CN110728664A - Fruit classification system based on computer vision - Google Patents
Fruit classification system based on computer vision Download PDFInfo
- Publication number
- CN110728664A CN110728664A CN201910930085.6A CN201910930085A CN110728664A CN 110728664 A CN110728664 A CN 110728664A CN 201910930085 A CN201910930085 A CN 201910930085A CN 110728664 A CN110728664 A CN 110728664A
- Authority
- CN
- China
- Prior art keywords
- fruit
- image
- defective
- fruits
- defect
- 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.)
- Pending
Links
- 235000013399 edible fruits Nutrition 0.000 title claims abstract description 258
- 230000002950 deficient Effects 0.000 claims abstract description 88
- 230000007547 defect Effects 0.000 claims abstract description 80
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 230000003993 interaction Effects 0.000 claims abstract description 6
- 230000005540 biological transmission Effects 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 25
- 238000003066 decision tree Methods 0.000 claims description 12
- 206010039509 Scab Diseases 0.000 claims description 9
- 238000005286 illumination Methods 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 208000025865 Ulcer Diseases 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 6
- 238000011084 recovery Methods 0.000 claims description 6
- 231100000397 ulcer Toxicity 0.000 claims description 6
- 238000003707 image sharpening Methods 0.000 claims description 3
- 238000003706 image smoothing Methods 0.000 claims description 3
- 239000013589 supplement Substances 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 235000021022 fresh fruits Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8466—Investigation of vegetal material, e.g. leaves, plants, fruits
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a fruit classification system based on computer vision, which is characterized by comprising an image acquisition device, a transmission device, a human-computer interaction interface, an image preprocessing module, a defect judgment module and a fruit classification module, wherein the image acquisition device is used for acquiring images of fruits; by acquiring an image; preprocessing an image; extracting a fruit image, and judging the defect of the fruit image to divide the fruit into a defective fruit and a normal fruit; and extracting a defect image and classifying the defective fruits. The defective fruits and the normal fruits are distinguished by the computer vision technology, manual classification is replaced, and labor cost is saved. Through after dividing into defect fruit and normal fruit with fruit, classify defect fruit once more, classify defect fruit according to defect kind, can utilize according to the different pertinence of defect kind to defect fruit to can carry out defect kind's data statistics, be convenient for analyze the major defect form of this kind of fruit.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a fruit classification system based on computer vision.
Background
China is a big agricultural country and a big fruit production country, and because of the development of machine vision technology, in fruit production, computer vision can replace manual work to carry out high-risk, high-strength and high-repeatability work, fruit classification is one of typical fruit classification, most of the fruit classification in the fruit production at present adopts manual classification, consumes a large amount of manpower, and a small number of fruit classification based on computer vision is also mainly used for distinguishing defective fruits from normal fruits and carrying out rough classification work, but the classification of defective fruits and further the analysis of the main defects of the fruits involve a small amount, and the classification of defective fruits can find out the main defects of the fruits in the season, further reflects some defects in the aspect of planting, carries out feedback analysis on fruit planting production and promotes fruit production.
Disclosure of Invention
In order to overcome the defects that manual fruit classification is needed, a large amount of manpower is consumed, and most of classifications only classify fruits without defects, the invention provides a computer vision-based fruit classification system which classifies fruits and classifies defects of defective fruits, can classify the main defects of the fruits and is beneficial to analyzing the reasons of the defective fruits.
In order to achieve the purpose, the invention adopts the following technical scheme:
a fruit classification system based on computer vision is characterized by comprising an image acquisition device, a transmission device, a human-computer interaction interface, an image preprocessing module, a defect judgment module and a fruit classification module; the operation steps of the system are as follows:
step A, collecting an image;
b, preprocessing an image;
step C, extracting a fruit image, judging the defect of the fruit image, and dividing the fruit into a defective fruit and a normal fruit;
and D, extracting a defect image and classifying the defective fruits.
