CN111091520A - Precision detection method for sowing performance parameters of hybrid rice bowl body tray - Google Patents
Precision detection method for sowing performance parameters of hybrid rice bowl body tray Download PDFInfo
- Publication number
- CN111091520A CN111091520A CN201811243534.1A CN201811243534A CN111091520A CN 111091520 A CN111091520 A CN 111091520A CN 201811243534 A CN201811243534 A CN 201811243534A CN 111091520 A CN111091520 A CN 111091520A
- Authority
- CN
- China
- Prior art keywords
- image
- seeding
- seed
- hybrid rice
- grid
- 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.)
- Withdrawn
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 67
- 235000007164 Oryza sativa Nutrition 0.000 title claims abstract description 40
- 235000009566 rice Nutrition 0.000 title claims abstract description 40
- 238000009331 sowing Methods 0.000 title claims abstract description 17
- 240000007594 Oryza sativa Species 0.000 title description 2
- 238000010899 nucleation Methods 0.000 claims abstract description 73
- 238000000034 method Methods 0.000 claims abstract description 44
- 241000209094 Oryza Species 0.000 claims abstract description 39
- 238000012545 processing Methods 0.000 claims abstract description 31
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 15
- 238000007637 random forest analysis Methods 0.000 claims abstract description 10
- 230000000694 effects Effects 0.000 claims abstract description 9
- 239000011248 coating agent Substances 0.000 claims abstract description 8
- 238000000576 coating method Methods 0.000 claims abstract description 7
- 238000000605 extraction Methods 0.000 claims abstract description 4
- 238000001914 filtration Methods 0.000 claims description 20
- 238000006243 chemical reaction Methods 0.000 claims description 15
- 238000004519 manufacturing process Methods 0.000 claims description 15
- 230000007797 corrosion Effects 0.000 claims description 9
- 238000005260 corrosion Methods 0.000 claims description 9
- 239000002689 soil Substances 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000007405 data analysis Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 239000000853 adhesive Substances 0.000 claims description 3
- 230000001070 adhesive effect Effects 0.000 claims description 3
- 230000000844 anti-bacterial effect Effects 0.000 claims description 3
- 239000003899 bactericide agent Substances 0.000 claims description 3
- 239000003795 chemical substances by application Substances 0.000 claims description 3
- 239000003337 fertilizer Substances 0.000 claims description 3
- 239000002917 insecticide Substances 0.000 claims description 3
- 239000005648 plant growth regulator Substances 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 238000005520 cutting process Methods 0.000 claims description 2
- 239000011800 void material Substances 0.000 claims description 2
- 238000011160 research Methods 0.000 abstract description 6
- 238000005457 optimization Methods 0.000 abstract description 4
- 235000013339 cereals Nutrition 0.000 description 16
- 238000010586 diagram Methods 0.000 description 6
- 230000011218 segmentation Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000002187 spin decoupling employing ultra-broadband-inversion sequences generated via simulated annealing Methods 0.000 description 2
- 238000003892 spreading Methods 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 230000012010 growth Effects 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Pretreatment Of Seeds And Plants (AREA)
Abstract
The invention relates to a precision detection method for hybrid rice pot tray seeding performance parameters, which is used in an intelligent constant seeding detection system developed by an open source cross-platform vision library OPENCV and comprising a Linux operating system, and realizes the precision detection of the seeding performance parameters of the pot tray when the hybrid rice is raised and seeded. It includes: seed coating pretreatment, seedling tray sowing operation, seedling tray original image acquisition, seed image processing, pot body tray grid image processing, finally characteristic parameter extraction, random decision forest algorithm establishment and accurate detection of sowing performance parameters. The method realizes accurate detection of the seeding performance parameters of hybrid rice pot seedling raising, has good detection effect, high speed and high algorithm precision, and lays a research foundation for subsequent seeding quantity optimization regulation and control and constant seeding.
Description
Technical Field
The invention belongs to agricultural intelligent detection, and particularly relates to a precision detection method for hybrid rice bowl body disc seeding performance parameters.
