CN106548480B - Quick agricultural product volume measuring device and measuring method based on machine vision - Google Patents
Quick agricultural product volume measuring device and measuring method based on machine vision Download PDFInfo
- Publication number
- CN106548480B CN106548480B CN201611202105.0A CN201611202105A CN106548480B CN 106548480 B CN106548480 B CN 106548480B CN 201611202105 A CN201611202105 A CN 201611202105A CN 106548480 B CN106548480 B CN 106548480B
- Authority
- CN
- China
- Prior art keywords
- volume
- agricultural product
- measured
- industrial cameras
- ccd industrial
- 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.)
- Active
Links
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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
-
- 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
- G06T2207/20081—Training; Learning
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a quick agricultural product volume measuring device and a measuring method based on machine vision, wherein three CCD industrial cameras are arranged according to a certain angle, the three CCD industrial cameras are connected with a camera driver and an image acquisition processor, image information of a training sample is obtained in parallel by utilizing the three CCD industrial cameras, images are segmented and edge characteristics are extracted, the maximum length size in three-face images and the maximum length size in the orthogonal direction thereof are obtained by an external rectangle method, PLSR modeling is carried out to obtain an agricultural product volume prediction model, and six size parameters of the three-face images of the agricultural product to be measured are substituted into the volume prediction model to quickly obtain the volume of the agricultural product. The low real-time performance of the three-dimensional virtual reproduction process is ignored through a pre-established volume prediction model, and the volume of the agricultural product to be measured is rapidly obtained through the volume prediction model in the volume measurement stage, so that necessary grading basis is provided for intelligent sorting of the agricultural product with multiple characteristics such as comprehensive color, volume and the like.
Description
Technical Field
The invention belongs to the field of agricultural product screening, and particularly relates to a machine vision-based agricultural product volume rapid measuring device and a measuring method.
Background
At present, the color feature detection technology of agricultural products based on machine vision is mature, and the color feature detection technology is successfully applied to commercial agricultural product classifying and screening devices, such as a rice color selector, a corn color selector and the like, and damaged grains, abnormal grains and other impurities in the agricultural products are removed according to the reflection degree of light of different colors, mildews, white abdomen and the like. The existing agricultural product grading screening method mainly uses a color selection technology with single color characteristics as a criterion, and does not comprehensively consider the characteristics of appearance, volume, especially volume and the like of agricultural products, so that the current agricultural product screening method based on the color selection technology has certain limitation, along with the improvement of living standard and the advancement of science and technology, the quality requirements of people on agricultural products are higher and higher, the sorting basis of the agricultural products is also more and more severe and refined, and obviously, the agricultural product screening mode with the color characteristics of the single sorting basis is difficult to adapt to the trend of more and more severe sorting and refinement. In order to realize the sorting of the agricultural products with higher quality, besides the detection of the color characteristics of the agricultural products, other characteristics such as volume and the like should be detected, so that the sorting of the agricultural products with higher grade is realized conveniently.
At present, the color feature detection technology of agricultural products based on machine vision is mature and is practically applied and popularized, and the appearance feature detection method based on machine vision is widely studied by a plurality of scholars, but the rapid volume feature detection method is still to be further perfected: the volume measurement based on the monocular vision system omits three-dimensional virtual reproduction, under the condition of regular and known appearance, the volume of the object to be measured is predicted by acquiring the dimensional characteristic parameters of the part, the real-time performance is high, but the information content in a single image is less, the volume measurement error is larger when the factors such as shape defect influence, and the volume measurement precision based on the three-dimensional reconstruction technology is higher, but the real-time performance is not high due to longer time consumption of the virtual reproduction process.
Disclosure of Invention
Aiming at the defect that the existing agricultural product sorting mode mainly uses single color characteristics as sorting basis, the invention provides a quick agricultural product volume measuring device based on machine vision.
