CN100495032C - Method and apparatus for detecting surface quality of egg - Google Patents
Method and apparatus for detecting surface quality of egg Download PDFInfo
- Publication number
- CN100495032C CN100495032C CNB2005101349025A CN200510134902A CN100495032C CN 100495032 C CN100495032 C CN 100495032C CN B2005101349025 A CNB2005101349025 A CN B2005101349025A CN 200510134902 A CN200510134902 A CN 200510134902A CN 100495032 C CN100495032 C CN 100495032C
- Authority
- CN
- China
- Prior art keywords
- egg
- eggs
- sample
- beasts
- birds
- 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
Images
Abstract
The invention relates to a method and apparatus for nondestructive inspecting the egg surface quality. It adopts three CCD cam to take the surface picture of the egg, uses the picture collecting card to transmit the picture to the computer, uses the elastic thumping rod to impact the surface of the egg shell, uses the acoustical meter and A/D transfer card to transmit the sound signal to the computer, the computer imitates the human brain sight and listening function and dose admixture and model identifying process to the data. It uses the computer software to quote the surface quality according to the picture information collected by computer sight system and the sound signal collected by the sound testing system and the established specialist quoting database.
Description
Technical field
The present invention relates to a kind of method and apparatus that detects at the birds, beasts and eggs surface quality, refer in particular to method and device based on computer vision and acoustics integration technology Non-Destructive Testing birds, beasts and eggs surface quality.
Background technology
Birds, beasts and eggs are rich in protein, fat, multivitamin and trace element, are animal foods important in people's daily life.Because eggshell is thin and easily broken, in case broken, bacterium can very fast intrusion and breeding, causes the corruption, rotten of birds, beasts and eggs, and contaminated environment also brings disaster to other birds, beasts and eggs, thus the breakage of birds, beasts and eggs to detect be one of important step in birds, beasts and eggs production, operation and the processing.Many birds, beasts and eggs that studies show that store and transport in purchase, and the main cause of losing in the process is owing to damaged egg, stain egg are entrained between the normal egg, causes due to the cross pollution.It is reported that the birds, beasts and eggs of the annual purchase of China are because the putrid and deteriorated loss that causes accounts for more than 10% of purchase volume.Along with the development of China's Foreign Trade, the standard of industrial egg and the outlet of bright egg is also more and more stricter, and not only the breakage to egg has requirement, but also require the birds, beasts and eggs cleaning surfaces, complete, size is consistent, rejects various unusual eggs (double-yolked egg, heavy shell egg, abnormal-shape egg etc.) sometimes as required.Therefore the surface quality of birds, beasts and eggs is detected and carry out classification and can reduce loss, can improve the storage crudy again, help operator's extra earning, help consumer's safety.At present, main dependence manually rejected the stain egg in the commercial production under light both at home and abroad, and by observing and rotating and bump birds, beasts and eggs mutually, damaged birds, beasts and eggs are discerned, rejected to the sound of listening eggshell to send.These class methods not only labour intensity are big, and production efficiency is low, and artificial breakage is big, and accuracy of detection is subject to the influence of artificial notice, muscle power, experience and working attitude and can not get basic assurance.
The foreign scholar has the human computer vision technique to carry out the research of birds, beasts and eggs breakage.Elster R.T, (Elster R.T, Goodrum J.W.Detection ofcracks in eggs using machine vision.Trans of ASAE, 1991,34 (1) such as Goodrum J.W; Goodrum J.W, Elster R.T, Machine vision for crack detection in rotating eggs.Trans of ASAE, 1992,35 (4): 1323-1328) egg is divided into two grades of certified products and substandard products, the differentiation accuracy is not high and speed is slower, and relevant patent is not appeared in the newspapers.Find that in their research computer vision technique detects damaged egg because the influence of stain etc. easily causes Lou adopts, and fine crack is difficult to detect, easily causes omission, error is bigger.
Utilizing the Acoustic detection technology is to develop a special kind of skill faster over nearly 30 years to the Quality Detection of agricultural product, the literary friend waits on September 23rd, 2002 earlier and has proposed a patented claim (Chinese patent application prospectus about " harmless stage equipment and the method for detecting automatically of quality of poultry eggs ", number of patent application: 0213983.5), utilize the acoustics hammering method to judge the breakage of birds, beasts and eggs, judge birds, beasts and eggs size and color of egg center by computer vision, but just detect with single method, and can not carry out classification, and do not relate to unusual egg (as heavy shell egg to the damaged situation of birds, beasts and eggs, abnormal-shape egg etc.) and the detection of stain egg.
At present, artificial sense assessment method or certain single detection method are adopted in the detection of birds, beasts and eggs surface quality both at home and abroad more, can not carry out more comprehensive objective evaluation to the quality of birds, beasts and eggs, there is limitation in detection, and precision is not enough.For example single computer vision technique can't detect micro-crack and unusual egg; Single acoustic technique can't be carried out classification to the damaged situation of birds, beasts and eggs, can't detect the stain on birds, beasts and eggs surface and abnormal-shape egg etc.