In the above scheme, the image acquisition device is used for selecting a proper industrial camera and a proper lens according to actual needs and adjusting the position and the angle of the camera to acquire the fruit image. The conveying device is used for completing fruit transportation work by combining parameters set by the human-computer interaction interface and the controller. The human-computer interaction interface is used for setting the starting, stopping and conveying speed of the conveying device. The image preprocessing device is used for smoothing and sharpening the fruit image and recovering information. The defect judging module is used for judging the defects of the fruit images and dividing the fruits into defective fruits and normal fruits. The fruit classification module is used for classifying the defect types of the defective fruits.
The defective fruits and the normal fruits are distinguished by the computer vision technology, manual classification is replaced, and labor cost is saved. Through after dividing into defect fruit and normal fruit with fruit, classify defect fruit once more, classify defect fruit according to defect kind, can utilize according to the different pertinence of defect kind to defect fruit to can carry out defect kind's data statistics, be convenient for analyze the major defect form of this kind of fruit.
Preferably, the equipment used in the step A comprises a CCD industrial camera and lens, an image acquisition card, a trigger, an industrial personal computer, an illumination box and a light source; the CCD industrial camera, the lens and the light source are matched and fixed with the lighting box, highlight on the surface of fruits and shadows of the fruits are reduced by matching the camera with light supplement of the light source, an image captured by the CCD industrial camera is transmitted to the industrial personal computer by the image capture card, the trigger is used for triggering the CCD industrial camera to capture images, and the CCD industrial lens is controlled by the industrial personal computer. The trigger is triggered and then sends a signal to the industrial personal computer, the industrial personal computer receives the signal and then controls the CCD industrial lens to take a picture, a background plate with obvious color difference between the color of the classified fruit is arranged in the illumination box, and the background plate is adopted to facilitate image processing.
Preferably, the image preprocessing step includes image smoothing, image sharpening and image information recovery, wherein a method of combining a linear smoothing filter and a median filter is adopted to smooth the image, and a high-pass filtering method is adopted to sharpen the image. The linear smoothing filter and the median filter are combined to inhibit noise in the original fruit image, and small gaps and gaps in the fruit image are connected, so that the fruit image can be extracted from the background. The image is sharpened through a high-pass filtering method, so that the edge of the image is clearer, the boundary between the background and the fruit image is conveniently identified, and the accuracy of fruit classification is improved.
Preferably, the specific steps of step C are as follows:
step C1, performing background segmentation on the image by a threshold method to obtain a fruit image;
and step C2, restoring information of the misjudged area, setting judgment values of R/B value and R/G value, calculating the average value of R, G, B values of the fruit image, calculating the R/B value of the fruit image and the R/G value of the fruit image, comparing the calculated R/B and R/G values of the fruit image with the judgment values, and judging the fruit image of the defective fruit.
The color difference between the fruit and the background plate is huge, the image can be subjected to background segmentation by adopting a threshold value method to obtain a fruit image, the difference between the surface color of the defect position of the defect fruit and the surface color of the normal fruit is large, so that the R, G, B value of the whole fruit can be greatly influenced, the average value of R, G, B values of the surface color of the normal fruit is obtained, then the critical R/B value and the R/G value of the normal fruit are obtained, the R/B value and the R/G value of the normal fruit are used as judgment values, the R/B value and the R/G value of the fruit image are calculated after the fruit image is obtained, and the R/B value and the R/G value of the image are compared with the judgment values to judge whether the fruit is defective.
Preferably, the specific steps of step D are as follows:
d1, performing information recovery on the edge of the fruit image of the defective fruit, and extracting the defective image of the defective fruit by adopting a threshold method;
and D2, classifying the defective fruits by adopting a decision tree method.
Preferably, the decision tree method comprises the following specific steps of:
step 1, selecting defect types and carrying out characteristic analysis;
setting cracked fruits, rottenness, fruit stalks, scabs and ulcers as root nodes, setting total defect number T, P, R/G value of defect area ratio and L/W ratio of length to width as four characteristic values, and setting preset values of the four characteristic values;
step 2, calling a fruit image and a defect image of the defective fruit, and calculating four characteristic values of the defective fruit according to the fruit image and the image information of the defect image of the defective fruit;
step 3, constructing a decision tree and classifying;
the method comprises the steps of comparing the total number of defects T of defective fruits with a preset value, judging whether the defective fruits are canker fruits or not, if not, comparing the defect area ratio P of the defective fruits with the preset value, judging whether the defective fruits are scab fruits or not, if not, comparing the R/G value of the defective fruits with the preset value, judging whether the defective fruits are peduncle fruits or not, if not, comparing the length-width ratio L/W of the defective fruits with the preset value, judging whether the defective fruits are cracked fruits or not, and if not, judging the defective fruits are rotten fruits.