In order to solve the technical problems, the invention adopts the following technical scheme: the utility model provides a quick measuring device of agricultural product volume based on machine vision, includes three CCD industry cameras, camera driver and the image acquisition processor of mutually angulation, and the camera driver meets with three CCD industry cameras simultaneously, and the image acquisition processor links to each other with three CCD industry cameras and camera driver simultaneously, the visual line of three CCD industry cameras intersect on same visual line intersection point, and the angle of three CCD industry cameras is based on the three face image coverage after the formation of image surpassing the whole three-dimensional surface of agricultural product that awaits measuring, and three CCD industry cameras can reciprocate along its visual line; each CCD industrial camera is also provided with a light source and a background plate on its opposite side.
When the volume of the agricultural product to be measured is measured, the agricultural product to be measured is placed on the intersection point of the visual lines of the three CCD industrial cameras, the distance between the three CCD industrial cameras and the agricultural product to be measured is adjusted, so that the three-sided images of the agricultural product to be measured after being imaged by the three CCD industrial cameras exceed the whole three-dimensional surface of the agricultural product to be measured, the three-sided images of the agricultural product to be measured are collected, the images are analyzed and processed by utilizing the camera driver and the image collecting processor, six size parameters of the three-sided images are obtained, and the six size parameters are substituted into a pre-established agricultural product volume prediction model to predict the volume of the agricultural product to be measured. The color of the background plate is obviously different from that of the agricultural product to be detected, so that the obvious difference between the foreground color and the background color of the imaging of the agricultural product to be detected can be ensured, and a necessary precondition is provided for rapid image segmentation.
In order to increase the degree of automation, the measuring device of the invention also comprises a conveying device, wherein the conveying device comprises a conveying belt and a conveying pipeline, the conveying belt corresponds to the upper end of the conveying pipeline, two switch-type detection sensors are arranged at the outlet of the conveying pipeline, and the switch-type detection sensors are connected with a camera driver; the three CCD industrial cameras are fixed below the conveying pipeline, and the visual line intersection points of the three CCD industrial cameras are located right below the conveying pipeline. And the agricultural products to be detected are transmitted by utilizing the transmission device, when the agricultural products fall into the transmission pipeline from the transmission belt and fall between the switch type detection sensors, the switch type detection sensors send trigger signals to the camera driver, and after proper time delay, when the agricultural products to be detected fall to the intersection point of the visual lines of the three industrial cameras, the camera driver triggers the three CCD industrial cameras to automatically collect the graphic information of the agricultural products to be detected.
According to the quick agricultural product volume measuring device based on machine vision, three CCD industrial cameras with angles are utilized to collect three images of agricultural products in parallel, the camera driver and the image collecting processor are utilized to process the images, the appearance characteristics of the agricultural products to be measured can be quickly obtained, and the pre-established agricultural product volume predicting model is combined to quickly predict the volume of the agricultural products to be measured, so that the quick and convenient agricultural product volume measuring device is quick and convenient, the predicted value is closer to the real value, and the real-time performance is high.
In order to solve the technical problems, the invention also provides a quick agricultural product volume measuring method based on machine vision, which adopts three high-speed industrial CCD industrial cameras which are at a certain angle to obtain images of agricultural products to be measured in parallel, and the volume characteristics of the agricultural products to be measured are quickly obtained through real-time processing of three-side images, so that a grading basis is provided for intelligent sorting of the agricultural products with multiple characteristics such as comprehensive colors, volume sizes and the like.
