CN104764402B - The visible detection method of citrus volume - Google Patents
The visible detection method of citrus volume Download PDFInfo
- Publication number
- CN104764402B CN104764402B CN201510107888.3A CN201510107888A CN104764402B CN 104764402 B CN104764402 B CN 104764402B CN 201510107888 A CN201510107888 A CN 201510107888A CN 104764402 B CN104764402 B CN 104764402B
- Authority
- CN
- China
- Prior art keywords
- citrus
- volume
- image
- rgb image
- extracted
- 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
- Image Analysis (AREA)
Abstract
The invention discloses a kind of visible detection method of citrus volume, comprise the following steps：Step 1: citrus image is split：The citrus RGB image that will be collected, using red feature 2R G B, by the automatic selected threshold of OTSU methods, splits citrus RGB image using by the threshold value that OTSU methods are chosen automatically；Step 2: by opening operation to the citrus RGB image denoising after above-mentioned segmentation；Step 3: three citrus image slices vegetarian refreshments number, major axis and minor axis length shape facilities are extracted in the citrus RGB image that above-mentioned steps two open after operation；Step 4: three the pixel number extracted using in above-mentioned step 3, major axis and minor axis length shape facilities estimate citrus volume as input using BP networks.The advantages of to realize the volume progress automatic, high precision identification to citrus.
Description
Technical field
The present invention relates to crops to differentiate field, and in particular, to a kind of visible detection method of citrus volume.
Background technology
Citrus is the important industrial crops in China, with the continuous development of agricultural modernization technology, is drawn in Orange Producing
The efficiency for entering image processing techniques raising citrus estimation is increasingly becoming an important developing direction.In recent years, except utilizing machine
Device vision helps mechanical hand positioning to realize the automatic picking of citrus and be detected two development to citrus quality using spectrum picture
Outside direction, the hot spot of research is increasingly becoming by introducing image processing techniques progress citrus profile measurement.But the prior art
Detection for citrus shape lacks the direct detection to its volume.
In existing research it is more by image in citrus girth area, fruit footpath size, maximum transverse diameter and minimum diameter, just
Hand over the size of the not two-dimensional parameter such as bending moment characterization citrus.These parameters are all based on the two dimension ginseng of the sectional area extraction of citrus image
Number, these parameters are not linearly related with citrus volume, therefore directly estimate citrus volume accuracy still using these parameters
It can not meet the needs of production.
The content of the invention
It is an object of the present invention in view of the above-mentioned problems, a kind of visible detection method of citrus volume is proposed, with realization pair
The volume of citrus carries out the advantages of automatic, high precision identification.
To achieve the above object, the technical solution adopted by the present invention is：
A kind of visible detection method of citrus volume, comprises the following steps：
Step 1: citrus image is split：The citrus RGB image that will be collected, using red feature 2R-G-B, passes through OTSU
The automatic selected threshold of method, splits citrus RGB image using by the threshold value that OTSU methods are chosen automatically；
Step 2: by opening operation to the citrus RGB image denoising after above-mentioned segmentation；
Step 3: citrus image slices vegetarian refreshments number, major axis are extracted in the citrus RGB image that above-mentioned steps two open after operation
With three shape facilities of minor axis length；
Step 4: three the pixel number extracted using in above-mentioned step 3, major axis and minor axis length shape facilities as
Input, estimates citrus volume using BP networks.
Preferably, using BP networks estimation citrus volume in above-mentioned steps four, BP networks intermediate layer is according to input number of nodes
Select 7 nodes.
Technical scheme has the advantages that：
Technical scheme, for citrus volume and the non-linear relevant situation of citrus cross-sectional image feature, from
And introduce the algorithm reduction estimation error of this non-linear estimation of neutral net.So as to which the volume reached to citrus carries out automatic height
The purpose of precision identification.
Below by drawings and examples, technical scheme is described in further detail.
Brief description of the drawings
Fig. 1 is the flow chart of the visible detection method of the citrus volume described in the embodiment of the present invention；
Fig. 2 a are citrus artwork；
Fig. 2 b are the citrus image after segmentation；
Fig. 2 c are out the citrus image after operation.
Embodiment
The preferred embodiment of the present invention is illustrated below in conjunction with attached drawing, it will be appreciated that described herein preferred real
Apply example to be merely to illustrate and explain the present invention, be not intended to limit the present invention.
As shown in Figure 1, a kind of visible detection method of citrus volume, comprises the following steps：
Artwork as shown in Figure 2 a, and has been done gray processing processing by citrus RGB image.