The invention has the advantages that (1) the fruits are classified based on the image processing of computer vision, thereby reducing the manpower input; (2) the fruit is classified into normal fruits and defective fruits, and then the defective fruits are specifically classified, so that the main defects of the fruits can be overcome, and the analysis of the reasons for generating the defective fruits is facilitated; (3) the solar energy utilization rate is high, and the energy is saved and the environment is protected; (4) the classification accuracy is high.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of a decision tree according to the present invention.
Detailed Description
The invention is further described with reference to the following figures and specific embodiments.
The specific embodiment is as follows: a fruit classification system based on computer vision is characterized by comprising an image acquisition device, a transmission device, a human-computer interaction interface, an image preprocessing module, a defect judgment module and a fruit classification module; the operation steps of the system are as follows:
step A, collecting an image;
b, preprocessing an image;
step C, extracting a fruit image, judging the defect of the fruit image, and dividing the fruit into a defective fruit and a normal fruit;
and D, extracting a defect image and classifying the defective fruits.
The equipment used in the step A comprises a CCD industrial camera, a lens, an image acquisition card, a trigger, an industrial personal computer, an illumination box and a light source; the CCD industrial camera, the lens and the light source are matched and fixed with the lighting box, highlight on the surface of fruits and shadows of the fruits are reduced by matching the camera with light supplement of the light source, an image captured by the CCD industrial camera is transmitted to the industrial personal computer by the image capture card, the trigger is used for triggering the CCD industrial camera to capture images, and the CCD industrial lens is controlled by the industrial personal computer.
The image preprocessing step comprises image smoothing and image sharpening, wherein the image is smoothed by adopting a method of combining a linear smoothing filter and a median filter, and the image is sharpened by adopting a high-pass filtering method.
The specific steps of the step C are as follows:
step C1, performing background segmentation on the image by a threshold method to obtain a fruit image;
and step C2, restoring information of the misjudged area, setting judgment values of R/B value and R/G value, calculating the average value of R, G, B values of the fruit image, calculating the R/B value of the fruit image and the R/G value of the fruit image, comparing the calculated R/B and R/G values of the fruit image with the judgment values, and judging the fruit image of the defective fruit.
The specific steps of the step D are as follows:
d1, performing information recovery on the edge of the fruit image of the defective fruit, and extracting the defective image of the defective fruit by adopting a threshold method;
d2, setting cracked fruits, rotten fruits, fruit stalks, scabs, ulcers and defect-free as root nodes, setting total number T of defects, P, R/G of defect area ratio and L/W of length-width ratio as four characteristic values, and setting preset values of the four characteristic values;
d3, calling the fruit image and the defect image of the defective fruit, and calculating four characteristic values of the defective fruit according to the fruit image and the image information of the defect image of the defective fruit;
and D4, comparing the total number of defects T of the defective fruit with the preset value thereof, judging whether the defective fruit is canker fruit, if not, comparing the defect area ratio P of the defective fruit with the preset value thereof, judging whether the defective fruit is scab fruit, if not, comparing the R/G value of the defective fruit with the preset value thereof, judging whether the defective fruit is peduncle fruit, if not, comparing the length-width ratio L/W of the defective fruit with the preset value thereof, judging whether the defective fruit is cracked fruit, and if not, judging the defective fruit is rotten fruit.