Summary of the invention
Technical matters
The objective of the invention is to overcome the defective of prior art, propose under a kind of online situation based on computer vision and acoustics integration technology method and apparatus at birds, beasts and eggs surface quality Non-Destructive Testing classification, particularly visual information and auditory information can be merged and carry out comprehensive distinguishing, can more accurately detect the damaged situation of birds, beasts and eggs, and stain and unusual egg etc., and carry out classification, improved quality of poultry eggs detects in the egg product processing industry quality and efficient.
Technical scheme
Purpose of the present invention realizes by the following method:
At first set up knowledge base, to the birds, beasts and eggs (egg that is detected, duck's egg, goose egg etc.), according to its examination criteria (as exporting bright egg national standard etc.), earlier the birds, beasts and eggs of some are carried out sensory evaluation and set up expert knowledge library by the professional, utilize the CCD camera to take the image of birds, beasts and eggs sample then, via image acquisition and import computing machine into, knocking the voice signal of eggshell generation then gathers by sound meter (microphone), collect intact egg and damaged egg (crack egg, decrease the shell egg, flow clear egg etc.) picture signal and voice signal, utilize Computer Analysis and characteristic information extraction, the brain of computer mould personification carries out fusion treatment to visual database and audible data storehouse then, set up the expert knowledge library expert decision data storehouse relevant with the birds, beasts and eggs surface quality, thereby by the quality of computer-made decision sample, quality qualities such as grade.
It is as follows to set up the step that standard model that the expert differentiates database detects:
1. during working sample, sample is placed on the hole of egg holder in the closed chamber, the egg holder divides levels with closed chamber, the lower floor source of giving out light, and light passes through the hole
Transmission sample;
2. the picture signal by CCD camera collection sample and import computing machine;
3. the stamp in the closed chamber impacts sample, produces voice signal and adopts in the computing machine by sound meter;
4. image is handled characteristic information extraction;
5. sound is handled characteristic information extraction;
6. egg holder Rotation Controllers makes pop-jump moving 90 degree of egg, and the drive sample revolves and turn 90 degrees;
7. the stamp in the closed chamber impacts sample, produces voice signal and adopts in the computing machine by sound meter;
8. sound is handled characteristic information extraction;
9. egg holder Rotation Controllers makes pop-jump moving 90 degree of egg, and the drive sample revolves and turn 90 degrees;
10. repeat an above-mentioned steps operation 2.～9.;
Described method based on computer vision and acoustic technique Non-Destructive Testing birds, beasts and eggs surface quality, it is characterized in that image information and voice signal information that computing machine extracts are merged and mode treatment, mathematics manipulation wherein adopts regretional analysis, Fourier transform, conventional analysis of statistical data method such as fuzzy mathematics and high precision real-time mode systems such as neural network, genetic algorithm come deal with data, and connect with knowledge base and to train, to learn, obtain expert's decision data storehouse, make the system of development can improve the precision of classification.
Described fusion is divided into the fusion of raw data fusion, characteristic fusion, decision-making data fusion many levels.Raw data fusion and characteristic fusion treatment mainly comprise processing and the fusion to picture signal and voice signal, and wherein Image Information Processing comprises background segment, the figure image intensifying, Threshold Segmentation, feature extraction etc., and sound signal processing comprises signals collecting, smoothing denoising, feature extraction etc.The characteristic fusion refers to extract eigenwert from each width of cloth image and each time voice signal, adopts discriminant analysis, neural network, regretional analysis, genetic algorithm etc. to merge on the eigenwert basis that all extract separately then.The fusion of decision data level then is to judge the characteristic fusion results that draws according to computer vision and acoustics, by using neural network, means such as fuzzy mathematics can be carried out the accurately comprehensive expert who judges to tested birds, beasts and eggs surface as crackle, profile, size even inner case and be differentiated database.
The device of described a kind of birds, beasts and eggs surface quality detection method is characterized in that being made up of four parts: gearing, Computer Vision Detection and analysis module, Acoustic detection and analysis module, pattern-recognition and Data Fusion system.Wherein gearing is made up of travelling belt, arrangements for speed regulation.The device of Computer Vision Detection and analysis module is by closed chamber, CCD camera, image pick-up card, light source, the egg holder, egg holder Rotation Controllers is formed, wherein, CCD camera, light source, the egg holder is in closed chamber inside, and image pick-up card is fixed on computer-internal, and the device of Acoustic detection and analysis module is by sound meter, the egg holder, egg holder Rotation Controllers, stamp knocks detecting device and closed chamber and forms.Sound meter wherein, the egg holder, stamp is in closed chamber inside.The data transmission that computer vision system and acoustics collection and disposal system collect is carried out analyzing and processing to identification and fused data disposal system in the computing machine.