Background
The planting area of hybrid rice in China accounts for 50 percent of the total planting area of rice, and is about 1500 ten thousand hectares. The hybrid rice has super tillering capability, so that the effective spike number can be increased, and according to the growth characteristic of the hybrid rice, low sowing quantity is generally required, and 2-3 grains/hole are ensured when hole sowing is used for precisely raising seedlings. Because the seeds need to be sown after pregermination before seedling raising operation, the length of the seed buds, the water content of the seeds and the size of the seeds can be changed inconsistently during sowing, and the sowing performance is influenced. Therefore, the seeding quantity of each hole of the pot body disc needs to be accurately detected in the seeding process, the change of the seeding state is timely found, and a basis is provided for the subsequent regulation and control of the seeding quantity or the reseeding work, so that the number of seeds in each hole of the seedling disc is basically kept consistent, and the requirement of intelligent constant-quantity precise seeding operation is met.
In the european and american countries, many scholars have conducted studies on the division and counting of adhered grains. ZHANG G et al, canada, divides the stuck grains by ellipse fitting and counts. The research on the partitioning and counting algorithms of the adhered grains is only limited to grains with low adhesion degree, and when the grains are overlapped or the adhesion degree is complex, the algorithm precision is not high. The domestic Qilong et al uses machine vision and virtual technology to realize the on-line detection of the hole, single grain and more than 3 conditions of double grain of the super rice seedling tray; in the aspect of image processing, the Wangxing et al detects the number of sowing grains in seedling holes of a seedling tray by using an area method and a watershed segmentation algorithm, and has higher detection efficiency on holes and single seeds; tanshiyan et al propose a method for identifying and classifying the seeding quantity of a super hybrid rice communicating region based on a mode classifier, select different characteristic value combinations and construct a BP neural network as the mode classifier of the seeding quantity of the rice communicating region, detect the number of grains with overlapped, crossed and adhered grains with low precision by aiming at the traditional gray average value, area method or ellipse fitting method, extract the multi-characteristic and analyze the characteristic preferentially for the communicating region, and identify and classify 6 conditions of broken rice/impurities, 1 grain, 2 grains, 3 grains, 4 grains and more than 5 grains in the communicating region, thereby realizing the accurate detection of the seeding quantity of the super hybrid rice. The Maxu et al propose an improved watershed segmentation algorithm and an improved SUSAN operator to realize segmentation and counting automatic detection of the adhered hybrid rice, and the improved SUSAN operator is based on the self-adaptive template circle radius and can effectively detect the corner points of the profile of the super hybrid rice communication area. The Dongwang et al combines the actual working condition of the efficient automatic seedling raising production line, and researches and develops an intelligent constant precision seeding device for raising seedlings by seedling trays according to the thought design of 'closed-loop feedback control', so that the seeding quantity is ensured in a constant range in real time, and the purpose of intelligent constant precision seeding is achieved.
At present, the seedling tray sowing performance detection technology mainly has a good detection effect on the hole and single seed sowing quantity of a seedling tray, but when the seeds are adhered and overlapped, the detection effect is poor, the number and the types of the detected seeds are simple, and when the sowing quantity is increased, the detection precision is sharply reduced; secondly, the detection device and the method are mostly used for blanket-shaped seedling trays (at present, the accurate detection of the seeding performance of the pot body tray has problems, the gray value of bed soil in the pot body tray is close to that of seeds, the binary image of the seeds cannot be well extracted when the image processing analysis is carried out, the serious reflection phenomenon exists when the image collection is carried out on the grids of the pot body tray, the seedling tray grids are not convenient to separate from the image), and the research on the accurate detection research aspect of the seeding amount of the pot body tray is fresh.
In summary, in order to realize the precise detection of the seeding rate during the seedling raising of the pot tray on the seedling raising and seeding production line, a precision detection method of the pot tray with high algorithm precision and good detection effect needs to be developed urgently, and a foundation is laid for the subsequent seeding rate control, optimization and constant seeding.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a precision detection method for the seeding performance parameters of hybrid rice pot trays, which realizes the precision detection of the seeding performance parameters of hybrid rice pot tray seedling raising, has high speed, high algorithm precision and good detection effect, and lays a research foundation for the subsequent optimal regulation and control of seeding quantity and constant seeding quantity.