Step 1: citrus image is split：The citrus RGB image that will be collected, using red feature 2R-G-B, passes through OTSU
The automatic selected threshold of method, splits citrus RGB image using by the threshold value that OTSU methods are chosen automatically；Citrus RGB image
I.e. each pixel represents the degree of pixel red, green, blue, citrus image such as Fig. 2 b institutes after segmentation by tri- values of R, G, B
Show.
Step 2: by opening operation to the citrus RGB image denoising after above-mentioned segmentation；The citrus obtained by image after segmentation
Target or relatively rough.Since there is the presence of carpopodium, defect and spot on citrus surface, the citrus image after segmentation can exist
Some cavities.Therefore use and open operation progress denoising.Open the citrus image after operation as shown in Figure 2 c.
Step 3: citrus image slices vegetarian refreshments number, major axis are extracted in the citrus RGB image that above-mentioned steps two open after operation
With three shape facilities of minor axis length；After handling citrus image, it will generally be extracted for estimation citrus volume
Parameters for shape characteristic.Rule of thumb, the bigger citrus of sectional area, length and width, its volume are bigger.So selection citrus figure
Citrus region corresponding pixel points number as in, grows comprising the elliptical long axis length of citrus picture point region equivalent and short axle
Three features are spent as characteristic parameter.
Step 4: three the pixel number extracted using in above-mentioned step 3, major axis and minor axis length shape facilities as
Input, estimates citrus volume using BP networks.
Using BP networks estimation citrus volume in step 4, BP networks intermediate layer selects 7 nodes according to input number of nodes.
To verify the validity of technical solution of the present invention, 20 citruses are measured using drainage, then using this technology side
Case carries out cubing.It is training sample BP network to randomly select wherein 17 citruses, then utilizes the BP nets after training
Network carries out volume detection to remaining 3 citruses.
For the precision of analysis estimation result, the related coefficient that vision-based detection volume measures volume with drainage is asked for, as a result
For 0.981793.This shows that estimating volume is closer to actual volume.Then variance inspection is carried out to estimation volume and actual volume
Test, significance probability is much larger than 5%, this shows that estimating volume and actual volume is not present significant difference.
Effect using BP network algorithms detection orange volume is fine, and training sample worst error is 3.63%；3 works
It is 4.36% for the citrus worst error of test sample.
In conclusion the invention has the characteristics that：
1）Introduce machine vision, neutral net carries out the detection of citrus volume.
2）It is proposed to use the citrus region corresponding pixel points number in citrus image, work as comprising citrus picture point region
Elliptical three parameters of long axis length and minor axis length are measured to be detected.
Finally it should be noted that：The foregoing is only a preferred embodiment of the present invention, is not intended to limit the invention,
Although the present invention is described in detail with reference to the foregoing embodiments, for those skilled in the art, it still may be used
To modify to the technical solution described in foregoing embodiments, or equivalent substitution is carried out to which part technical characteristic.
Within the spirit and principles of the invention, any modification, equivalent replacement, improvement and so on, should be included in the present invention's
Within protection domain.
Claims (1)
1. a kind of visible detection method of citrus volume, it is characterised in that comprise the following steps：
Step 1: citrus image is split：The citrus RGB image that will be collected, using red feature 2R-G-B, by OTSU methods from
Dynamic selected threshold, splits citrus RGB image using by the threshold value that OTSU methods are chosen automatically, citrus RGB image, that is, every
A pixel represents the degree of pixel red, green, blue by tri- values of R, G, B；
Step 2: by opening operation to the citrus RGB image denoising after above-mentioned segmentation；
Step 3: citrus image slices vegetarian refreshments number, major axis and short are extracted in the citrus RGB image that above-mentioned steps two open after operation
Three shape facilities of shaft length；After handling citrus image, some shape facilities ginseng is extracted for estimation citrus volume
Number, the bigger citrus of sectional area, length and width, its volume is bigger, therefore selects the citrus region in citrus image to correspond to
Pixel number, comprising the elliptical long axis length of citrus picture point region equivalent and three features of minor axis length as feature
Parameter；
Step 4: three the pixel number extracted using in above-mentioned step 3, major axis and minor axis length shape facilities as input,
Estimate citrus volume using BP networks；BP networks intermediate layer selects 7 nodes according to input number of nodes；
For citrus volume and the non-linear relevant situation of citrus cross-sectional image feature, so as to introduce neutral net, this is non-thread