Image detection closed chamber in the described Computer Vision Detection device is divided into two-layer up and down: lower floor is the Lights section, the upper strata is that birds, beasts and eggs detect the darkroom, the egg holder of two-layer centre contains roller bearing, be empty in the middle of the egg holder, drive the rotation of birds, beasts and eggs by the rotation of roller bearing, there are three CCD cameras on the surface of birds, beasts and eggs, and the CCD camera links to each other with image pick-up card in being inserted in computing machine by cable, and forms real-time computer vision collecting and disposal system with computing machine.
The image acquisition of computer vision and sound signal collecting process are all finished in closed chamber, and closed chamber lower floor light source adopts incandescent lamp, and light source is transmitted on the birds, beasts and eggs by the hollow of egg holder.For every piece of birds, beasts and eggs, roller bearing rotation four times, each 90 degree, stamp impacts birds, beasts and eggs four times, and sound meter is gathered four voice signals, and CCD camera collection secondary image.
During work, arrangements for speed regulation make travelling belt that tested birds, beasts and eggs are sent into given pace and detect closed chamber inside, and operation suspension, by the image of CCD camera shooting egg, import image pick-up card into through cable and handle, and handle by the software analysis of computer-internal.After image acquisition was finished for the first time, the Acoustic detection device was started working, and knocked controller and drove the position, equator that stamp impacts egg, and signal imports the A/D transition card by the sound meter collection into by cable, handled by the software analysis of computer-internal.Egg holder Rotation Controllers control egg holder rotation then drives egg and rotates 90 degree, as preceding carrying out the sound signal collecting process second time; After sound signal collecting finished for the second time, egg holder Rotation Controllers control egg holder rotation continued to make egg to rotate 90 degree, and at this moment camera system is carried out the image acquisition second time, and sound system carries out sound signal collecting process for the third time; Then egg holder Rotation Controllers control egg holder rotation continues to make egg to rotate 90 degree, carries out the collection of the 4th voice signal, then the signals collecting end-of-job.So far, to every piece of egg, the CCD camera is gathered the egg image altogether 2 times, the voice signal after 4 eggs of sound meter collection are hit.The data that at every turn obtain are carried out fusion treatment by the software of computer-internal, the last judgement classification results of exporting the birds, beasts and eggs surface quality on screen.
Beneficial effect of the present invention
1, the method that a kind of birds, beasts and eggs surface quality detects, vision and auditory system based on the birds, beasts and eggs surface quality Non-Destructive Testing anthropomorphic dummy of computer vision and acoustics integration technology, what obtain is not the visual information of birds, beasts and eggs sample and the simple superposition of auditory information, but apish information fusion ability, vision and auditory information are merged, method for classifying modes during with high-precision real such as neural network, deal with data such as genetic algorithm, has artificial intelligence, can be used for the quality judging and the classification of agricultural product, the judge personnel can be assisted or replace to detection method of mentioning and device.
2 of the present invention with single computer visions are compared with the acoustics detection technique, the information that obtains more comprehensively, its reliability, repeatability and adaptability are higher, comparing the accuracy of not only judging the birds, beasts and eggs breakage with conventional method can reach more than 95%, and can carry out classification to the damaged situation of birds, beasts and eggs, can attach simultaneously and detect birds, beasts and eggs size, profile, stain, inner blood cake etc., more comprehensive to the birds, beasts and eggs quality assessment, be more suitable for the needs of producing in modern industry.
3, computer vision technique is merged in the present invention and acoustic technique can be carried out comparatively fast and comprehensively Non-Destructive Testing to agricultural product such as birds, beasts and eggs, both can liberate the labour, get rid of people's interference caused by subjective factors, and can judge and classification quality of poultry eggs in real time quickly and accurately again.
Description of drawings
Fig. 1: technical scheme synoptic diagram of the present invention;
Fig. 2: computer vision and acoustic technique merge synoptic diagram;
Fig. 3: application example of the present invention (at egg) technology path synoptic diagram;
Fig. 4: application example device synoptic diagram of the present invention;
Fig. 5: each area identification synoptic diagram of application example of the present invention;
Fig. 6: application example Computer Vision Detection partial devices synoptic diagram of the present invention;
Fig. 7: application example of the present invention rotates synoptic diagram;
Fig. 8: application example acoustic technique of the present invention test section device synoptic diagram;
Each part description in the accompanying drawing is as follows
Among the figure: 1, travelling belt; 2, closed chamber; 3, the CCD camera; 4, sound meter; 5, the A/D transition card; 6, image pick-up card; 7, computing machine; 9, arrangements for speed regulation; 10, the egg holder; 11, egg holder Rotation Controllers; 12, stamp; 13, knock controller; 14, birds, beasts and eggs
Embodiment
The present invention has versatility to the lossless detection method and the device of birds, beasts and eggs, with the egg is that example describes, other Poultry and Eggs products can be with reference to the method for this example, specifically at the evaluation criterion of the sample of being surveyed, set up a new knowledge base and expert and differentiate the storehouse, just can test such birds, beasts and eggs.