The invention relates to a hybrid rice bowl body disc seeding performance parameter precision detection method, which is used in an intelligent constant seeding detection system developed by an open source cross-platform vision library OPENCV and comprising a Linux operating system; the accurate detection of the seeding performance parameters of the bowl body plate is realized when the hybrid rice is raised and seeded. The method specifically comprises (1) seed coating pretreatment; (2) seeding operation is carried out on a seedling tray; (3) collecting original images of seedling trays; (4) seed image processing, specifically to image color space conversion, seed image gray scale, binarization, expansion corrosion operation, and image connected region acquisition; (5) the pot body disc grid image processing specifically relates to image color space conversion, grid image gray scale and binarization, pixels of each row and each column on a binarization image are added by using a projection method, row and column pixels and peak points are calculated to obtain horizontal and vertical grid lines, and a detection area or a seedling hole is preferably positioned; (6) extracting characteristic parameters, establishing a random decision forest algorithm, and finally, accurately detecting the sowing performance parameters of the pot body disc.
The collection and analysis processing of the images of the seedling tray and the seeds are completed on an automatic seedling raising production line containing an intelligent constant seeding quantity detection system. The detection system is arranged behind a bed soil paving device and a precision seeding device on a seedling raising production line, and comprises an image acquisition module with a high-definition camera and an image processing and analyzing module. During working, the seedling tray after soil spreading and seeding passes through the intelligent constant seeding amount detection system, the high-definition camera carries out image acquisition operation, and the image processing and analyzing module processes the acquired image.
Step (4) in claim 1 above relates specifically to seed image processing: s1) acquiring the effective seed image, and performing color space conversion on the image. Converting the RGB image into HSV color space, setting S, V component threshold, performing data analysis and filtration on the collected original image, and separating the image of the seed from the original image; s2) carrying out gray scale and binarization on the image of the seed separated in the previous step, carrying out expansion corrosion form filtering operation processing, and filtering out interference points on the non-seed image to finally obtain a binarized image of the seed; s3) carrying out boundary scanning on the seed binary image to obtain a seed connected region.
Step (5) in claim 1 above relates specifically to the processing of the grid image of the bowl disk: s1) performing color space conversion on the original image of the pot seedling grid. Converting the collected RGB image into Lab color space, setting a threshold value only for the b component, carrying out data analysis and filtration on the original image, and separating out a grid image of the pot body disc seed holes; s2) carrying out gray scale, binaryzation and expansion corrosion form filtering operation processing on the pot body seedling grid image extracted in the previous step, filtering out interference points of a non-grid image, and finally obtaining a grid binaryzation image; s3) adding pixels of each row and each column on the binary image by using a projection method, calculating the pixels of the row and the column and peak points to obtain transverse and longitudinal grid lines, and preferably positioning a detection area and a seedling hole.
The step (6) in the claim 1 is specifically related to extracting characteristic parameters, establishing a random decision forest algorithm, and finally, accurately detecting and analyzing the seeding effect. S1) selecting the preferred location detection area and seedling pit according to claim 4 (S3), i.e. the connected area image of the cuttable seed; s2) obtaining the characteristic parameters of the connected region in each sliced slice image; s3) predicting the characteristic parameters according to the random forest tree established during sample learning to obtain the seed number in each slice image; s4), the number of seeds in each slice, the total amount of seeds in all slices and the void ratio thereof can be analyzed statistically.
The method (S3) of claim 5, wherein the method specifically includes the step of performing feature extraction on the parameters of the obtained seed connected region, such as area, perimeter, invariant moment and the like. The invariant moment mainly represents the geometric characteristics of an image area and has invariance of rotation, translation and scale. Invariant moments can be used as an important feature to represent the object, and the moment based on the area is less likely to be misclassified by using the moment ratio of the boundary sequence, so that the calculation amount is reduced.
The method (S3) of claim 4, wherein the detecting region and the seedling hole are preferably located by projection. Wherein the projection method comprises the following steps: and respectively projecting the binary image of the grid of the pot body disc along the horizontal direction and the vertical direction, wherein a series of wave crests and wave troughs are formed in the vertical direction and the horizontal direction and respectively represent the number of pixels of the grid of the pot body disc in corresponding rows and columns.
In order to match with a precision detection system for the seeding quantity during the pot body tray seedling raising on an automatic seedling raising production line, the hybrid rice seeds need to be subjected to seed coating pretreatment. Specifically, the adhesive or the film-forming agent is used for wrapping the components such as the bactericide, the insecticide, the micro-fertilizer, the plant growth regulator and the like outside the seeds.