The invention discloses a quick agricultural product volume measuring method based on machine vision, which comprises the following steps of:
(1) Establishing a volume prediction model: selecting a plurality of agricultural products with known volumes as training samples, placing the training samples on the intersection points of visual lines of three CCD industrial cameras of the measuring device, utilizing three CCD industrial cameras which are mutually angled to collect three-side images of each training sample in parallel, carrying out image segmentation and edge feature extraction on the collected three-side images, respectively obtaining the maximum length size in the three-side images and the maximum length size in the orthogonal direction through an external rectangle method, combining a partial least square regression algorithm to obtain an agricultural product volume prediction model taking six size parameters in the three-side images as input variables, wherein the following formula is shown,
V＝α 0 +α 1 A 1 +α 2 B 1 +α 3 A 2 +α 4 B 2 +α 5 A 3 +α 6 B 3
wherein V is the volume of the agricultural product to be measured, alpha 0 Regression intercept, alpha 1 、α 2 、……、α 6 -regression coefficients between the original independent variable and the original dependent variable;
(2) Six size parameters of the agricultural product to be measured are obtained: placing the agricultural product to be detected on the intersection point of the visual lines of three CCD industrial cameras, collecting three-sided images of the agricultural product to be detected, carrying out image segmentation and edge feature extraction on the collected three-sided images, and then respectively obtaining six dimension parameters A of the three-sided images of the agricultural product to be detected by an external rectangle method 1 、B 1 、A 2 、B 2 And A 3 、B 3 ；
(3) Calculating the volume of the agricultural product to be measured: substituting the six dimensional parameters of the three-sided image obtained in the step (2) into the volume prediction model of the step (1), and calculating the volume V of the agricultural product to be detected.
According to the invention, by referring to the application of machine vision in three-dimensional reconstruction, image information of agricultural products is obtained in parallel by adopting three CCD industrial cameras at an angle, appearance characteristics of the agricultural products are obtained by carrying out methods such as rapid segmentation, edge characteristic extraction and the like on output images of the three CCD industrial cameras, in order to avoid low real-time performance of a volume measurement method based on a three-dimensional reconstruction technology, maximum length dimensions in three-face images which are output in parallel and maximum length dimensions in orthogonal directions of the three-face images are respectively obtained by adopting an external rectangle method, a dimension reduction algorithm-partial least square regression algorithm (PLSR) in sample training and machine learning is combined to obtain an agricultural product volume prediction model taking six dimension parameters in the three-face images as input variables, the volume of the agricultural products to be detected is verified by using the volume prediction model, and the volume of the agricultural products to be detected obtained by using the volume prediction model is more approximate to a true value; and substituting the six dimensional parameters of the three-sided image of the agricultural product to be measured into a volume prediction model to quickly acquire the volume of the agricultural product, thereby realizing quick measurement of the volume of the agricultural product to be measured. The low real-time performance of the three-dimensional virtual reproduction process is ignored through the agricultural product volume prediction model established in the sample training stage, and the volume of the agricultural product to be measured is rapidly acquired through the volume prediction model in the volume measurement stage, so that necessary grading basis is provided for intelligent sorting of the agricultural product with multiple characteristics such as comprehensive color, volume and the like: on the basis of keeping the existing agricultural product color selection technology, the intelligent sorting device is added with the agricultural product volume feature detection to obtain the multi-aspect features of the agricultural products to be detected, the grading basis of the agricultural products is finer, and the optimization of the high-quality agricultural products can be ensured.
According to the invention, through obtaining the longest dimension and the appearance dimension parameters in the orthogonal direction, the rotation invariance of the agricultural product to be measured in the conveying process can be ensured, and the volume characteristic acquisition is obtained through a volume prediction model which is independent of the actual volume measurement process and is established in the sample training stage, so that the multi-characteristic measurement efficiency of the agricultural product to be measured can be greatly improved.
Drawings
FIG. 1 is a schematic diagram of a machine vision-based agricultural product volume rapid measurement device of the present invention.
FIG. 2 is a flow chart of the machine vision based method for rapidly measuring the volume of agricultural products of the present invention.
FIG. 3 is six dimensional parameters of a three-sided image of an agricultural product to be measured obtained by the circumscribed rectangle method of the present invention.
Detailed Description
The technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments.