Embodiment 1:
Consult Fig. 3, be the system schema synoptic diagram that detects at egg eggshell crackle, related request according to bright egg export inspection national standard, detection means is carried out grade estimation and classification routinely earlier, be sample with these eggs then, use the cannot-harm-detection device to detect, set up the expert decision data storehouse relevant with knowledge base based on computer vision and acoustics integration technology.
Computer Vision Detection among the figure and processing comprise the image acquisition of egg, the image pre-service, graphical analysis and feature extraction, can judge whether crackle exists reaches the crackle size cases, wherein image acquisition is taken the egg sample in the closed chamber by the CCD camera, import computing machine into through image pick-up card, wherein sealed inside has light source, and image acquisition can be fixed on computer-internal.By the pulpit egg holder rotation of computing machine, and then drive the egg rotation, all surfaces image of egg is gathered egg holder Rotation Controllers.
Acoustic technique detection system among the figure comprises the collection of signal, the pre-service of signal, feature extraction, comparative analysis, result's output.Signal impacts the egg surface by stamp and produces, and is gathered by sound meter, imports the A/D transition card into by cable and handles, and imports computing machine then into and further analyzes.One piece of egg is gathered voice signal altogether 4 times, import computing machine into and carry out analyzing and processing.
Judge the damaged situation of egg between the Computer Vision Detection analysis and processing module among the figure and acoustic technique check and analysis and processing module according to the information of gathering separately, adopt according to the software systems in the computing machine then integration technology such as neural network and with knowledge base contrast, synthetic determination classification egg surface quality.
Information fusion among the figure at first is that the raw data in computer vision or the Acoustic detection merges, be the fusion of each characteristic information of map-area in acquired information then, be comprehensive the two result data at last, utilize neural network, fuzzy mathematics etc. to carry out Decision Fusion, obtain expert's decision data storehouse, the quality status of accurate description egg.
The hardware unit synoptic diagram of the embodiment of the invention as shown in Figure 4, Computer Vision Detection System is by closed chamber (2), CCD camera (3), image pick-up card (6), light source (8), egg holder compositions such as (10), CCD camera (3) wherein, light source (8), egg holder (10) are fixed on closed chamber (2) inside, and image pick-up card (6) is fixed on computing machine (7) inside.Wherein the situation of this partial devices in closed chamber (2) as shown in Figure 6.The acoustic technique pick-up unit comprises sound meter (4) as shown in Figure 8, A/D transition card (5), and egg holder Rotation Controllers (11), stamp (12) knocks controller (13).Sound meter (4) wherein, egg holder (10), stamp (12) knocks controller (13) and is fixed on closed chamber (2) inside, and A/D transition card (5) is fixed on the PCI slot of computing machine (7) mainboard.Wherein, equator on the egg surface, blunt end and most advanced and sophisticated position are as shown in Figure 6.Wherein, as shown in Figure 7, egg holder Rotation Controllers (11) control egg holder (10) rotation make CCD camera (3) can gather the full detail on egg (14) surface, and sound meter (4) can be gathered the voice signal of egg different parts.
During work, arrangements for speed regulation (9) make travelling belt (1) import egg into closed chamber (2) inside with given pace, and operation suspension, are taken the image of egg by CCD camera (3), import image pick-up card (6) into through cable and handle, handle by the software analysis that computing machine (7) is inner.After image acquisition is finished for the first time, the Acoustic detection device is started working, and knocks controller (13) and drives the position, equator that stamp (12) impacts egg (14), and signal is gathered by sound meter (4), import A/D transition card (5) into by cable, handle by the software analysis that computing machine (7) is inner.Egg holder Rotation Controllers (11) control egg holder (10) rotation then drives egg (14) and rotates 90 degree, carries out the sound signal collecting process second time; After for the second time sound signal collecting finished, egg holder Rotation Controllers (11) control egg holder (10) rotation continued to make egg (14) to rotate 90 degree, carried out image acquisition and the sound signal collecting process for the third time second time; Egg holder Rotation Controllers (11) control egg holder (10) rotation then drives egg (14) and rotates 90 degree, carries out the voice signal gatherer process the 4th time.To every piece of egg (14), each CCD camera (3) is gathered 2 egg images, obtain 6 width of cloth images altogether, and sound meter (4) is gathered the voice signal after 4 eggs (14) are hit, the software of last computer-internal carries out fusion treatment to all data, exports last egg surface quality and judges classification results.
Embodiment 2:
1 test material
The non-polluted egg that test material provides for source, Jiangsu Province wound fowl industry Development Co., Ltd, the chicken kind is the Luo Man laying hen.The new fresh hen egg in 1-2 days postpartum, being divided into is two batches, the egg of all test usefulness is judged as crackle egg or intact egg all by the human expert scrutiny.First egg is for being used for setting up knowledge base and expert's decision data storehouse.Have 290 pieces, 150 pieces in wherein intact egg, 140 pieces in shell shake egg, wherein a part of shell shake egg are that intact egg is artificially made crackle and got, crackle size and distributing more at random.Second batch of egg uses for the precision that the check expert differentiates database, has 237 pieces, and the result of artificial scrutiny is 126 pieces in an intact egg, 111 pieces in shell shake egg.System level result's accuracy rate draws by comparing with artificial cognition.