Property estimation algorithm reduce estimation error；So as to achieve the purpose that the volume to citrus carries out automatic, high precision identification；
Introduce machine vision, neutral net carries out the detection of citrus volume.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510107888.3A CN104764402B (en) | 2015-03-11 | 2015-03-11 | The visible detection method of citrus volume |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510107888.3A CN104764402B (en) | 2015-03-11 | 2015-03-11 | The visible detection method of citrus volume |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104764402A CN104764402A (en) | 2015-07-08 |
| CN104764402B true CN104764402B (en) | 2018-04-27 |
Family
ID=53646382
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510107888.3A Expired - Fee Related CN104764402B (en) | 2015-03-11 | 2015-03-11 | The visible detection method of citrus volume |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104764402B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107358627B (en) * | 2017-07-12 | 2020-06-09 | 西北农林科技大学 | Fruit size detection method based on Kinect camera |
| CN108038879A (en) * | 2017-12-12 | 2018-05-15 | 众安信息技术服务有限公司 | A kind of volume of food method of estimation and its device |
| CN108262267B (en) * | 2017-12-29 | 2019-10-18 | 北京农业智能装备技术研究中心 | More fruit detection methods and device in a kind of sorted fruits |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0727760A3 (en) * | 1995-02-17 | 1997-01-29 | Ibm | Produce size recognition system |
| CN1275020C (en) * | 2005-03-28 | 2006-09-13 | 浙江大学 | Multi dimension energy detection method and apparatus for fruit shape |
| CN101726251A (en) * | 2009-11-13 | 2010-06-09 | 江苏大学 | Automatic fruit identification method of apple picking robot on basis of support vector machine |
| CN102654463A (en) * | 2012-04-13 | 2012-09-05 | 北京农业信息技术研究中心 | Watermelon quality NDT (non-destructive testing) method and device |
| CN202649135U (en) * | 2012-06-29 | 2013-01-02 | 常熟理工学院 | Image acquisition device for citrus classification and detection |
| JP6245463B2 (en) * | 2013-01-17 | 2017-12-13 | 株式会社横崎製作所 | Sorting conveyor device |
| CN103148781A (en) * | 2013-01-26 | 2013-06-12 | 广西工学院鹿山学院 | Grapefruit size estimating method based on binocular vision |
| CN103162627A (en) * | 2013-03-28 | 2013-06-19 | 广西工学院鹿山学院 | Method for estimating fruit size by citrus fruit peel mirror reflection |
-
2015
- 2015-03-11 CN CN201510107888.3A patent/CN104764402B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN104764402A (en) | 2015-07-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yiyang | The design of glass crack detection system based on image preprocessing technology | |
| CN109191459B (en) | Automatic identification and rating method for continuous casting billet macrostructure center segregation defect | |
| CN104463159B (en) | A kind of image processing method and device for positioning iris | |
| CN105158258A (en) | Computer vision-based bamboo strip surface defect detection method | |
| Lü et al. | Leaf area measurement based on image processing | |
| CN108181316B (en) | Bamboo strip defect detection method based on machine vision | |
| CN104156951B (en) | A kind of white blood cell detection method for BAL fluid smear | |
| Mustafa et al. | Image processing of an agriculture produce: Determination of size and ripeness of a banana | |
| CN108876860B (en) | Image calibration method for measuring bubble offset of tube level | |
| CN114170208A (en) | Paper product defect detection method based on artificial intelligence | |
| CN107742307A (en) | Based on the transmission line galloping feature extraction and parameters analysis method for improving frame difference method | |
| TWI624807B (en) | Iterative analysis of medical images | |
| CN110349153B (en) | Image method tobacco peduncle content detection method based on multi-feature fusion | |
| CN104764402B (en) | The visible detection method of citrus volume | |
| CN106127735B (en) | A kind of facilities vegetable edge clear class blade face scab dividing method and device | |
| CN105225243B (en) | One kind can antimierophonic method for detecting image edge | |
| CN111145205B (en) | Pig body temperature detection method based on infrared image under multiple pig scenes | |
| CN107316287A (en) | A kind of defect identification method in rectangle ferrite magnetic disk sheet face | |
| CN114332081B (en) | Textile surface abnormity determination method based on image processing | |
| CN104614386A (en) | Lens defects type identification method | |
| JP2008267943A (en) | Crack detecting method | |
| CN106600615B (en) | A kind of Edge-Detection Algorithm evaluation system and method | |
| CN107993243A (en) | A kind of wheat tillering number automatic testing method based on RGB image | |
| Nguyen et al. | A novel automatic concrete surface crack identification using isotropic undecimated wavelet transform | |
| Bermúdez et al. | Image analysis for automatic feature estimation of the Mangifera indica fruit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190809 Address after: 545000 Workshop No. 225, No. 19 Chuyang Road, Liudong New District, Liuzhou City, Guangxi Zhuang Autonomous Region Patentee after: Guangxi Grapefruit Network Technology Co.,Ltd. Address before: 545006 Liuzhou City East District Road, central city, No. 268, No. Patentee before: Guangxi University of Science and Technology |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180427 |