The quick agricultural product volume measuring device based on machine vision comprises a conveying device, a switch-type detection sensor 3, three high-speed colorful CCD industrial cameras 41, 42 and 43 which are mutually angled, wherein camera lenses of the three CCD industrial cameras 41, 42 and 43 all adopt variable-focus lenses, a camera driver 8 is driven by a driving plate taking an FPGA as a core to realize camera driving and related algorithms (such as image filtering, image segmentation and other preprocessing, volume prediction algorithms and the like), an image acquisition processor 7 is realized by a processing plate taking a DSP device as a core, and the camera driver 8 and the image acquisition processor 7 are connected with the CCD industrial cameras 41, 42 and 43. The conveying device comprises a conveying belt 1 and a conveying pipeline 2, wherein the conveying belt 1 corresponds to the upper end of the conveying pipeline 2, two switch-type detection sensors 3 are arranged at the position of an outlet O ' of the conveying pipeline 2, the switch-type detection sensors 3 are connected with a camera driver 8, the camera driver 8 is simultaneously connected with three CCD industrial cameras 41, 42 and 43, the three CCD industrial cameras 41, 42 and 43 are fixed below the conveying pipeline 2, visual lines L1, L2 and L3 of the three CCD industrial cameras 41, 42 and 43 are intersected at the same point O ', and a visual line intersection point O ' is located right below the conveying pipeline 2. The angles theta 1, theta 2 and theta 3 of the three CCD industrial cameras 41, 42 and 43 are based on the fact that the coverage of the imaged three-side images exceeds the whole three-dimensional surface of the agricultural product to be measured, and the three CCD industrial cameras 41, 42 and 43 can move back and forth along the visual lines L1, L2 and L3 of the three CCD industrial cameras; each of the CCD industrial cameras 41, 42, 43 is further provided with a light source 51, 52, 53 and a background plate 61, 62, 63 on opposite sides thereof, the light sources 51, 52, 53 are LED light sources, the stability is good and the service life is long, the colors of the background plates 61, 62, 63 are obviously different from the colors of the agricultural products to be detected, the obvious difference between the foreground color and the background color of the imaging of the agricultural products to be detected can be ensured, and necessary preconditions are provided for rapid image segmentation.
The quick agricultural product volume measuring method based on machine vision has a flow shown in fig. 2, and comprises the following steps:
(1) Establishing a volume prediction model: a plurality of agricultural products with known volumes are selected as training samples and placed on the conveyor belt 1, distances between the intersections O' of the three CCD industrial cameras 41, 42 and 43 with the visual lines with high speed color are adjusted and calibrated respectively, and as shown in figure 1, three-face images of the three CCD industrial cameras 41, 42 and 43 after the training samples are imaged exceed the whole three-dimensional surface of the training samples. When the training sample is sent into the conveying pipeline 2 by the conveyor belt 1, when the training sample falls to the O 'position, the high-speed switch type detection sensor 3 sends a trigger signal to the camera driver 8, after proper time delay, when the training sample falls to the O' position, the camera driver 8 triggers the three CCD industrial cameras 41, 42 and 43 to collect three-side images of the training sample in parallel and transmit the three-side images to the image collecting processor 7, and the image collecting processor 7 carries out graying treatment, image segmentation and edge feature extraction on the images. In order to ensure rotation invariance of agricultural products in the process of conveying, a circumscribed rectangle method is adopted, the respective longest dimension in three-sided images and the maximum dimension parameter in the orthogonal direction are obtained by combining the calibration parameters of a camera, as shown in fig. 3, and are used as input variables, and a volume prediction model is built by combining a dimension reduction algorithm-partial least squares regression algorithm (PLSR) in the process of sample training based on machine learning, wherein the following formula is shown:
V＝α 0 +α 1 A 1 +α 2 B 1 +α 3 A 2 +α 4 B 2 +α 5 A 3 +α 6 B 3
wherein: v-volume of agricultural product to be measured, alpha 0 Regression intercept, alpha 1 、α 2 、…、α 6 Regression coefficients between original independent variables and original dependent variables, A 1 、B 1 、A 2 、B 2 And A 3 、B 3 -six dimensional parameters of the three-sided image of the agricultural product to be measured.