2 test units
2.1 computer vision
Camera: Japanese JVCTK-C1381 type CCD camera, resolution is 752 * 582, is output as simulating signal;
Image pick-up card: Canadian Matrox Π image pick-up card;
Camera bellows: 80 * 80 * 100cm
3The big wooden case of self-control, inwall is opaque black material;
Light source: 25W incandescent lamp;
Computer system (PC): P41.7G processor, 256MB internal memory, NVIDIA GeForce MX440 64MB video card;
The image processing software system: image processing software is the image processing software Image processing 1.0 of processing of farm products teaching and research room of food science and technology institute of Agricultural University Of Nanjing independent development.
2.2 Acoustic detection
Stamp: knocker is a carefully rod of steel elasticity, and an end is spherical, is about 60mm;
SD150 dynamic test and Signal Analysis System: ring Electronic Instrument, Limited unites and produces in dynamo-electric Science ﹠ Technology Center of University Of Tianjin and the Tianjin
Digital sound level meter: HS5633A
Computer system (PC): P41.7G processor, 256MB internal memory, NVIDIA GeForce MX440 64MB video card;
3 test methods and procedures
3.1 the phase one
To remove the stain on surface for samples tested egg cleaning earlier, be crackle egg and intact egg by human expert classification egg then, is numbered respectively.Wherein, carry out the test of intact egg earlier, collect the computer vision information and the acoustics signal message of intact egg, artificially make crackle then, and utilize computer vision and acoustic method to detect the surface quality of crackle egg respectively.
3.2 Computer Vision Detection method and step
In the darkroom, egg is crossed in transmittance, utilizes the image of camera collection egg pros and cons, gathers 6 width of cloth images altogether 2 times.Image processing step is: 1. background removal; 2. image transitions (comprising form and size); 3. gray scale transforms; 4. medium filtering; 5. Threshold Segmentation; 6. zone marker; 7. feature extraction.
3.3 acoustic detection method and step
Stamp knocks the position, equator of tested egg, and all eggs (intact egg and crackle egg) are all knocked 4 times, knocks once every 90 degree approximately along clockwise direction.Knock in the spectrogram of back record PC output the frequency values of amplitude maximum, i.e. characteristic response frequency except that noise at every turn.The position of sound meter is on the opposite of beating point, and egg is disposed across on the pad.The acquisition process step of voice signal is: 1. impact egg; 2. signals collecting; 3. signal analysis; 4. characteristic frequency is extracted; 5. data processing; 6. result's output.
4 knowledge bases and expert differentiate the foundation and the checking of database
4.1 Computer Vision Detection
Analyze the image of intact egg and crackle egg respectively, obtain the characteristic information of image.Extracted the characteristic information of the slit region of crackle egg image by processing, respectively with regional circularity, area, regional major diameter, minor axis and the line of apsides recently judge crackle.It is as follows to utilize statistical method to stipulate that the expert differentiates being judged to be of crackle in the database:
Definition: circularity (R:Ratio); Area (A:Area); Major diameter (L:LongPath); The ratio (LS:LongPath/ShortPath) of minor axis (S:ShortPath) line of apsides.
The feature of crackle satisfies:
①A＝-315.693+726.579R+4.2775L+9.0556S+9.0288LS
②L＝43.7841-140.39R+0.1443A+1.1673LS
3. if A 〉=100 pixels, then R≤0.2 or LS 〉=3
4. if A＜100 pixels, then R≤0.35 or LS 〉=3
Wherein: 0.0＜R≤0.398,1.5≤LS≤33.7
19≤A≤1871,8≤L≤270,1≤S≤76 (unit: pixel)
4.2 Acoustic detection
Analyze the acoustic signal information of intact egg and crackle egg, find one piece of intact flawless egg, knock the different parts in equator, the spectrogram that obtains is quite similar, and the characteristic response frequency values is approaching, and the CV value is little; The egg that crackle is arranged, difference is bigger each other for four characteristic frequency numerical value, and the CV value is relatively large.Wherein, CV refers to the coefficient of variation, promptly the result that the characteristic response frequency numerical value that obtains for four times carries out analysis of variance is knocked at the position, equator.In the process that detects classification, the detection of fine crack is difficulty relatively, becomes big and fine crack is easy in the storage transportation, and in order to guarantee detecting of crackle egg as far as possible, the threshold value of classification is low for well, so threshold value (CV) is made as 1.Determine the damaged hierarchical algorithms that detects of egg thus, all eggs of definition U={ }, the intact egg of A={ }, B={ shell shake egg }, function is expressed as follows:
4.3 differentiate the foundation of database
Setting up the principle of differentiating database is: each regional circularity after the extraction egg Flame Image Process, the information of the ratio of area, regional major diameter, minor axis and the line of apsides and acoustics signal characteristic information (CV), import these variablees, utilize fuzzy mathematics and neural network to judge and train.