Compared with the prior art, the invention has the following beneficial effects:
(1) and after the effective seed image is acquired, performing color space conversion on the image. Converting the RGB image into HSV color space, setting S, V component threshold, analyzing and filtering the collected original image, and separating the image of the seed from the original image;
(2) and carrying out color space conversion on the original image of the pot seedling grid. Converting the collected RGB image into Lab color space, setting a threshold value only aiming at the b component, carrying out data analysis and filtration on the original image, and better separating out a grid image of the pot body disc seed holes;
(3) and adding pixels of each row and each column on the binary image of the grid of the pot body disc by using a projection method to obtain pixels of the row and the column and peak points, obtaining transverse and longitudinal grid lines, and preferably positioning a detection area and seedling holes. The detection method can realize the precise detection of the seeding amount when the pot body tray is used for raising rice seedlings on a rice seedling raising and seeding production line, has high algorithm precision and good detection effect, and lays a foundation for subsequent seeding amount control, optimization and constant seeding.
Drawings
FIG. 1 is a working principle diagram of a hybrid rice bowl body disc seeding performance parameter precision detection method.
FIG. 2 is a schematic diagram of a sample learning process after seed image acquisition according to the present invention.
FIG. 3 is an original image of a bowl plate containing seeds collected according to the present invention.
FIG. 4 is a seed image after color space conversion according to the present invention.
FIG. 5 is a seed binarization image of the present invention.
FIG. 6 is a diagram illustrating the expansion erosion operation performed on the seed binarized image according to the present invention.
FIG. 7 is a color space converted grid image of a bowl disk of the present invention.
FIG. 8 is a binary image of the grid of the inventive pot disc.
FIG. 9 is a row peak diagram of the binary image of the grid of the potted disk according to the present invention after summing the pixels of each row.
FIG. 10 is a column peak diagram after summation of each column of pixels of the binary image of the grid of the bowl body disk.
FIG. 11 is a schematic diagram of a seed connected region image cut according to grid lines.
FIG. 12 is an image of a cut seed connected region cell of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific examples described herein are intended to be illustrative only and are not intended to be limiting.
The collection, analysis and processing of the images of the seedling tray and the seeds are completed on an automatic seedling raising production line containing an intelligent constant seeding quantity detection system. The detection system is arranged behind a bed soil paving device and a precision seeding device on a seedling raising production line, and comprises an image acquisition module with a high-definition camera and an image processing and analyzing module. During working, the seedling tray after soil spreading and seeding passes through the intelligent constant seeding amount detection system, the high-definition camera carries out image acquisition operation, and the image processing and analyzing module processes the acquired image. The precision seeding device containing the intelligent constant seeding amount detection system is preferably a 2SJB-500 type rice seedling raising precision seeding production line developed and designed by southern China agricultural university, and the seedling tray adopts the most common rice pot seedling soft plastic seedling tray (composite tray containing embedded soft plastic seedling tray) with the length, width, height and hole number of 406 holes, wherein the length, width and height of the rice pot seedling soft plastic seedling tray are 585mm, 285mm and 22 mm.
Before image acquisition, the height of a high-definition camera is adjusted, the working speed of a production line is adjusted to be 500 disks/h, namely 7.2 s/disk, limiting plates on two sides of a seedling disk on the production line are adjusted, when the seedling disk reaches a shooting area, a computer controls a camera to shoot images every 3s, each seedling disk acquires 2 images, and no overlapped shooting area exists between the images. In addition, the spatial light intensity uniformity of the detection system needs to be adjusted, and the selection of the hybrid rice with uniform color is beneficial to extracting a complete-shaped communication area, so that the detection accuracy is improved.
Preferably, in order to match with a precision detection system for the seeding quantity during pot plate seedling raising on an automatic seedling raising production line, the hybrid rice seeds need to be subjected to seed coating pretreatment. Specifically, the adhesive or the film-forming agent is used for wrapping the components such as the bactericide, the insecticide, the micro-fertilizer, the plant growth regulator and the like outside the seeds. In order to make the color of the coated seeds obviously different from the color of the bed soil, the red seed coating agent is used to treat the seeds in the specific embodiment.