The known volume of the training sample is utilized to verify the volume prediction model, if the accuracy meets the requirement, the volume prediction model is qualified, and the method can be used for measuring the volume of the agricultural product to be measured; if the precision does not meet the requirement, the PLSR modeling is needed to be carried out again until the precision meets the requirement;
(2) Six size parameters of the agricultural product to be measured are obtained: collecting three-sided images of agricultural products to be detected according to the method in the step (1), carrying out image segmentation and edge feature extraction on the collected three-sided images, and then respectively obtaining six dimension parameters A of the three-sided images of the agricultural products to be detected by a circumscribed rectangle method 1 、B 1 、A 2 、B 2 、A 3 、B 3 ；
(3) Calculating the volume of the agricultural product to be measured: six dimension parameters A of the three-side image obtained in the step (2) 1 、B 1 、A 2 、B 2 、A 3 、B 3 Substituting the obtained volume prediction model in the step (1) to calculate the volume V of the agricultural product to be detected.
The volume of potatoes measured by the method of the invention was 88.60cm 3 The volume of the same potato measured by the water discharge method is 89.28cm 3 The error was 0.76%.
The following describes a specific procedure for the establishment of the volume prediction model with reference to fig. 2 and 3:
the dimensionality reduction algorithm-partial least square method (PLS) in machine learning combines the functions of multiple regression analysis, principal component analysis, correlation analysis and the like, and Partial Least Square Regression (PLSR) based on the partial least square method principle can be used for solving the problems that multiple correlations among independent variables or sample capacity is smaller than the number of the variables and the like in the multiple regression analysis, has the advantages that other regression analysis does not have in numerous regression analysis, is widely applied to a plurality of fields at present and has good effects.
Assuming that the volume of the agricultural product to be measured is V, six dimensional parameters are A as shown in figure 3 1 、B 1 、A 2 、B 2 And A 3 、B 3 For ease of analysis, the following variables were substituted: let the single dependent variable y=v, the six dimensional parameters be independent variables, and set to: x is x 1 ＝A 1 ,x 2 ＝B 1 ,x 3 ＝A 2 ,x 4 ＝B 2 ,x 5 ＝A 3 ,x 6 ＝B 3 。
Selecting u different agricultural products with known volumes as training samples, and obtaining sample data X and Y of independent variables and dependent variables, wherein X u×6 To explain the matrix, Y u×1 Is a response matrix. The sample matrix X is normalized according to the descending order and the cross effectiveness principle, and the variance Var (t i ) Sum covariance Cov (t) i Y) all as large a component t as possible 1 ,t 2 ,…,t h (h.ltoreq.6), then by establishing y and t 1 ,t 2 ,…,t h And finally obtaining y and x according to regression equation of (2) 1 ,x 2 ,…,x h Is a regression equation of (2).
The matrix X and the dependent variable Y are subjected to standardization processing to obtain a standardized variable matrix E 0 And column vector ζ 0 ：
Wherein the method comprises the steps of
Mu in the middle jx 、S jx -jth argument x j Sample mean and sample standard deviation, mu y 、S y Dependent variable y j Is a sample mean and a sample standard deviation;
from E 0 Extract 1 st component:
and execute E 0 And zeta 0 For component 1 t 1 Is a regression of:
wherein the method comprises the steps of
In p 1 、r 1 Regression coefficient, E 1 、ζ 1 -residual matrices and vectors of regression equations;
continuing to extract component 2 t 2 And execute E 1 And zeta 1 For component t 2 2 Is a regression of:
wherein the method comprises the steps of
In p 2 、r 2 Regression coefficient, E 2 、ζ 2 -residual matrices and vectors of regression equations;
continuously extracting the components to obtain m components t 1 ，t 2 ，…，t m And execute ζ 0 Regression on m components, namely:
ζ 0 ＝r 1 t 1 +r 2 t 2 +r 3 t 3 +…+r m t m finally, the volume regression model is restored to the form of an original variable to obtain the volume regression model of the agricultural product to be measured:
y＝α 0 +α 1 x 1 +α 2 x 2 +...+α 6 x 6 ，
namely the volume V of the agricultural product to be measured and 6 dimension parameters A 1 、B 1 、A 2 、B 2 、A 3 、B 3 The relation between them is:
V＝α 0 +α 1 A 1 +α 2 B 1 +α 3 A 2 +α 4 B 2 +α 5 A 3 +α 6 B 3 ，
alpha in the formula 0 Regression intercept, alpha 1 、α 2 、…、α 6 -regression coefficients between the original independent variable and the original dependent variable.