Differentiate the judgement and the classification principle of database:
Judge: to the detected egg of input, obtain the image on surface, carry out Flame Image Process, extract and handle each regional circularity in the image of back, the numerical value of the ratio of area, regional major diameter, minor axis and the line of apsides judges that this egg is crackle egg or intact egg; Analyze the voice signal of this egg simultaneously,, determine that this piece egg is crackle egg or intact egg according to the algorithm of judging.Relatively two kinds of results that method is judged then carry out classification if the result is the same, if the result is different, judge that then this piece egg is a crack egg.
Classification: if two kinds of methods judge that all this egg is a crack egg, judge that then this piece egg is three grades,, judge that then this piece egg is an one-level if all be judged as intact egg; If acoustic method judges that this piece egg is a crack egg, and computer vision methods is judged as intact egg, think that then this piece egg has micro-crack, be decided to be secondary, if acoustic method judges that this piece egg is intact egg, and computer vision methods is judged as the crackle egg, then thinks the crackle egg, is decided to be three grades.
4.4 the expert differentiates the precision test of database
With second batch of egg is detected object, enters verifying attachment respectively, carries out computer vision and Acoustic detection, and provides the judgement classification results by system.By differentiating the precision of database with human expert result of determination contrast evaluation experts, the result is as shown in table 1 below, and this method and apparatus reaches more than 95% detecting of crack egg as can be seen, and the differentiation accuracy rate of egg has by the gross also been reached 93%.
Table 1 apparatus and method of the present invention detect the precision of egg cracks
Claims (1)
1, the method that a kind of birds, beasts and eggs surface quality detects, it is characterized in that: at first set up knowledge base, the birds, beasts and eggs of elder generation to being detected, related request according to fresh hen egg export inspection national standard, detection means is carried out grade estimation and classification routinely earlier, utilize the CCD camera to take the image of birds, beasts and eggs sample then, via image acquisition and import computing machine into, knocking the voice signal of eggshell generation then gathers by sound meter, collect the picture signal and the voice signal of egg, utilize Computer Analysis and characteristic information extraction, thereby by the quality of computer-made decision sample, grade quality quality:
1) during working sample, sample is placed on the hole of egg holder in the closed chamber, the egg holder divides levels with closed chamber, the lower floor source of giving out light, and light is by the hole transmission sample;
2) picture signal by CCD camera collection sample and import computing machine;
3) stamp in the closed chamber impacts sample, produces voice signal and adopts in the computing machine by sound meter;
4) image is handled characteristic information extraction;
5) sound is handled characteristic information extraction;
6) egg holder Rotation Controllers makes pop-jump moving 90 degree of egg, and the drive sample revolves and turn 90 degrees;
7) stamp in the closed chamber impacts sample, produces voice signal and adopts in the computing machine by sound meter;
8) sound is handled characteristic information extraction;
9) egg holder Rotation Controllers makes pop-jump moving 90 degree of egg, and the drive sample revolves and turn 90 degrees;
10) operation above-mentioned steps 2 of repetition)～9);
Aforementioned calculation machine vision-based detection comprises:
With regional circularity, area, regional major diameter, minor axis and the line of apsides recently judge crackle egg and intact egg,
Definition: circularity R; Area A; Major diameter L; Minor axis S; The ratio LS of the line of apsides;
The feature of crackle egg satisfies:
①A＝-315.693+726.579R+4.2775L+9.0556S+9.0288LS
②L＝43.7841-140.39R+0.1443A+1.1673LS
3. if A 〉=100 pixels, then R≤0.2 or LS 〉=3
4. if A＜100 pixels, then R≤0.35 or LS 〉=3
Wherein: 0.0＜R≤0.398,1.5≤LS≤33.7
19≤A≤1871,8≤L≤270,1≤S≤76; Unit: pixel
Above-mentioned Acoustic detection comprises:
CV refers to the coefficient of variation, promptly the result that the characteristic response frequency numerical value that obtains for four times carries out analysis of variance is knocked at the position, equator, determines the damaged hierarchical algorithms that detects of egg thus, all eggs of definition U={ }, the intact egg of A={ }, B={ shell shake egg }, function is expressed as follows:
11) differentiate result and grade in conjunction with the quality of this sample in the expert knowledge library:
Judge: to the detected egg of input, extract and handle each regional circularity in the image of back, the numerical value of the ratio of area, regional major diameter, minor axis and the line of apsides judges that this egg is crackle egg or intact egg; Analyze the voice signal of this egg simultaneously, determine that this piece egg is crackle egg or intact egg; Relatively two kinds of results that method is judged then carry out classification if the result is the same, if the result is different, judge that then this piece egg is a crack egg;
Classification: if two kinds of methods judge that all this egg is a crack egg, judge that then this piece egg is three grades,, judge that then this piece egg is an one-level if all be judged as intact egg; If acoustic method judges that this piece egg is a crack egg, and computer vision methods is judged as intact egg, think that then this piece egg has micro-crack, be decided to be secondary, if acoustic method judges that this piece egg is intact egg, and computer vision methods is judged as the crackle egg, then thinks the crackle egg, is decided to be three grades.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101349025A CN100495032C (en) | 2005-12-30 | 2005-12-30 | Method and apparatus for detecting surface quality of egg |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101349025A CN100495032C (en) | 2005-12-30 | 2005-12-30 | Method and apparatus for detecting surface quality of egg |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1804620A CN1804620A (en) | 2006-07-19 |
| CN100495032C true CN100495032C (en) | 2009-06-03 |
Family
ID=36866710
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101349025A Expired - Fee Related CN100495032C (en) | 2005-12-30 | 2005-12-30 | Method and apparatus for detecting surface quality of egg |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100495032C (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI644616B (en) * | 2017-03-17 | 2018-12-21 | Method for judging the hatching shape of poultry eggs by using image processing | |
| US10956809B1 (en) | 2019-11-21 | 2021-03-23 | Wang Lian | Artificial intelligence brain |
| RU2759733C1 (en) * | 2020-12-16 | 2021-11-17 | Общество с ограниченной ответственностью Инженерно-технический центр "УКАВТ" | Method for detecting substandard products in egg tray |
Families Citing this family (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101122591B (en) * | 2007-08-16 | 2010-08-18 | 中国农业大学 | Method for detecting eggshell composition deposition rule |
| CN101234381B (en) * | 2008-03-07 | 2011-09-07 | 天津市华核科技有限公司 | Granular material sorting classifying method based on visual sense recognition |
| CN101413928B (en) * | 2008-11-14 | 2011-11-16 | 江苏大学 | Fowl egg crack rapid on-line nondestructive detection device and method based on acoustic characteristic |
| CN101750419B (en) * | 2009-12-21 | 2012-10-10 | 核工业理化工程研究院华核新技术开发公司 | Open-type visual detection method based on featuring matching |
| CN101915823B (en) * | 2010-07-05 | 2014-02-26 | 华中农业大学 | Ultraviolet detection method for shell bacteria-carrying eggs and device thereof |
| CN101915767B (en) * | 2010-07-13 | 2011-10-19 | 南京农业大学 | Method for systematically recognizing egg cracks |
| CN101916366B (en) * | 2010-08-09 | 2013-05-01 | 黑龙江八一农垦大学 | Brooding-before clear egg distinguishing method and device thereof |
| CN101975357B (en) * | 2010-09-25 | 2012-07-04 | 东华大学 | Self-adaptive light source lighting system and method for high-precision image measuring device |
| CN102144575A (en) * | 2011-05-04 | 2011-08-10 | 浙江省农业科学院 | Method for comprehensively evaluating quality of goose eggshell |
| CN102866236B (en) * | 2012-08-31 | 2015-06-24 | 广东省农业科学院农业生物技术研究所 | Texture evaluation method for rapidly discriminating quality of rice noodles |
| ES2673558T3 (en) * | 2012-09-04 | 2018-06-22 | Katholieke Universiteit Leuven | Method and apparatus for examining eggs |
| CN103529128B (en) * | 2013-09-30 | 2016-09-21 | 天津工程机械研究院 | A kind of fatigue crack on-line detecting system and detection method |
| CN103592367B (en) * | 2013-10-23 | 2016-03-02 | 江苏大学 | A kind of Portable poultry egg quality detection method |
| FR3023617B1 (en) * | 2014-07-11 | 2018-06-15 | Tronico | NON - INVASIVE DEVICE FOR DETERMINING THE FERTILITY AND / OR SEX OF AN EGG AND CORRESPONDING METHOD. |
| CN105159248B (en) * | 2015-08-05 | 2019-01-29 | 东莞理工学院 | A method of classifying to industrial products based on machine vision |
| CN105432182B (en) * | 2015-08-14 | 2019-05-24 | 南京农业大学 | Split clever rice paddy seed acoustic detecting device |
| CN105494168B (en) * | 2016-01-26 | 2018-03-27 | 蔡海睿 | A kind of device and method observed and record chicken hatching experiment |
| CN107127162B (en) * | 2016-02-27 | 2020-01-14 | 二商健力广平生物科技有限公司 | Automatic detection and separation device for damaged eggs |
| CN109191461A (en) * | 2018-10-22 | 2019-01-11 | 广东工业大学 | A kind of Countryside Egg recognition methods and identification device based on machine vision technique |
