treatment of gathered image, connected region in mark bianry image, the size in this image connectivity region is indicated according to image pixel, according to the area of the garlic image gathered and the data of garlic transverse diameter, utilize simple linear regression analysis to carry out mathematical modeling, finally according to you and linear equation classification is carried out to the garlic of pre-classification.The method of the present invention's image zooming-out, CCD has very high precision when measuring length, obtains image, do not damage garlic, use simple linear regression analysis Modling model simultaneously by scanning, and the method is simple, and speed is fast, saves grading time.
Description
(1) technical field
The present invention relates to garlic classification technique field, particularly a kind of garlic stage division based on machine vision and simple linear regression analysis.
(2) background technology
The export volume of current China garlic increases, but the growth of collecting is not as the growth of export volume, causes the main cause of this phenomenon to be that garlic postharvest treatment means fall behind, so that the mixed level such as garlic is mixed, and good time uneven.Quality mixes, and causes value low, cannot reach the standard of international export, thus also cannot make inquiries international market.Along with the increase of import and export volume, this loss is also increasing year by year.Traditional mechanical classification method easily produces extruding or scratches at mesh edge.
Intelligent Recognition based on machine vision replaces the visual identity of people to have its very large advantage and long-range development prospect, the most general with the application in precision agriculture in agricultural, at present for aspect researchs such as plant growth supervision, weed identification, agricultural robot, agricultural remote sensing analysis, farm products area, Quality Detection, animal behavior tracking.Utilize machine vision classification to replace manual grading skill, mechanical dimension's classification, weight grading to be the inexorable trend that robotization classification develops, be mainly reflected in following several respects:
(1) machine vision effectiveness of classification is high.The hierarchy system of computer vision adopts CCD camera as sensor, and the CCD demarcated has very high precision when measuring area, and oneself is widespread use industrially, can meet the requirement of garlic classification.
(2) non-contact detecting process is belonged to by Machine Vision Detection.During detection, sensor CCD obtains image by scanning, does not damage garlic.And traditional mechanical classification method easily produces extruding or scratches at mesh edge.
(3) classification process controls by software algorithm.Intelligentized Software for Design, classification can be carried out under knowledge based and rule, and low to supplemental equipment requirement, physical construction is simple.
Therefore, study machine vision technique, for garlic classification, there is potential using value and good development prospect.
(3) summary of the invention
The present invention in order to make up the deficiencies in the prior art, provide a kind ofly to detect accurately, work efficiency is high based on the garlic stage division of machine vision and simple linear regression analysis.
The present invention is achieved through the following technical solutions:
A kind of garlic stage division based on machine vision and simple linear regression analysis, it is characterized in that: image acquisition is carried out to garlic, then binary conversion treatment is carried out to gathered image, connected region in mark bianry image, the size in this image connectivity region is indicated according to image pixel, and according to sample image, square measure is amounted to into square centimeter; According to the area of the garlic image gathered and the data of garlic transverse diameter, utilize simple linear regression analysis to carry out mathematical modeling, finally according to the linear equation of matching, classification is carried out to the garlic of pre-classification.
Specifically comprise the steps:
(1) choose each a few head of Three Estate garlic respectively, utilize same camera to take pictures to garlic, require that shooting height and camera pixel want consistent;
(2) carry out binaryzation pre-service to the image gathered, clearing draw white garlic region area, using the area that the obtains feature as garlic hierarchy system;
(3) according to the area obtaining garlic image, with the transverse diameter of known garlic, simple linear regression analysis Modling model is utilized;
(4) image information is extracted to pre-classification garlic, calculate pre-classification garlic area, gained area is brought in a linear equation built up, obtains the transverse diameter of garlic, carry out classification according to transverse diameter data.
Its more excellent scheme is:
In step (2), the area in described garlic region represents with square centimeter.
In step (3), garlic transverse diameter millimeter represents.
In step (4), garlic classification is divided into Three Estate according to the size of garlic transverse diameter: one-level garlic: transverse diameter is more than or equal to 5cm, secondary garlic: transverse diameter is more than or equal to 4cm, three grades of garlics: transverse diameter is more than or equal to 3cm.
Garlic classification is divided into Three Estate according to the size of its transverse diameter, xsect due to garlic is approximately circular or oval, so the transverse diameter of garlic and its cross-sectional area have certain correlativity, the present invention takes garlic cross sectional image by sensor CCD, by gathered image binaryzation, the area in the white garlic region calculated by software, then can set up the relation between garlic area and garlic transverse diameter according to unary linear regression equation.Choose garlic three rank each rank garlics some, obtain its garlic cross-sectional area by image, be designated as:
, obtain its transverse diameter and be designated as,
.Simple linear regression analysis predicted method is according to independent variable
and dependent variable
correlationship, set up
with
equation of linear regression carry out the method predicted.
First whole measurement data brings into by the present invention respectively
in:
The present invention adopts least square fitting unary linear regression equation, the basic meaning of least square method in theory of errors is: in the repetitive measurement with equally accurate, ask the most reliable (the most believable) value time, be when each measured value residual sum of squares (RSS) for minimum be tried to achieve value.During least-squares algorithm linear fitting to measurement data, be that all measurement data points are all marked in coordinate diagram, with the straight line of least square fitting, the residual sum of squares (RSS) between its each data point and fitting a straight line is minimum.
To linear equation
, be minimum by residual sum of squares (RSS), can obtain according to all measurement data
Above formula is right respectively
with
get partial derivative to obtain:
Necessary condition is
Then have
After arrangement
(
Simultaneous solution obtains:
Will
with
substitute into
, i.e. the linear equation of handy least square fitting.
The method of the present invention's image zooming-out, sensor CCD has very high precision when measuring length, obtains image, do not damage garlic, use simple linear regression analysis Modling model simultaneously by scanning, and the method is simple, and speed is fast, saves grading time.
(4) accompanying drawing explanation
Below in conjunction with accompanying drawing, the present invention is further illustrated.
Fig. 1 is process chart of the present invention;
Fig. 2 is the relation linear graph of simple linear regression analysis matching, this curve transverse axis coordinate is the area of garlic, unit is square centimeter, ordinate of orthogonal axes is the maximum transverse diameter of garlic, unit is millimeter, and straight line is the straight line that the unary linear regression equation set up according to area and its maximum transverse diameter of garlic is drawn;
Fig. 3 is garlic cross-sectional view.
(5) embodiment
Concrete grammar of the present invention is as follows:
(1) choose each a few head of Three Estate garlic respectively, utilize same sensor CCD to take pictures to garlic, require that shooting height and camera pixel want consistent.And measure garlic transverse diameter size, be designated as
, obtain its garlic cross-sectional area by image, be designated as:
, data are substituted into, formula
(1)
(2)
Will
with
substitute into
, i.e. the linear equation of handy least square fitting.
(2) the same sensor of predicted picture is taken pictures, require that shooting height and camera pixel want consistent.By the image area obtained
substitute into linear equation
obtain garlic transverse diameter size, be designated as
.
(3) three grade scales 5cm, 4cm and 3cm of the garlic transverse diameter size obtained and garlic are compared, then confirm hierarchical categories.
Claims (2)
1. the garlic stage division based on machine vision and simple linear regression analysis, it is characterized in that: image acquisition is carried out to garlic, then binary conversion treatment is carried out to gathered image, connected region in mark bianry image, the size in this image connectivity region is indicated according to image pixel, and according to sample image, square measure is amounted to into square centimeter; According to the area of the garlic image gathered and the data of garlic transverse diameter, utilize simple linear regression analysis to carry out mathematical modeling, finally according to the linear equation of matching, classification is carried out to the garlic of pre-classification; Specifically comprise the steps: that (1) chooses each a few head of Three Estate garlic respectively, utilize same camera to take pictures to garlic, require that shooting height and camera pixel want consistent; (2) carry out binaryzation pre-service to the image gathered, clearing draw white garlic region area, using the area that the obtains feature as garlic hierarchy system; (3) according to the area obtaining garlic image, with the transverse diameter of known garlic, simple linear regression analysis Modling model is utilized; (4) image information is extracted to pre-classification garlic, calculate pre-classification garlic area, gained area is brought in a linear equation built up, obtains the transverse diameter of garlic, carry out classification according to transverse diameter data.
2. the garlic stage division based on machine vision and simple linear regression analysis according to claim 1, it is characterized in that: in step (3), garlic transverse diameter millimeter represents.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310413428.4A CN103471514B (en) | 2013-09-12 | 2013-09-12 | A kind of garlic stage division based on machine vision and simple linear regression analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310413428.4A CN103471514B (en) | 2013-09-12 | 2013-09-12 | A kind of garlic stage division based on machine vision and simple linear regression analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103471514A CN103471514A (en) | 2013-12-25 |
CN103471514B true CN103471514B (en) | 2016-02-03 |
Family
ID=49796468
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310413428.4A Expired - Fee Related CN103471514B (en) | 2013-09-12 | 2013-09-12 | A kind of garlic stage division based on machine vision and simple linear regression analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103471514B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103914849B (en) * | 2014-04-18 | 2018-07-06 | 扬州福尔喜果蔬汁机械有限公司 | A kind of detection method of red jujube image |
CN106503749B (en) * | 2016-11-04 | 2019-06-14 | 北京农业信息技术研究中心 | A kind of automatic grading method and its system of caviar |
CN110458106A (en) * | 2019-08-13 | 2019-11-15 | 深圳市睿海智电子科技有限公司 | A kind of intelligent analysis method and intellectual analysis device of tomato growth state |
CN110472557B (en) * | 2019-08-13 | 2023-06-02 | 深圳市睿海智电子科技有限公司 | Tomato growth monitoring method and device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101153851A (en) * | 2007-06-15 | 2008-04-02 | 中国农业大学 | Apple detection classification method based on machine vision |
CN102419147A (en) * | 2011-08-17 | 2012-04-18 | 徐州市仁和农业科技咨询服务有限公司 | Garlic metering device |
CN103027368A (en) * | 2012-12-06 | 2013-04-10 | 青岛农业大学 | Automatic garlic root cutting machine based on image processing technique |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4734650B2 (en) * | 2006-10-31 | 2011-07-27 | 国立大学法人 岡山大学 | Defect detection method and apparatus for cream solder printing |
JP5514131B2 (en) * | 2011-01-31 | 2014-06-04 | 日立Geニュークリア・エナジー株式会社 | Image processing method, image processing apparatus, and underwater inspection apparatus equipped with the same |
-
2013
- 2013-09-12 CN CN201310413428.4A patent/CN103471514B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101153851A (en) * | 2007-06-15 | 2008-04-02 | 中国农业大学 | Apple detection classification method based on machine vision |
CN102419147A (en) * | 2011-08-17 | 2012-04-18 | 徐州市仁和农业科技咨询服务有限公司 | Garlic metering device |
CN103027368A (en) * | 2012-12-06 | 2013-04-10 | 青岛农业大学 | Automatic garlic root cutting machine based on image processing technique |
Non-Patent Citations (2)
Title |
---|
基于图像处理与GUI的大蒜切脐的研究;赵丽清等;《青岛农业大学学报(自然科学版)》;20111231;第28卷(第4期);第305-308页 * |
基于机器视觉的苹果最大横切面直径分级方法;陈艳军等;《农业工程学报》;20120131;第28卷(第2期);第284-288页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103471514A (en) | 2013-12-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Luo et al. | Vision-based extraction of spatial information in grape clusters for harvesting robots | |
CN104501720B (en) | Non-contact object size and range image measuring instrument | |
CN112164062A (en) | Wasteland information extraction method and device based on remote sensing time sequence analysis | |
CN103471514B (en) | A kind of garlic stage division based on machine vision and simple linear regression analysis | |
CN103983190B (en) | Visual inspection classification device and method for appearance dimensions of group origin duck eggs | |
CN103278090A (en) | Visual measurement method for volume of irregular object | |
CN104501735A (en) | Method for observing three-dimensional deformation of side slope by utilizing circular marking points | |
CN104020475A (en) | Power line extracting and modeling method based on airborne LiDAR data | |
CN104880160B (en) | Two-dimensional-laser real-time detection method of workpiece surface profile | |
CN104268853A (en) | Infrared image and visible image registering method | |
Ma et al. | Automatic branch detection of jujube trees based on 3D reconstruction for dormant pruning using the deep learning-based method | |
CN105678768A (en) | Machine vision-based tire tread detection method | |
CN103512494A (en) | Visual inspection system and method for scale micro changes of plant fruits | |
CN103236043B (en) | A kind of plant organ point cloud restoration method | |
Liu et al. | A lightweight method for grape berry counting based on automated 3D bunch reconstruction from a single image | |
CN109764809A (en) | A method of fruit tree canopy volume is calculated based on two-dimensional laser sensor in real time | |
CN103077398A (en) | Livestock group number monitoring method based on embedded natural environment | |
CN113920138A (en) | Cow body size detection device based on RGB-D camera and detection method thereof | |
CN104036514A (en) | Circle detection method based on histogram peak value search | |
CN106202002A (en) | A kind of for detecting the method whether series of hydrological parameter makes a variation | |
Miao et al. | Measurement method of maize morphological parameters based on point cloud image conversion | |
CN105654479A (en) | Multispectral image registering method and multispectral image registering device | |
CN103533332A (en) | Image processing method for converting 2D video into 3D video | |
CN103697833B (en) | Agricultural product shape detecting method and device | |
CN102608349A (en) | Intelligent greenhouse nectarine fruit growth rate wireless monitoring system based on embedded dual-camera platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160203 Termination date: 20170912 |
|
CF01 | Termination of patent right due to non-payment of annual fee |