Claims (5)
1. Quick measuring device of agricultural product volume based on machine vision, its characterized in that: the system comprises three CCD industrial cameras which are mutually angled, a camera driver and an image acquisition processor, wherein the camera lenses of the three industrial cameras adopt variable-focus lenses, the camera driver is driven by a driving board taking an FPGA as a core to realize camera driving and related algorithms, the image acquisition processor is realized by a processing board taking a DSP device as a core to acquire and process images, the camera driver is connected with the three CCD industrial cameras, the image acquisition processor is simultaneously connected with the three CCD industrial cameras and the camera driver, visual lines of the three CCD industrial cameras are intersected at the same visual intersection point, when the volume of agricultural products to be measured is measured, the agricultural products to be measured are placed on the visual line intersection point of the three CCD industrial cameras, the three CCD industrial cameras and the agricultural products to be measured are moved forwards and backwards along the visual lines of the three CCD industrial cameras, the distance between the three CCD industrial cameras and the agricultural products to be measured is adjusted, the three images of the three CCD industrial cameras after imaging the agricultural products to be measured exceed the whole three-dimensional surface to be measured, the three images of the agricultural products to be measured are acquired, the three-dimensional parameters of the three-dimensional images of the agricultural products to be measured are obtained by analyzing and processing the images by the camera driver and the image acquisition processor, six dimensional parameters of the three-dimensional images are obtained, and are substituted into a pre-established volume prediction model to be predicted;
each CCD industrial camera is also provided with a light source and a background plate on its opposite side.
2. The quick measuring device for the volume of agricultural products based on machine vision according to claim 1, further comprising a conveying device, wherein the conveying device comprises a conveying belt and a conveying pipeline, the conveying belt corresponds to the upper end of the conveying pipeline, two switch-type detection sensors are arranged at the outlet of the conveying pipeline, and the switch-type detection sensors are connected with a camera driver; the three CCD industrial cameras are fixed below the conveying pipeline, and the visual line intersection points of the three CCD industrial cameras are located right below the conveying pipeline.
3. The quick agricultural product volume measuring method based on machine vision is characterized by comprising the following steps of:
(1) Establishing a volume prediction model: selecting a plurality of agricultural products with known volumes as training samples, placing the training samples on the intersection points of visual lines of the measuring device according to claim 1 or 2, utilizing three CCD industrial cameras which are mutually angled to collect three-side images of each training sample in parallel, carrying out image segmentation and edge feature extraction on the collected three-side images, then respectively obtaining the maximum length size in the three-side images and the maximum length size in the orthogonal direction through an external rectangle method, and obtaining an agricultural product volume prediction model taking six size parameters in the three-side images as input variables by combining a partial least square regression algorithm;
(2) Six size parameters of the agricultural product to be measured are obtained: placing the agricultural product to be measured on the intersection point of the visual lines of the measuring device according to claim 1 or 2, collecting three-side images of the agricultural product to be measured, performing image segmentation and edge feature extraction on the collected three-side images, and then respectively obtaining six dimension parameters A of the three-side images of the agricultural product to be measured by a circumscribed rectangle method 1 、B 1 、A 2 、B 2 And A 3 、B 3 ；
(3) Calculating the volume of the agricultural product to be measured: substituting the six dimensional parameters of the three-sided image obtained in the step (2) into the volume prediction model of the step (1), and calculating the volume V of the agricultural product to be detected.