| CN109859197A (en) * | 2019-02-01 | 2019-06-07 | 佛山市南海区广工大数控装备协同创新研究院 | One kind being based on characteristic matching roasted eggs detection method |
| CN110161121B (en) * | 2019-06-17 | 2020-01-17 | 江苏仁康蛋业有限公司 | Online nondestructive test system for poultry eggs |
| CN110150182B (en) * | 2019-06-17 | 2021-06-08 | 江苏仁康蛋业有限公司 | Egg shell detection system |
| CN110376142B (en) * | 2019-08-20 | 2022-07-01 | 陕西中医药大学 | Detection method for quality grade of traditional Chinese medicine safflower |
| CN112990623A (en) * | 2019-12-12 | 2021-06-18 | 阿里巴巴集团控股有限公司 | Honey, target object quality analysis method, equipment and storage medium |
| CN112162575B (en) * | 2020-09-13 | 2021-10-22 | 江苏深农智能科技有限公司 | Intelligent shed environment control method and system based on data model |
| CN113793328A (en) * | 2021-09-23 | 2021-12-14 | 中国农业大学 | Light-weight egg shape recognition method based on SE-ResNet structure |
| CN114080999B (en) * | 2021-11-06 | 2023-02-28 | 天津市广源畜禽养殖有限公司 | Intelligent screening partial shipment system of egg |
| CN114563266A (en) * | 2022-01-27 | 2022-05-31 | 西安理工大学 | In-situ dynamic camera shooting equipment for crack propagation process in mechanical test |
| CN115436488B (en) * | 2022-08-31 | 2023-12-15 | 南京智慧基础设施技术研究院有限公司 | Self-guiding self-adapting mobile detection system and method based on vision and voiceprint fusion |
| CN116148357B (en) * | 2023-04-21 | 2023-07-18 | 华裕农业科技有限公司 | Egg crack automatic detection and classification system and method capable of accurately sorting eggs |
| CN117256514B (en) * | 2023-11-17 | 2024-04-23 | 河北玖兴农牧发展有限公司 | Method for removing unqualified hatching eggs for sorting system |
-
2005
- 2005-12-30 CN CNB2005101349025A patent/CN100495032C/en not_active Expired - Fee Related
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI644616B (en) * | 2017-03-17 | 2018-12-21 | Method for judging the hatching shape of poultry eggs by using image processing | |
| US10956809B1 (en) | 2019-11-21 | 2021-03-23 | Wang Lian | Artificial intelligence brain |
| WO2021101439A1 (en) * | 2019-11-21 | 2021-05-27 | Lian Wang | Artificial intelligence brain |
| RU2759733C1 (en) * | 2020-12-16 | 2021-11-17 | Общество с ограниченной ответственностью Инженерно-технический центр "УКАВТ" | Method for detecting substandard products in egg tray |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1804620A (en) | 2006-07-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100495032C (en) | Method and apparatus for detecting surface quality of egg | |
| CN101144780B (en) | Pork freshness intelligent detection device | |
| CN101852768B (en) | Workpiece flaw identification method based on compound characteristics in magnaflux environment | |
| Patel et al. | Color computer vision and artificial neural networks for the detection of defects in poultry eggs | |
| CN1245630C (en) | Agricultural and animal product nondestrctive detection method based on electronic visual sense and smell sense fusion technology and its device | |
| CN202092599U (en) | Device for testing watermelon quality in online nondestructive manner | |
| CN111709333B (en) | Tracing early warning system based on abnormal excrement of cage-raised chickens and health monitoring method | |
| CN107123114A (en) | A kind of cloth defect inspection method and device based on machine learning | |
| Pan et al. | Eggshell crack detection based on computer vision and acoustic response by means of back-propagation artificial neural network | |
| CN109543679A (en) | A kind of dead fish recognition methods and early warning system based on depth convolutional neural networks | |
| CN101923059A (en) | Automatic cloth inspection method and system | |
| CN104764744A (en) | Visual inspection device and method for inspecting freshness of poultry eggs | |
| CN109307675A (en) | A kind of product appearance detection method and system | |
| CN106248680A (en) | A kind of engine commutator quality detecting system based on machine vision and detection method | |
| CN108318494B (en) | The red online vision detection and classification devices and methods therefor for proposing fruit powder | |
| JP2019149173A (en) | Network of intelligent machines | |
| CN107045013A (en) | Beef taste quality automatic classification detector and its detection method | |
| CN106353336A (en) | Lens coating automatic detection system | |
| CN112691939A (en) | Intelligent fruit and vegetable quality perception sorting robot system | |
| CN112560896A (en) | Fruit quality screening and classifying system based on image processing | |
| CN109937912B (en) | Egg classification system and method based on machine vision | |
| CN109342378B (en) | Colony growth state detection device and method based on multi-modal imaging technology | |
| CN204613160U (en) | A kind of bird egg crack detection device | |
| CN105911000A (en) | Characteristic wave band based blood spot egg on-line detection method | |
| CN104833724A (en) | Poultry eggshell crack detection method and device |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090603 Termination date: 20161230 |