In this embodiment, the image acquisition device is a rectangular illumination box, the CCD industrial camera is installed right above the illumination box, the light source is 6 fluorescent lamps, the light source has a continuous and linear mixed spectrum, the CCD industrial camera has the characteristics of a high-sensitivity color-compensating filter, a programmable shutter from 1/60 to 1/10000 seconds, an automatic shutter, etc., the image acquisition card is placed in an industrial control machine to remotely control the CCD industrial camera, the fruit studied in this embodiment is orange, the background in the illumination box is selected to be blue, the classification flow of this embodiment is shown in fig. 1,
step 1, collecting an image;
step 2, preprocessing an image;
in this embodiment, the image is smoothed by a method of combining a linear smoothing filter and a median filter, and then sharpened by a high-pass filtering method
Step 3, performing background segmentation on the image by a threshold method to obtain a fruit image;
and 4, recovering information of the misjudged area, setting judgment values of the R/B value and the R/G value, calculating an average value of R, G, B values of the fruit image, calculating the R/B value of the fruit image and the R/G value of the fruit image, comparing the calculated R/B and R/G values of the fruit image with the judgment values, and judging the fruit image of the defective fruit.
Step 5, performing information recovery on the edge of the fruit image of the defective fruit, and extracting the defective image of the defective fruit by adopting a threshold value method;
step 6, selecting defect types and carrying out characteristic analysis;
setting cracked fruits, rottenness, fruit stalks, scabs and ulcers as root nodes, setting total defect number T, defect area ratio P, R/G value and length-width ratio L/W as four characteristic values, and setting preset values of the four characteristic values;
step 7, calling a fruit image and a defect image of the defective fruit, and calculating four characteristic values of the defective fruit according to the fruit image and the image information of the defect image of the defective fruit;
step 8, constructing a decision tree and classifying;
comparing the total number of defects T of the defective fruit with a preset value thereof to judge whether the defective fruit is canker fruit or not, if not, comparing the defect area ratio P of the defective fruit with the preset value thereof to judge whether the defective fruit is scab fruit or not, if not, comparing the R/G value of the defective fruit with the preset value thereof to judge whether the defective fruit is peduncle fruit or not, if not, comparing the length-width ratio L/W of the defective fruit with the preset value thereof to judge whether the defective fruit is cracked fruit or not, and if not, judging that the defective fruit is rotten fruit.
After the image is collected, the background needs to be cut off, a threshold value method is adopted for cutting off, wherein when the threshold value method is adopted, a threshold value is dynamically generated for each image according to the background and the surface color characteristics of each fruit, then the segmentation is carried out, and the calculation formula of the threshold value is as follows:
T＝R+G-α*B
where T is the threshold, R, G, B is the red vector, green vector, blue vector of the pixel, respectively. α is a dynamic coefficient, and is calculated by the formula:
α＝(∑R+∑G)/∑B
then, a scatter diagram is made, the fact that the image is subjected to background segmentation by using T-40 to obtain a fruit image is determined, because the defect of the orange surface is black or gray in most cases, the image is mistaken for the background to be segmented when the background is segmented, and the image needs to be subjected to information recovery, and the principle is as follows: let M be the set of pixels of the captured image, B and S be the set of pixels of the background image and the set of pixels of the fruit image, respectively, then,
S∪B＝M
and:
B＝T∪F
wherein T is the correct background image pixel point set, F is the mistakenly recognized background fruit image pixel point set, then,
let p (x, y) be belonged to B, construct a circle with p as the center, L as the radius length, r as the center, if
For the
Otherwise p (x, y) is ∈ F. And scanning the shot image by taking p as a circle center and a line segment with the length of L in a counterclockwise mode by taking 1 degree as a step length, stopping scanning if all points of one line segment are found to be background points, and judging that p is a background image point, otherwise, judging that p is a fruit image point.