4. A machine vision based agricultural product volume rapid measurement method according to claim 3, characterized in that: the volume prediction model is shown as follows:
V＝α 0 +α 1 A 1 +α 2 B 1 +α 3 A 2 +α 4 B 2 +α 5 A 3 +α 6 B 3
wherein alpha is 0 Regression intercept, alpha 1 、α 2 、……、α 6 -regression coefficients between the original independent variable and the original dependent variable.
5. A machine vision based agricultural product volume rapid measurement method according to claim 3, characterized in that: and (3) verifying the agricultural product volume prediction model in the step (1) by using the known volume of the training sample, and if the verification result does not meet the precision requirement, carrying out PLSR modeling again until the verification result meets the precision requirement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611202105.0A CN106548480B (en) | 2016-12-23 | 2016-12-23 | Quick agricultural product volume measuring device and measuring method based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611202105.0A CN106548480B (en) | 2016-12-23 | 2016-12-23 | Quick agricultural product volume measuring device and measuring method based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106548480A CN106548480A (en) | 2017-03-29 |
| CN106548480B true CN106548480B (en) | 2023-05-26 |
Family
ID=58397819
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611202105.0A Active CN106548480B (en) | 2016-12-23 | 2016-12-23 | Quick agricultural product volume measuring device and measuring method based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106548480B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107410933B (en) * | 2017-04-05 | 2020-06-05 | 浙江省海洋开发研究院 | Fish processing device and control method |
| CN107080148B (en) * | 2017-04-05 | 2020-06-05 | 浙江省海洋开发研究院 | Aquatic product processing system and control method thereof |
| CN107545247B (en) * | 2017-08-23 | 2020-05-12 | 北京伟景智能科技有限公司 | Stereo cognition method based on binocular recognition |
| CN107957246B (en) * | 2017-11-29 | 2019-12-17 | 北京伟景智能科技有限公司 | binocular vision-based method for measuring geometric dimension of object on conveyor belt |
| CN110602355A (en) * | 2018-05-25 | 2019-12-20 | 上海翌视信息技术有限公司 | Image acquisition method |
| CN113538321A (en) * | 2020-03-31 | 2021-10-22 | 华为技术有限公司 | Vision-based volume measurement method and terminal equipment |
| CN111833317A (en) * | 2020-06-30 | 2020-10-27 | 佛山科学技术学院 | Industrial product specification detection method and equipment based on augmented reality |
| CN112697068A (en) * | 2020-12-11 | 2021-04-23 | 中国计量大学 | Method for measuring length of bubble of tubular level bubble |
| CN113177949B (en) * | 2021-04-16 | 2023-09-01 | 中南大学 | Large-size rock particle feature recognition method and device |
| CN113674006A (en) * | 2021-08-31 | 2021-11-19 | 广州侨益科技有限公司 | Agricultural product logistics traceability system based on visual calculation |
| CN116465888A (en) * | 2023-03-17 | 2023-07-21 | 南通锐越信息科技有限公司 | Agricultural product quality Internet of things monitoring method and device based on machine visual effect, electronic equipment and computer readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101907453A (en) * | 2010-07-23 | 2010-12-08 | 北京农业信息技术研究中心 | Online measurement method and device of dimensions of massive agricultural products based on machine vision |
| CN103278090A (en) * | 2013-05-14 | 2013-09-04 | 陕西科技大学 | Visual measurement method for volume of irregular object |
| CN104154877A (en) * | 2014-09-03 | 2014-11-19 | 中国人民解放军国防科学技术大学 | Three-dimensional reconstruction and size measurement method of complex convex-surface object |
| CN105675539A (en) * | 2016-01-07 | 2016-06-15 | 北京市农林科学院 | Comprehensive evaluation method of quality of agricultural products |
| CN206331487U (en) * | 2016-12-23 | 2017-07-14 | 蚌埠学院 | A kind of agricultural product volume rapid measurement device based on machine vision |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE112013003338B4 (en) * | 2012-07-02 | 2017-09-07 | Panasonic Intellectual Property Management Co., Ltd. | Size measuring device and size measuring method |
| EP3537343B1 (en) * | 2013-07-02 | 2020-10-21 | Owl Navigation Inc. | Method and system for a brain image pipeline and brain image region location and shape prediction |