With reference to fig. 1 and 2, the method is applied to an intelligent constant seeding detection system developed by an open-source cross-platform vision library OPENCV and comprising a Linux operating system; the accurate detection of the seeding performance parameters of the pot body plate is realized when the seedling is raised and the seeding is carried out. The method specifically comprises (1) seed coating pretreatment; (2) seeding operation is carried out on a seedling tray; (3) collecting original images of seedling trays; (4) seed image processing, specifically to image color space conversion, seed image gray scale, binarization, expansion corrosion operation, and image connected region acquisition; (5) the pot body disc grid image processing specifically relates to image color space conversion, grid image gray scale and binarization, pixels of each row and each column on a binarization image are added by using a projection method, row and column pixels and peak points are calculated to obtain horizontal and vertical grid lines, and a detection area or a seedling hole is preferably positioned; (6) extracting characteristic parameters, establishing a random decision forest algorithm, and finally, accurately detecting the sowing performance parameters of the pot body disc.
The seed image processing process is shown in connection with fig. 2-6. S1) acquiring the effective seed image, and performing color space conversion on the image. Converting the RGB image into HSV color space, setting S, V component threshold, performing data analysis and filtration on the collected original image, and separating the image of the seed from the original image; s2) carrying out gray scale and binarization on the image of the seed separated in the previous step, carrying out expansion corrosion form filtering operation processing, and filtering out interference points on the non-seed image to finally obtain a binarized image of the seed; s3) carrying out boundary scanning on the seed binary image to obtain a seed connected region.
The bowl disk grid image processing is shown in connection with fig. 7-10. S1) performing color space conversion on the original image of the pot seedling grid. Converting the collected RGB image into Lab color space, setting a threshold value only for the b component, carrying out data analysis and filtration on the original image, and separating out a grid image of the pot body disc seed holes; s2) carrying out gray scale, binaryzation and expansion corrosion form filtering operation processing on the pot body seedling grid image extracted in the previous step, filtering out interference points of a non-grid image, and finally obtaining a grid binaryzation image; s3) adding pixels of each row and each column on the binary image by using a projection method, calculating the pixels of the row and the column and peak points to obtain transverse and longitudinal grid lines, and preferably positioning a detection area and a seedling hole.
Preferably, the present embodiment selects 9 holes × 13 holes in the shooting area as the detection area. The projection method is to project the binary image of the grid of the pot body disc along the horizontal direction and the vertical direction respectively, wherein the vertical direction and the horizontal direction both have a series of wave crests and wave troughs which respectively represent the pixel number of the grid of the pot body disc in corresponding rows and columns.
And (3) combining the processes shown in the figures 11 and 12, extracting characteristic parameters, establishing a random decision forest algorithm, and finally accurately detecting and analyzing the sowing effect. S1) according to the above mentioned optimization, positioning the detection area or seedling hole, that is, cutting the connected area image of the seed; preferably, the seed image of the complete mesh portion is retained and the marginal incomplete mesh portion is discarded and left unprocessed. S2) obtaining the characteristic parameters of the connected region in each sliced slice image; preferably, the method specifically relates to feature extraction of parameters such as the area, the perimeter, the invariant moment and the like of the acquired seed connected region. The invariant moment mainly represents the geometric characteristics of an image area and has invariance of rotation, translation and scale. Invariant moments can be used as an important feature to represent the object, and the moment based on the area is less likely to be misclassified by using the moment ratio of the boundary sequence, so that the calculation amount is reduced. The moment invariant data of the present invention can be image data of an original image, reduced by half, horizontally inverted, and rotated by 45 °. S3) predicting the characteristic parameters according to the random forest tree established during sample learning to obtain the seed number in each slice image; s4) specifically, the seeding performance parameters such as the number of seeds in each slice, the total amount of seeds of all slices, the hole rate of the total amount of seeds and the like can be statistically analyzed; by analyzing the images of the seeds in the plurality of sowed bowl plates, the overall sowing performance parameters can be obtained.
The present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents and are included in the scope of the present invention.
Claims (8)
1. A precision detection method for hybrid rice pot plate seeding performance parameters is used in an intelligent constant seeding detection system developed by a Linux operating system and an open-source cross-platform vision library OPENCV, and realizes the precision detection of the seeding performance parameters of the pot plate when hybrid rice is raised and seeded. Characterized in that the method comprises the following steps:
(1) pre-treating the seed coating;
(2) seeding operation is carried out on a seedling tray;
(3) collecting original images of the seedling tray, including collecting seed images and collecting grid images of the pot body tray;
(4) processing the collected seed image, specifically relating to image color space conversion, seed image gray scale and binarization, expansion corrosion operation and image connected region acquisition;
(5) processing the collected grid image of the pot body disc, in particular to image color space conversion, grid image gray scale and binarization, adding pixels of each row and each column on the binarization image by using a projection method, calculating row and column pixels and peak points to obtain transverse and longitudinal grid lines and preferably positioning a detection area or a seedling hole;
(6) and extracting characteristic parameters and establishing a random decision forest algorithm, and finally, accurately detecting the sowing performance parameters of the hybrid rice bowl body plate.
2. The method for precisely detecting the seeding performance parameters of the hybrid rice bowl body disc as claimed in claim 1, wherein the method comprises the following steps: the collection and analysis processing of the images of the seedling tray and the seeds are completed on an automatic seedling raising production line containing an intelligent constant seeding amount detection system, the detection system is installed behind a bed soil paving device and a precise seeding device on a seedling raising production line, the detection system comprises an image collection module containing a high-definition camera and an image processing and analysis module, and when the seedling tray after soil paving and seeding passes through the intelligent constant seeding amount detection system, the high-definition camera carries out image collection operation and transmits the image collection operation to the image processing and analysis module to process the collected images.
3. The method for precisely detecting the seeding performance parameters of hybrid rice bowl plates as claimed in claims 1 and 2, wherein: step (4) in claim 1 above relates specifically to seed image processing:
s1) acquiring the effective seed image, and performing color space conversion on the image. Converting the RGB image into HSV color space, setting S, V component threshold, performing data analysis and filtration on the collected original image, and separating the image of the seed from the original image;
s2) carrying out gray scale and binarization on the image of the seed separated in the previous step, carrying out expansion corrosion form filtering operation processing, and filtering out interference points on the non-seed image to finally obtain a binarized image of the seed;
s3) carrying out boundary scanning on the seed binary image to obtain a seed connected region.
4. The method for precisely detecting the seeding performance parameters of the hybrid rice bowl body plate according to claim 1 or 2, which is characterized in that: step (5) in claim 1 above relates specifically to the processing of the grid image of the bowl disk: s1) carrying out color space conversion on the original grid image of the pot seedlings, converting the collected RGB image into Lab color space, setting a threshold value only for the component b, carrying out data analysis and filtration on the original image, and separating out grid images of the pot seedling tray holes;
s2) carrying out gray scale and binaryzation on the pot seedling grid image extracted in the previous step, then carrying out expansion corrosion form filtering operation processing, filtering out interference points of a non-grid image, and finally obtaining a grid binaryzation image;
s3) adding pixels of each row and each column on the binary image by using a projection method, calculating the pixels of the rows and the columns and peak points to obtain horizontal and vertical grid lines, and preferably positioning a detection area and a seedling hole.
5. The method for precisely detecting the seeding performance parameters of hybrid rice bowl body discs as claimed in claim 4, characterized in that: the method of claim 1, wherein the steps of (6) extracting characteristic parameters and establishing a random decision forest algorithm, and finally, the accurate detection and analysis of the seeding effect specifically comprise:
s1) cutting out a communicated region image of the seeds according to the preferred positioning detection region and the seedling hole;
s2) obtaining the characteristic parameters of the connected region in each sliced slice image;
s3) predicting the characteristic parameters according to the random forest tree established during sample learning to obtain the seed number in each slice image;
s4), the number of seeds in each slice, the total amount of seeds in all slices and the void ratio thereof can be analyzed statistically.
6. The method for precisely detecting the seeding performance parameters of hybrid rice bowl plates as claimed in claim 5, wherein the method comprises the following steps: the method specifically comprises the steps of carrying out feature extraction on parameters such as the area, the perimeter and the invariant moment of the connected region, wherein the invariant moment is used for representing the geometric features of the image region and has invariance of rotation, translation and scale, the invariant moment can be used as an important feature to represent an object, and the error classification with lower probability can be obtained by using the moment ratio of the boundary sequence based on the moment of the area, so that the calculated amount is reduced.
7. The method for precisely detecting the seeding performance parameters of hybrid rice bowl plates as claimed in claim 4, wherein the method comprises the following steps: the projection method is to project the binary image of the grid of the pot body disc along the horizontal direction and the vertical direction respectively, wherein the vertical direction and the horizontal direction both have a series of wave crests and wave troughs which respectively represent the number of the pixels of the grid of the pot body disc in corresponding rows and columns.
8. The method for precisely detecting the seeding performance parameters of the hybrid rice bowl body disc as claimed in claim 1, wherein the method comprises the following steps: in order to match with a precision detection system for the seeding quantity during the pot body tray seedling raising on an automatic seedling raising production line, the hybrid rice seeds need to be subjected to seed coating pretreatment. Specifically, the adhesive or the film-forming agent is used for wrapping the components such as the bactericide, the insecticide, the micro-fertilizer, the plant growth regulator and the like outside the seeds.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811243534.1A CN111091520A (en) | 2018-10-24 | 2018-10-24 | Precision detection method for sowing performance parameters of hybrid rice bowl body tray |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811243534.1A CN111091520A (en) | 2018-10-24 | 2018-10-24 | Precision detection method for sowing performance parameters of hybrid rice bowl body tray |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111091520A true CN111091520A (en) | 2020-05-01 |
Family
ID=70392718
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811243534.1A Withdrawn CN111091520A (en) | 2018-10-24 | 2018-10-24 | Precision detection method for sowing performance parameters of hybrid rice bowl body tray |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111091520A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112102258A (en) * | 2020-08-28 | 2020-12-18 | 无锡卡尔曼导航技术有限公司 | Air-suction type seeder seeding detection method based on machine vision |
-
2018
- 2018-10-24 CN CN201811243534.1A patent/CN111091520A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112102258A (en) * | 2020-08-28 | 2020-12-18 | 无锡卡尔曼导航技术有限公司 | Air-suction type seeder seeding detection method based on machine vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108037123B (en) | Precision detection method for sowing performance parameters of hybrid rice bowl body tray | |
| Sun et al. | Image processing algorithms for infield single cotton boll counting and yield prediction | |
| CN109447945B (en) | Quick counting method for basic wheat seedlings based on machine vision and graphic processing | |
| Tong et al. | Machine vision techniques for the evaluation of seedling quality based on leaf area | |
| CN109255757B (en) | Method for segmenting fruit stem region of grape bunch naturally placed by machine vision | |
| Jiang et al. | Detection of maize drought based on texture and morphological features | |
| CN105095958A (en) | Cocoon counting method | |
| CN105095957A (en) | Silkworm cocoon counting method based on image segmentation | |
| CN108323389B (en) | The rice transplanting rice shoot spacing in the rows of rice transplanter and the detection method and device of cave rice shoot number | |
| CN112614147B (en) | Crop seedling stage plant density estimation method and system based on RGB image | |
| Selvi et al. | Weed detection in agricultural fields using deep learning process | |
| CHANG et al. | Quantifying muskmelon fruit attributes with A-TEP-based model and machine vision measurement | |
| CN103632157A (en) | A method for counting seeds of a wheatear portion per wheat | |
| CN114818909A (en) | Weed detection method and device based on crop growth characteristics | |
| CN114694047A (en) | Corn sowing quality evaluation method and device | |
| CN107464232B (en) | Image detection method for planting quality of unmanned rice transplanter | |
| CN112580671A (en) | Automatic detection method and system for multiple development stages of rice ears based on deep learning | |
| Li et al. | Design and experiment of intelligent sorting and transplanting system for healthy vegetable seedlings | |
| CN111091520A (en) | Precision detection method for sowing performance parameters of hybrid rice bowl body tray | |
| CN116897668B (en) | Electric-drive crop sowing and fertilizing control method and system | |
| Zhang et al. | Application of computer vision technology in agricultural field | |
| CN111886982B (en) | Detection method of dry land planting operation quality real-time detection system | |
| CN115661647A (en) | Timed inspection method for tree planting growth conditions | |
| Zhang et al. | Online Recognition of Small Vegetable Seed Sowing Based on Machine Vision | |
| Yihang et al. | Automatic recognition of rape seeding emergence stage based on computer vision technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| DD01 | Delivery of document by public notice | ||
| DD01 | Delivery of document by public notice |
Addressee: Yang Chenxi Document name: Notice of Preliminary Confirmation Opinion on Abnormal Patent Application |
|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200501 |