The ID3 decision tree algorithm is adopted to construct a decision tree, and the process is that firstly, the similar collection is carried out on the current sub-cases, secondly, the mutual information of the attributes of each set is calculated, the attribute A with the maximum mutual information is selected, the sub-cases are divided into a plurality of subsets according to the attribute A, and the algorithm is called recursively until each subset has only one category. Selecting the total number T of defects, the P, R/G value of the defect area ratio and the L/W ratio as four characteristic values as input characteristics to construct a decision tree, wherein the selection basis can be that after the defects to be classified are determined, the color characteristics, the shape characteristics and the texture characteristics of the orange defects are qualitatively analyzed by a certain amount of samples to determine the main characteristic values for distinguishing the defects, and technicians in the field can determine the characteristic values by experiments, in the embodiment, the defects to be classified are set as dehiscent fruits, rottenness, fruit stalks, scabs and ulcers, and the color characteristics, the shape characteristics and the texture characteristics of the orange defects are qualitatively analyzed to determine that the total number T of the defects, the P, R/G value of the defect area ratio and the L/W ratio are respectively selected as the superior characteristic values, then, a certain sample is adopted to carry out information entropy-increasing calculation, and the characteristic values are determined to be sorted into the total defect number T, the defect area ratio P, R/G value and the length-width ratio L/W, so that the schematic diagram of the decision tree is obtained and is shown in FIG. 2.
The system is used for classifying the oranges, the comprehensive accuracy rate is over 80 percent, the accuracy rate of fruit stalks and ulcers is 100 percent, the misjudgment rate of fresh fruits is about 10 percent, the accuracy rate of the classification system is high on the whole, and the manpower is liberated.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and all simple modifications, alterations and equivalents of the above embodiments according to the technical spirit of the present invention are still within the protection scope of the technical solution of the present invention.
Claims (6)
1. A fruit classification system based on computer vision is characterized by comprising an image acquisition device, a transmission device, a human-computer interaction interface, an image preprocessing module, a defect judgment module and a fruit classification module; the operation steps of the system are as follows:
step A, collecting a fruit image;
b, preprocessing an image;
step C, extracting a fruit image, judging the defect of the fruit image, and dividing the fruit into a defective fruit and a normal fruit;
and D, extracting a defect image and classifying the defective fruits.
2. The fruit classification system based on computer vision as claimed in claim 1, wherein the equipment used in step a comprises a CCD industrial camera and lens, an image acquisition card, a trigger, an industrial personal computer, an illumination box and a light source; the CCD industrial camera, the lens and the light source are matched and fixed with the lighting box, highlight on the surface of fruits and shadows of the fruits are reduced by matching the camera with light supplement of the light source, an image captured by the CCD industrial camera is transmitted to the industrial personal computer by the image capture card, the trigger is used for triggering the CCD industrial camera to capture images, and the CCD industrial lens is controlled by the industrial personal computer.
3. The computer vision based fruit classification system of claim 1, wherein the image preprocessing step comprises image smoothing and image sharpening, wherein the image is smoothed by a combination of a linear smoothing filter and a median filter, and wherein the image is sharpened by a high pass filter.
4. The computer vision based fruit sorting system of claim 1, wherein the specific steps of step C are as follows:
step C1, performing background segmentation on the image by a threshold method to obtain a fruit image;
and step C2, restoring information of the misjudged area, setting judgment values of R/B value and R/G value, calculating the average value of R, G, B values of the fruit image, calculating the R/B value of the fruit image and the R/G value of the fruit image, comparing the calculated R/B and R/G values of the fruit image with the judgment values, and judging the fruit image of the defective fruit.
5. The computer vision based fruit sorting system of claim 1, wherein the specific steps of step D are as follows:
d1, performing information recovery on the edge of the fruit image of the defective fruit, and extracting the defective image of the defective fruit by adopting a threshold method;
and D2, classifying the defective fruits by adopting a decision tree method.
6. The computer vision based fruit classification system of claim 5, wherein the decision tree method for classifying defective fruits comprises the following specific steps:
step 1, selecting defect types and carrying out characteristic analysis;
setting cracked fruits, rottenness, fruit stalks, scabs and ulcers as root nodes, setting total defect number T, P, R/G value of defect area ratio and L/W ratio of length to width as four characteristic values, and setting preset values of the four characteristic values;
step 2, calling a fruit image and a defect image of the defective fruit, and calculating four characteristic values of the defective fruit according to the fruit image and the image information of the defect image of the defective fruit;
step 3, constructing a decision tree and classifying;
the method comprises the steps of comparing the total number of defects T of defective fruits with a preset value, judging whether the defective fruits are canker fruits or not, if not, comparing the defect area ratio P of the defective fruits with the preset value, judging whether the defective fruits are scab fruits or not, if not, comparing the R/G value of the defective fruits with the preset value, judging whether the defective fruits are peduncle fruits or not, if not, comparing the length-width ratio L/W of the defective fruits with the preset value, judging whether the defective fruits are cracked fruits or not, and if not, judging the defective fruits are rotten fruits.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910930085.6A CN110728664A (en) | 2019-09-27 | 2019-09-27 | Fruit classification system based on computer vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910930085.6A CN110728664A (en) | 2019-09-27 | 2019-09-27 | Fruit classification system based on computer vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110728664A true CN110728664A (en) | 2020-01-24 |
Family
ID=69219523
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910930085.6A Pending CN110728664A (en) | 2019-09-27 | 2019-09-27 | Fruit classification system based on computer vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110728664A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112949438A (en) * | 2021-02-19 | 2021-06-11 | 武汉工程大学 | Fruit visual classification method and system based on Bayesian network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008249413A (en) * | 2007-03-29 | 2008-10-16 | Fujinon Corp | Defect detection method and device |
| CN101912847A (en) * | 2010-08-02 | 2010-12-15 | 扬州福尔喜果蔬汁机械有限公司 | Fruit grading system and method based on DSP machine vision |
-
2019
- 2019-09-27 CN CN201910930085.6A patent/CN110728664A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008249413A (en) * | 2007-03-29 | 2008-10-16 | Fujinon Corp | Defect detection method and device |
| CN101912847A (en) * | 2010-08-02 | 2010-12-15 | 扬州福尔喜果蔬汁机械有限公司 | Fruit grading system and method based on DSP machine vision |
Non-Patent Citations (1)
| Title |
|---|
| 庞江伟: "基于计算机视觉的脐橙表面常见缺陷种类识别的研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112949438A (en) * | 2021-02-19 | 2021-06-11 | 武汉工程大学 | Fruit visual classification method and system based on Bayesian network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107486415B (en) | Thin bamboo strip defect online detection system and detection method based on machine vision | |
| Qing et al. | Automated detection and identification of white-backed planthoppers in paddy fields using image processing | |
| Safren et al. | Detection of green apples in hyperspectral images of apple-tree foliage using machine vision | |
| CN114757900B (en) | Artificial intelligence-based textile defect type identification method | |
| CN111860533A (en) | Image recognition method and device, storage medium and electronic device | |
| CN106584209A (en) | Real-time online monitoring method for tool wear based on cloud manufacturing | |
| CN108181316B (en) | Bamboo strip defect detection method based on machine vision | |
| CN111275679A (en) | Solar cell defect detection system and method based on image | |
| CN111080691A (en) | Infrared hot spot detection method and device for photovoltaic module | |
| CN109145848B (en) | Wheat ear counting method | |
| CN102915432A (en) | Method and device for extracting vehicle-bone microcomputer image video data | |
| CN110675395A (en) | Intelligent on-line monitoring method for power transmission line | |
| CN101976350A (en) | Grain storage pest detection and identification method based on video analytics and system thereof | |
| CN104198497A (en) | Surface defect detection method based on visual saliency map and support vector machine | |
| CN104048966B (en) | The detection of a kind of fabric defect based on big law and sorting technique | |
| WO2006113979A1 (en) | Method for identifying guignardia citricarpa | |
| CN112044808A (en) | Household garbage recognition system | |
| US20230360411A1 (en) | Cherry picking and classifying method and device based on machine vision | |
| CN115294377A (en) | System and method for identifying road cracks | |
| CN111563519A (en) | Tea leaf impurity identification method based on Stacking weighted ensemble learning and sorting equipment | |
| CN114029943A (en) | Target grabbing and positioning method and system based on image data processing | |
| CN115841633A (en) | Power tower and power line associated correction power tower and power line detection method | |
| CN202815869U (en) | Vehicle microcomputer image and video data extraction apparatus | |
| CN114120218A (en) | River course floater monitoring method based on edge calculation | |
| CN114723668A (en) | Cathode copper quality detection method and system based on area-array camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200124 |
|
| RJ01 | Rejection of invention patent application after publication |