-
2016
- 2016-12-23 CN CN201611202105.0A patent/CN106548480B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101907453A (en) * | 2010-07-23 | 2010-12-08 | 北京农业信息技术研究中心 | Online measurement method and device of dimensions of massive agricultural products based on machine vision |
| CN103278090A (en) * | 2013-05-14 | 2013-09-04 | 陕西科技大学 | Visual measurement method for volume of irregular object |
| CN104154877A (en) * | 2014-09-03 | 2014-11-19 | 中国人民解放军国防科学技术大学 | Three-dimensional reconstruction and size measurement method of complex convex-surface object |
| CN105675539A (en) * | 2016-01-07 | 2016-06-15 | 北京市农林科学院 | Comprehensive evaluation method of quality of agricultural products |
| CN206331487U (en) * | 2016-12-23 | 2017-07-14 | 蚌埠学院 | A kind of agricultural product volume rapid measurement device based on machine vision |
Non-Patent Citations (2)
| Title |
|---|
| P. Eisert et al..Multi-hypothesis, volumetric reconstruction of 3-D objects from multiple calibrated camera views.《1999 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258)》.2002,第3509-3512页. * |
| 袁雷明.基于多视成像及近红外光谱技术的巨峰葡萄品质无损检测研究.《中国博士学位论文全文数据库工程科技Ⅰ辑》.2016,(第8期),第B024-46页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106548480A (en) | 2017-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106548480B (en) | Quick agricultural product volume measuring device and measuring method based on machine vision | |
| Kamal et al. | Wood defects classification using laws texture energy measures and supervised learning approach | |
| CN109269951B (en) | Image-based flotation tailing ash content, concentration and coarse particle content detection method | |
| Pramunendar et al. | A classification method of coconut wood quality based on Gray Level Co-occurrence matrices | |
| CN104256882B (en) | Based on reconstituted tobacco ratio measuring method in the pipe tobacco of computer vision | |
| CN104198497B (en) | Surface defect detection method based on visual saliency map and support vector machine | |
| CN108287010B (en) | Crab multi-index grading device and method | |
| Alipasandi et al. | Classification of three varieties of peach fruit using artificial neural network assisted with image processing techniques. | |
| CN108985170A (en) | Transmission line of electricity hanger recognition methods based on Three image difference and deep learning | |
| CN207516257U (en) | A kind of wheat seed Image-capturing platform based on machine vision | |
| KR101743270B1 (en) | The method for separating and recognizing individual plankton using deep learning in a clumped or dispersed microscopic plankton image | |
| CN105651776A (en) | Device and method for automatically grading beef carcass meat yield based on computer vision | |
| CN108318494B (en) | The red online vision detection and classification devices and methods therefor for proposing fruit powder | |
| CN112184648A (en) | Piston surface defect detection method and system based on deep learning | |
| CN104198325B (en) | Stem ratio measuring method in pipe tobacco based on computer vision | |
| CN113688751A (en) | Method and device for analyzing alum blossom characteristics by using image recognition technology | |
| CN101510295B (en) | Design method for machine vision system based on PCIe and Vision Assistan | |
| CN206331487U (en) | A kind of agricultural product volume rapid measurement device based on machine vision | |
| CN109115775A (en) | A kind of betel nut level detection method based on machine vision | |
| CN110097540A (en) | The visible detection method and device of polygon workpeace | |
| CN112676195B (en) | Color sorting device and method for solid wood floor based on linear array CMOS camera | |
| CN113333329A (en) | Cigarette defect detection system based on deep learning | |
| CN109937912A (en) | A kind of egg categorizing system and method based on machine vision | |
| CN107679581B (en) | Method for processing gas flow distribution based on characteristic values of infrared image pixel matrix | |
| CN102680488A (en) | Device and method for identifying massive agricultural product on line on basis of PCA (Principal Component Analysis) |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |