CN102564593A - Plant growth condition monitoring system based on compute vision and internet of things - Google Patents

Plant growth condition monitoring system based on compute vision and internet of things Download PDF

Info

Publication number
CN102564593A
CN102564593A CN2011104517988A CN201110451798A CN102564593A CN 102564593 A CN102564593 A CN 102564593A CN 2011104517988 A CN2011104517988 A CN 2011104517988A CN 201110451798 A CN201110451798 A CN 201110451798A CN 102564593 A CN102564593 A CN 102564593A
Authority
CN
China
Prior art keywords
blade
image
internet
things
plant
Prior art date
Application number
CN2011104517988A
Other languages
Chinese (zh)
Other versions
CN102564593B (en
Inventor
李庆武
彭文
马国翠
曹晔锋
霍冠英
周妍
黄河
Original Assignee
河海大学常州校区
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 河海大学常州校区 filed Critical 河海大学常州校区
Priority to CN 201110451798 priority Critical patent/CN102564593B/en
Publication of CN102564593A publication Critical patent/CN102564593A/en
Application granted granted Critical
Publication of CN102564593B publication Critical patent/CN102564593B/en

Links

Abstract

The invention discloses a plant growth condition monitoring system based on compute vision and internet of things. The plant growth condition monitoring system is characterized by comprising a high speed digital signal processor (DSP) computer vision module for shooting plant leaves and acquiring and analyzing the shot images, a control execution module for watering plants, and an internet of things module for transmitting the plant leaf information to a remote terminal through a remote server and providing the information for workers to consult and operate, wherein the high speed DSP computer vision module, the control execution module and the internet of things module are connected through data. The traditional detection system is improved, a computer vision-based DSP image processing technology and an improved processing algorithm are used for acquiring and processing leaf images of luxurious flowers and trees, the image acquiring and processing speed and accuracy are improved, the automation level of the system is improved and the production cost is reduced; and the acquired data is timely transmitted to the workers at the terminal through the remote server by the internet of things technology.

Description

Vegetation growth state monitoring system based on computer vision and Internet of Things

Technical field

The present invention relates to a kind of based on computer vision and Internet of Things the vegetation growth state monitoring system and utilize this system to carry out the method that the famous flower and tree blade detects and realize, belong to the application of digital image processing techniques at agriculture detection range.

Background technology

The blade of plant is the important photosynthetic organ of plant, also is the main path that plant carries out transpiration.Photosynthesis, i.e. luminous energy synthesis, the blade that is plant utilizes photosynthetic pigments under visible light radiation, carbon dioxide (or sulfuretted hydrogen) and water are transformed into organism, and discharges the biochemical process of oxygen (or hydrogen).Photosynthesis is the summation of the metabolic response of a series of complicacies, is the basis that whole organic sphere is depended on for existence, also is the important media of earth carbon oxygen cycle.

We are not difficult to find, in today that precision agriculture constantly develops, carry out the precision agriculture practice and not only can make rational use of resources, improve irrigation quality, and can reduce production costs, protect environment, raising crop competitive power.Meanwhile, the various parameters of research plant leaf blade are for growth and development of plant, irrigable culture, and crop yield etc. all has crucial meaning.Make up real-time, quick, effective plant leaf blade analytical approach,, all have important practical significance thereby instruct the arable farming density and the rational application of fertilizer to irrigate for adjusting group structure, making full use of the photo-thermal resource.

Traditional vegetation growth state monitoring method too relies on manually-operated, needs flowers and trees cultivation personnel deeply to plant booth inside and carries out information acquisition and fertigation, so monitoring inefficiency and cost are higher.Along with the continuous development of computer technology in recent years, the continuous maturation of digital image processing techniques can consider to obtain and analyze by computer vision the external appearance characteristic of flowers and trees blade.Computer vision is to use a kind of simulation to biological vision of computing machine and relevant device, and its main task obtains corresponding information through gathering picture or video exactly, thereby it is analyzed and handles.At present; With digital image processing techniques is that the based computer vision detection technology has been applied to the various aspects in the social production; Like recognition of face, vehicle identification, industrial detection and medical image analysis or the like; Its distinctive unartificial property, accuracy and stability, and to the adaptability of environmental change, be that other traditional detection methods are difficult to match in excellence or beauty.

Can the quality of image processing algorithm be the computer vision technique key in application of succeeing.Compare with otherwise application, the color of flowers and trees blade to be detected, texture, shape and size all are its important external appearance characteristics.But the growing environment of plant causes image-forming condition undesirable greatly because of illumination variation, and blade can not have regular as industrial part in the outward appearance performance and can be descriptive simultaneously, so brought great difficulty for the identification of objectives.Since the last century the nineties; A lot of scholars are utilizing computer vision technique plant leaf blade to be carried out launched on the direction of check and analysis research: [Ji Shouwen; Wang Rongben, Chen Jiajuan, etc. the research [J] of appliance computer image processing techniques identification corn weeds in field in seedling stage. EI; 2001; 17 (2): 201-204] utilize image processing techniques to obtain shape description parameters such as projected area of blade, leaf length, Ye Kuan in the literary composition, the corn and the weeds in seedling stage are discerned, thereby effective foundation is provided for accurate spraying herbicide; [Mao Hanping; Xu Guili, Li Pingping, etc. the identification [J] that wanes based on the tomato nutrient element of computer vision. EI; 2003; 34 (2): 73-75] in the literary composition by the carry out feature extraction of computer vision to tomato leaf, set up the framework that the tomato nutritional deficiency symptom is carried out pattern-recognition through the binary tree sort method, thereby tomato nitrogen stress and potassium deficiency initial stage blade that naked eyes be difficult for to be differentiated discerned; [Sin N.; Casady W.W.; Costello T.A.Machine-vision-based Nitrogen management models for rice [J] .Transaction of the ASAE.1996; 39 (5): 1899-1904] use image segmentation to obtain the morphological feature of paddy rice in the literary composition, and combine, invented a kind of method of utilizing morphological feature such as blade area to judge the N nutritional deficiency situation of plant with the vegetation growth state analysis.

Internet of Things is considered to the information technology revolution again after computing machine, Internet technology as an emerging in recent years infotech.Technology of Internet of things makes people can understand the information of all environment, equipment, facility in the reality whenever and wherever possible through by computer internet, sensor and terminal device, has broken away from the restriction of time and region.Internet of Things is of many uses, has spreaded all over numerous areas such as intelligent transportation, environmental protection, public safety, personal health, arable farming and industry monitoring at present.The widespread use of technology of Internet of things in agricultural is expected to bring for agricultural the change of epoch-making significance: people not only can understand growing environment and the growing way situation of crop through online, and can make judgement with the disease and pest situation to the arid situation of crop and also in time take appropriate measures.In the famous flower and tree growth monitoring that the agricultural technology of Internet of things is changed in modern times huge development prospect is arranged also.As everyone knows, upgrowth situation of plant and illumination condition, environment temperature and factors such as humidity, soil moisture content are all closely related.All the time, the flowers and trees industry can only depend on the experience management owing to lack accurate measurement means.This extensive way to manage has greatly restricted the raising of flowers and trees growth qualities, uses agriculture technology of Internet of things then can improve the efficient of famous flower and tree industry in producing and managing significantly.

China is in that to utilize computer vision to carry out the technology ability that plant detects aspect analysis at the early-stage.The flow process of general computer vision system detection system is: adopt CCD camera collection image; Subsequently it is sent into and carry out the A/D conversion in the data collecting card; Call corresponding application by operating system behind the data entering computing machine thereby these data are handled the extraction characteristic, accomplish detection task target.But because the data transmission of this type systematic needs the clock period of labor; And general computing machine does not adopt the hardware configuration of specialty data are handled to be optimized; Cause algorithm consuming time more, greatly reduced the real-time and the detection efficiency of this system.Therefore need to adopt special high-speed dsp chip to come substituting for computer to carry out data processing to Flame Image Process.Optimization process algorithm how, seeking simple detection method fast is to improve the key of system availability.

Summary of the invention

The present invention is directed to the demand that flowers and trees leaf growth situation detects automatically, proposed a kind of vegetation growth state monitoring system based on computer vision and Internet of Things.This system's utilization is carried out IMAQ and processing to the high-speed dsp computer vision module of digital picture; High-speed dsp chip in the module is as a kind of special microprocessor; Its distinctive Harvard structure and stream line operation can make it that treatment of picture arithmetic speed is far surpassed general micro controller, have improved the real-time that detects greatly; On the other hand; Through technology of Internet of things is combined with the blade detection technique; Not only can realize the real-time monitoring of famous flower and tree upgrowth situation and irrigation in time; More can relevant information be conveyed to the staff that flowers and trees are cultivated enterprise through remote server and terminal device, thereby reach the purpose of Remote, make things convenient for them in time the famous flower and tree in the booth to be irrigated.

Technical scheme of the present invention provides a kind of vegetation growth state monitoring system based on computer vision and Internet of Things; It is characterized in that: the control and executive module that it comprises high-speed dsp computer vision module that plant leaf blade is taken and taken image is gathered and analyzed, plant is watered; And send plant leaf blade information to Internet of Things module that remote terminal supplies staff's reference and operation through remote server, data connection between said high-speed dsp computer vision module, control and executive module and the Internet of Things module.

Preferably, said high-speed dsp computer vision module comprises the industrial camera that is positioned at directly over the plant to be monitored, is used to the DSP image processing module that obtains the ccd image acquisition module of the taken image of said industrial camera and include the high-speed dsp chip TMS320C6000 that is used for Flame Image Process.

Preferably, said industrial camera is wrapped with one deck waterproof membrane all over the body, and said industrial camera is other to be provided with lighting source, and said lighting source is a led light source.

Preferably, said ccd image acquisition module and DSP image processing module all are loaded on the portable dsp board card of encapsulation type.

Preferably; Said control and executive module comprises: a plurality of sprinkling irrigation shower nozzles that are positioned at plant to be monitored top and the aqueduct that supplies water for said sprinkling irrigation shower nozzle; Said sprinkling irrigation shower nozzle is controlled by strong power controller, and said strong power controller all is connected industrial computer with said high-speed dsp computer vision module, Internet of Things module.

Preferably, the step of its work is:

1) said industrial camera is regularly taken the coloured image of the blade of plant to be monitored;

2) said ccd image acquisition module sends said DSP image processing module to after obtaining said coloured image;

3) said DSP image processing module detects the plant leaf blade in the said coloured image, if find plant hydropenia then notify said control and executive module that plant is watered.

Preferably, in the said step 3), said DSP image processing module carries out in the following detection one or more to the plant leaf blade in the said coloured image: leaf color detects, the blade texture detects and blade shape detects.

Preferably, the step of said leaf color detection is:

1) plant leaf blade in the said coloured image being removed petiole handles;

2) with said coloured image by the RGB color space conversion to the hsv color space, set H component and S component variation scope and the two divided equally;

3) create the two-dimensional histogram that each dimension is divided equally according to the data statistics on H plane and S plane;

4) utilize former said coloured image said two-dimensional histogram to be converted to the color 2 D histogram of the rgb space of blade;

5) setting threshold through the flowers and trees blade in the said color 2 D histogram is analyzed, determines plant and whether belongs to exsiccosis.

Preferably, the step of said blade texture detection is:

1) with said coloured image gray processing and adopt mean filter;

2) set dynamic threshold to different types of flowers and trees blade, the edge in algorithm detection of utilization dual threshold and the connection layout picture;

3) the texture quantity of blade unit area is added up, judge through the density degree of texture whether plant belongs to exsiccosis.

Preferably, said blade shape detects and comprises that blade shape parameter detecting and blade area detect, and the step of wherein said blade shape parameter detecting is:

1) plant leaf blade in the said coloured image being removed petiole handles;

2), and extract the profile of plant leaf blade with the gray level image binaryzation of the plant leaf blade in the said coloured image;

3) mean value of the pixel coordinate of the profile through calculate extracting plant leaf blade obtains blade centre of form point coordinate, and is that maximum circumradius of blade and minimum inscribed circle radius are asked in the center of circle with the centroid point, with the two ratio as the blade shape parameter;

4) compare through the blade shape parameter that calculates and the blade shape parameter under the normal condition, judge whether plant belongs to exsiccosis.

Preferably, the step of said blade area detection is:

1) with the gray level image binaryzation of the plant leaf blade in the said coloured image, obtains binary image;

2) all gray-scale values of the said binary image of traversal statistics are the sum of all pixels of 0, with its blade area as plant leaf blade;

3) historical variations of the blade area of this plant leaf blade relatively judges that if blade area diminishes this plant leaf blade is for belonging to exsiccosis.

Preferably, wherein the plant leaf blade in the said coloured image being removed the petiole processed steps comprises:

1) at first will comprise the leaf image binaryzation of petiole, and with its rotate to be blade upwards, petiole position down, to obtain bianry image;

2) said bianry image is carried out transversal scanning from bottom to top, to confirm the breadth extreme of petiole part;

3) through to the continuously transversal scanning of black pixel point and number statistics in the bianry image, with less than step 2) in the part of definite petiole part breadth extreme be petiole, thereby determine the definite position of petiole;

4) with the petiole image inverse that obtains, itself and former said coloured image are carried out or operate, obtained rejecting the blade coloured image of petiole.

The present invention is scientific and reasonable; Taken into full account the real-time that detects; And traditional detection system improved; Use the leaf image that obtains and handle famous flower and tree based on the DSP image processing techniques and the Processing Algorithm after the improvement of computer vision, improved the speed and the accuracy rate of Image Acquisition and processing.After detecting the plant that need water, can utilize industrial computer to control the sprinkling irrigation shower nozzle and accomplish corresponding operation, improved the automatization level of system, reduced production cost.In addition, technology of Internet of things has been introduced the analysis monitoring system of plant leaf blade.Send the data of collecting to the terminal works personnel in time through remote server, broken away from the restriction on traditional flowers and trees breeding method time and the region, be very easy to the remote monitoring of staff flowers and trees.When the staff need water to the flowers and trees in the booth, only need in office, send instruction and get final product through PC or smart mobile phone Remote industrial computer.Should quite good detecting real-time and stronger stability, accuracy be arranged based on the vegetation growth state monitoring system and the method for computer vision and Internet of Things, also support Remote simultaneously, cultivate enterprise's use so be fit to very much famous flower and tree.

Description of drawings

Fig. 1 is the system architecture synoptic diagram of the vegetation growth state monitoring system based on computer vision and Internet of Things of the present invention;

Fig. 2 is the system algorithm schematic flow sheet of the vegetation growth state monitoring system based on computer vision and Internet of Things of the present invention;

Among Fig. 1: 1.LED light source (100w), 2. industrial camera, 3. high-speed dsp chip TMS320C6000,4. aqueduct, 5. sprinkling irrigation shower nozzle, 6. strong power controller, 7. industrial computer, 8. remote server, 9. PC, 10. smart mobile phone.

Embodiment

Following specific embodiments of the invention is described in further detail.

As depicted in figs. 1 and 2; The hardware components of a kind of vegetation growth state monitoring system based on computer vision and Internet of Things of the present invention comprises: the high-speed dsp computer vision module that the flowers and trees blade is taken and taken image is gathered and analyzed; The control and executive module that the lack of water flowers and trees are watered, and send blade information to three parts of Internet of Things module that the terminal supplies staff's reference and operation through remote server:

One, high-speed dsp computer vision module comprises industrial camera, special-purpose ccd image acquisition module and DSP image processing module.Industrial camera is positioned at about 1 meter directly over the flowers and trees to be detected, and the camera next door is furnished with the led light source of 100w as illumination, and camera is enclosed with one deck waterproof membrane all over the body.Special-purpose ccd image acquisition module is used to obtain by the taken image of industrial camera; Then contain the high-speed dsp chip TMS320C6000 that is specifically designed to Flame Image Process in the DSP image processing module; The two all is loaded on the dsp board card, and integrated circuit board is that encapsulation type is portable.

Two, control and executive module comprises industrial computer, strong power controller, aqueduct and sprinkling irrigation shower nozzle.Industrial computer is placed on flowers and trees and cultivates in the booth, directly links to each other with high-speed dsp computer vision module through pci interface, and controls the sprinkling irrigation shower nozzle through strong power controller and water.Aqueduct is made by resistant material, is positioned at about 0.2 meter, dead ahead of plant.The sprinkling irrigation shower nozzle is placed on the aqueduct, can be provided with one at a certain distance according to actual conditions, and main being responsible for sprayed to the flowers and trees of lack of water according to the signal that industrial computer transmits.

Three, the Internet of Things module comprises terminal devices such as remote server, PC and smart mobile phone.Remote server is arranged on flowers and trees and cultivates in enterprise's office block; It is the bridge block between flowers and trees monitoring point and the client terminal; The flowers and trees blade information that the interior industrial computer of booth transmits is collected and stored to main being responsible for, and is communicated with member in the net, and to terminal device data, services is provided at any time.PC and smart mobile phone belong to terminal device; Main being responsible for is shown to the staff that flowers and trees are cultivated enterprise with the plant information that server sends; Industrial computer in cultivating booth sends instruction in real time simultaneously, makes things convenient for them that the upgrowth situation of famous flower and tree is carried out remote monitoring.

Vegetation growth state monitoring system based on computer vision and Internet of Things of the present invention is following to the concrete performing step that blade detects:

Industrial camera photographs a color image frame of current flowers and trees blade, and view data is obtained by special-purpose ccd image acquisition module, carries out the blade testing by the DSP image processing module afterwards.Testing mainly comprises: leaf color detects, and the blade texture detects and blade shape detects three parts.In a single day find that the flowers and trees blade is in poor shape in the whole detection link, the lack of water phenomenon is arranged, just will produce trigger pip, this signal of industrial computer analysis is subsequently also controlled the sprinkling irrigation shower nozzle through strong power controller and is accomplished the task of watering.In addition, the staff that famous flower and tree is cultivated enterprise can obtain plant leaf blade information in real time through PC and smart mobile phone, and the industrial computer in the Long-distance Control booth sends the sprinkling irrigation instruction.Wherein:

One, after high-speed dsp computer vision module gets access to the coloured image of the frame flowers and trees blade that the current industrial camera photographs, through after the analog to digital conversion with image data storage in the IMAQ data storage area of high-speed dsp computer vision module.

What two, at first will carry out is the image pre-service.With the new coloured image gray processing of taking of industrial camera, then adopt the method for gray scale nonlinear transformation that its degree of comparing is stretched, again through the salt-pepper noise in the medium filtering filtering image.Use adaptive algorithm that gray level image is carried out Threshold Segmentation then, blade (comprising petiole) part is changed to black and background is changed to white.Again itself and initial coloured image are carried out the blade coloured image that OR operation can obtain wiping out background and comprise petiole at last.

Three,, brought difficulty for check and analysis subsequently because the phyllome picture that industrial camera obtains comprises petiole mostly.The present invention proposes a kind of petiole elimination method based on transversal scanning technology: at first will comprise the leaf image binaryzation of petiole, and with its rotate to be blade upwards, petiole position down.Next utilize transversal scanning to confirm the breadth extreme of petiole part, and then through the transversal scanning and the number of continuous black pixel point in the bianry image are added up the definite position that determines petiole.With isolated petiole image inverse, itself and the blade coloured image that comprises petiole are carried out OR operation, the blade coloured image that can obtain rejecting petiole and have complete display leaf margin information at last.

Four, carry out the color detection of blade below.Here at first with the blade coloured image that obtains in the step 3 by the RGB color space conversion to the hsv color space, set the variation range of H component and S component and the two realization divided equally.Then create the two-dimensional histogram that each dimension is divided equally, finally be transformed into the color 2 D histogram that rgb space is drawn the blade coloured image according to the data statistics on H plane and S plane.Through the color 2 D histogram of blade the flowers and trees blade under the different conditions is analyzed at last, whether belonged to exsiccosis thereby setting threshold determines plant.

The texture that five, next will carry out blade detects.The detection effect of blade textural characteristics is closely related with factors such as floristics, illumination condition and industrial camera resolution; So the present invention is provided with fixing led light source; And, used a kind of Canny operator detection method based on dynamic threshold to common flowers and trees blade.This method is at first with blade coloured image gray processing and adopt mean filter, to reduce the influence to subsequent algorithm.Then set dynamic threshold to different types of flowers and trees blade, and the edge in algorithm detection of utilization dual threshold and the connection layout picture.At last the texture quantity of blade unit area is added up, judge whether lack of water of plant through the density degree of texture.

What six, will carry out at last is the SHAPE DETECTION of blade, comprises that blade shape parameter detecting and blade area detect two parts.At first realize the detection of blade shape parameter: blade gray level image binaryzation at first to obtaining in the step 5, and make profile and extract.Then obtain blade centre of form point coordinate, and be maximum circumradius and minimum inscribed circle radius that blade is asked in the center of circle, with the ratio value defined blade shape parameter of the two with the centroid point through the mean value that calculates the contour pixel coordinate.At last in different time point mensuration and calculate the form parameter of blade, through with normal condition under the blade shape parameter relatively judge whether lack of water of plant.

Seven, blade area also is the important parameter of reflection vegetation growth state, degree of water shortage.Flowers and trees blade area detection method in the native system is different from the area detection method of General System.Because the blade real area is closely related with factors such as object of reference size, camera shooting angle and distances, so be difficult to obtain the accurate area of blade.So used a kind of blade area analytical approach among the present invention, confirmed the RP of plant hydropenia characteristic through the variation of comparing the flowers and trees blade area based on pixel.

Eight, be combined in detected leaf color distribution, texture density degree, form parameter and blade area change to judge whether flowers and trees are in exsiccosis in the step 4, five, six, seven, control the sprinkling irrigation shower nozzle through industrial computer simultaneously it is watered.And, then send to terminal device in time through remote server for the leaf color that has obtained in the above-mentioned steps, texture, shape and area information, realized the Internet of Things remote monitoring that famous flower and tree is cultivated.

In conjunction with instance specifically:

As shown in fig. 1, famous flower and tree is cultivated in the booth, is provided with led light source 1 and industrial camera 2 directly over the flowers and trees blade to be detected.This moment, industrial camera photographed the coloured image of a frame flowers and trees blade; View data is obtained by the ccd image acquisition module in the high-speed dsp chip 2 simultaneously; And give the DSP image processing module in the chip with this data transfer; Leaf image data to photographing are carried out high speed processing, thus the result of obtaining.After detecting completion, the high-speed dsp chip sends the view data result to industrial computer 7 through its built-in pci interface.Industrial computer sends processing signals to strong power controller 6 through built-in I/O interface on the one hand, and strong power controller will be according to the processing signals of receiving, 5 pairs of qualified flowers and trees of blade of sprinkling irrigation shower nozzle that control is placed on the aqueduct 4 are watered; On the other hand; Industrial computer sends the plant leaf blade information that collects to famous flower and tree and cultivates the remote server 8 in enterprise's office building; By server the blade data are collected and stored, and to terminal device PC 9 and smart mobile phone 10 data, services is provided in real time, make things convenient for flowers and trees to cultivate the enterprise work personnel and carry out remote monitoring; And the industrial computer in booth sends instruction at any time, and control sprinkling irrigation shower nozzle is accomplished the task of watering.System algorithm flow process of the present invention is as shown in Figure 2, and its concrete steps comprise:

One, the high-speed dsp vision module gets access to the coloured image I of the current taken frame flowers and trees blade of industrial camera 1,, deposit in the IMAQ data storage area in this module through after the analog to digital conversion.

What two, at first will carry out is the image pre-service.The flowers and trees blade coloured image I that industrial camera is newly photographed 1Gray processing then utilizes the method for gray scale nonlinear transformation that its degree of comparing is stretched, and can give prominence to the leaf characteristic in the gray level image like this, and nonlinear function uses exponential function to accomplish, and comes the salt-pepper noise in the filtering gray level image through medium filtering again.Use adaptive algorithm that gray level image is carried out Threshold Segmentation then, blade (comprising petiole) part is changed to black and background is changed to white, can obtain to comprise the blade bianry image I of petiole 2Again itself and initial coloured image are carried out the blade coloured image I that OR operation can obtain wiping out background and comprise petiole at last 3

Three,, brought difficulty for analytical work subsequently because the phyllome picture that industrial camera obtains comprises petiole mostly.The method of petiole generally all is the opening operation that adopts in the morphology processing in the rejecting image that uses in the past, promptly earlier it is carried out erosion operation, carries out dilation operation again and accomplishes.This method not only be difficult to realize separating fully of plant leaf blade and petiole in actual mechanical process, and the sawtooth information of blade edge is also had damage in various degree.To the problem that is occurred in the said method, the present invention proposes a kind of petiole elimination method: at first with the blade bianry image I that obtains comprising petiole in the step 2 based on the transversal scanning technology 2Rotate to blade up, petiole position vertically downward, begin from the image bottom, from down to up, scanning element point one by one from left to right, the row number note that first black pixel point place that scans is walked crosswise is made i, i.e. first pixel of petiole.Then at bianry image I 2[i-19, i] amount to the black pixel point numbers in the every row of statistics in 20 row, draw petiole width maximal value max through the mode that compares one by one.The bianry image of lining by line scan from top to bottom, from left to right then is a petiole with the continuous black picture element location determination that is less than or equal to max of counting out.Here it must be emphasized that it is the number of " continuously " pixel; Because the blade of a lot of plant leaf blade images and petiole have situation about being distributed in same the walking crosswise; Mistake just occurs if the mode that adopts statistics to walk crosswise interior black pixel point sum is judged, the petiole in causing partly walking crosswise can't be identified.Also need separate the petiole that has identified at last and obtain petiole bianry image I with inverse 4Again with itself and the blade coloured image I that comprises petiole 3Carry out OR operation, the blade coloured image I that can obtain rejecting petiole and have complete display leaf margin information 5This separation method that the present invention designed has all produced good effect through repeatedly practice to various types of plant leaf blade images.

Four, the following leaf color that at first carries out detects.Color detection is the important component part of foliar analysis; The ability of the plant synthesize chlorophyll (chl) that lack of water or sunshine are not enough is severely limited; Cause that leaf green content presents negative growth on the blade; The pigment of other colors just displays gradually, occurs so visually just have the blade of colors such as yellow, brownish red.To rejecting the blade coloured image I of petiole 5Handle, need to draw its color histogram.At first with its by the RGB color space conversion to the hsv color space; Set H component variation scope (0,180), S component variation scope (0,255) is divided into 16 grades with the H component, and the S component is divided into 8 grades.Create the two-dimensional histogram that each dimension is divided equally according to the data statistics on H plane and S plane then, obtain the color that current histogram is represented, be transformed into the color 2 D histogram that rgb space is drawn leaf image by initial pictures.Analyze the lack of water situation of plant leaf blade at last through the pixels statistics mean value that calculates color histogram.The present invention here analyzes to the fruit taro blade coloured image that closes of different conditions, and concrete statistics is as shown in table 1.Listed respectively in the table 1 be in new life, normal growth, slightly lack of water, extremely lack of water and withered state down close the really assembly average of taro leaf image pixel in the RGB passage.Be not difficult to find that G passage and R passage average present the trend of successively decreasing and increasing progressively respectively through data in the analytical table, and the variation of B passage is less relatively.There is no harm in setting threshold T [G]=148 and T [R]=120, when the average of G passage and R passage can satisfy condition G<T [G] and R>T [R] simultaneously, judge that it is in exsiccosis when closing fruit taro blade, control sprinkling irrigation shower nozzle waters then.

Table 1 plant leaf blade colouring information and vegetation growth state relation

What five, next will carry out is that the blade texture detects.The detection of texture is the important component part of foliar analysis equally, and leaf curling appears in plant under the lack of water condition easily, vein is gathered and phenomenon such as jaundice, so can realize the analysis to the plant hydropenia situation through the detection to the blade textural characteristics.Here mainly be to the main texture number in the leaf image unit area: at first to rejecting the blade coloured image I of petiole 5Gray processing then carries out the blade gray level image I after mean filter obtains smoothly 6Use a kind of Canny operator detection method then, to different flowers and trees leaf images, through setting dynamic dual threshold T based on dynamic threshold 1(the control edge connects) and T 2(controlling the initialization of strong edge) realized the detection to the blade texture.Can make T generally speaking 1: T 2Can obtain comparatively desirable effect at=1: 3, for example: the threshold value T that closes fruit taro blade 1: T 2Can get 75: 225, begonia blade threshold value was generally got 88: 264, Liriodendron blade threshold value then get 69: 207 comparatively suitable.Single length is regarded as blade master texture greater than the situation of 500 pixels in the interior profile of image.If it is I that blade gray level image completion texture detects the image that obtains 7, use the unit template of one 256 * 256 size this moment, statistics blade texture image I 7Main texture bar in the unit template is counted K.Count the density degree that Aver [K] analyzes the blade texture through the main texture bar in the unit of account area again, thereby judge this moment, whether plant was in exsiccosis.List unit area master's texture bar number of the involutory respectively fruit taro of the present invention blade normal condition, exsiccosis and withered state in the table 2 and got the testing result that obtains behind four position effective digitals.Can set decision threshold T [Aver [K]]=3.5 this moment equally, then judges plant hydropenia when testing result during greater than this value, and needs water to it at once.

Table 2 plant leaf blade texture information and vegetation growth state relation

What six, will carry out at last is the SHAPE DETECTION of blade, comprises that blade shape parameter detecting and blade area detect two parts.The flowers and trees blade under exsiccosis, usually demonstrate in the leaf margin turn over, vein is gathered and a series of characteristics such as blade fold, and these characteristics finally can show in the variation of blade shape, this also provides strong foundation for we judge the blade water shortage status.The present invention at first to the change of blade shape characteristic, realizes the detection to the blade shape parameter: at first with the flowers and trees blade gray level image I that obtains in the step 5 6Binaryzation obtains the desirable bianry image I of blade 7Then its scaling to 800 * 800 sizes are carried out profile and extract, obtain the contour images I of blade 8Add up the coordinate figure of blade profile pixel then successively, calculate the mean value of profile coordinate again, be the flowers and trees blade centroid point coordinate P (x, y).Next on blade profile, find out some A and the minimum some B of distance apart from maximum with centroid point P respectively, then the length of AP is the maximum circumradius I of blade profile 1, the length of BP is the minimum inscribed circle radius L of blade profile 2Here through calculating L 1And L 2Ratio define blade shape parameter lambda: λ=L 1/ L 2

Parameter lambda can be used as the parameter of real-time reflection leaf growth situation, and can not receive the influence of blade rotation, translation and scaled, but relevant with the kind of flowers and trees blade to be detected.The present invention will drip kwan-yin here as research object; Through the kwan-yin that drips under the normal condition being statically placed in dry environment and detecting and calculate form parameter λ, the GUANYINYE plate shape parameter lambda of dripping of two groups of detections under different conditions have been listed in the table 3 at different time point.Can reach a conclusion through analyzing: the blade shape parameter lambda under the normal condition is in the 2.2-2.5 scope; And the form parameter λ of exsiccosis lower blade is generally more than 2.6; So can be as criterion, control sprinkling irrigation shower nozzle waters to the flowers and trees of lack of water.

Table 3 plant leaf blade form parameter and vegetation growth state relation

Seven, blade area also is the important parameter of reflection vegetation growth state, and the characteristic that blade area significantly reduces can appear in plant under the water shortage status.Flowers and trees blade area detection method in the native system is different from the area detection method of General System.Because the blade real area is closely related with factors such as object of reference size, camera shooting angle and distances, so be difficult to obtain the accurate area value of blade.Therefore the present invention has used the blade area assay method based on pixel: read in the desirable bianry image I of the blade that obtains in the step 6 7, and two initial value m are set and n is 0.Each pixel all is a point in the coordinate system in the image, has two dimensional character.Begin to travel through the view picture bianry image from first point: if pixel (x, y) gray-scale value equals 255, explains that this pixel belongs to the pixel of background, and m is from increasing 1; If pixel (x, y) gray-scale value equals 0, explains that this pixel belongs to the pixel of blade, and n is from increasing 1.Finish up to traversal, can draw promptly that the blade sum of all pixels is n in the bianry image, and the sum of all pixels of entire image is (m+n).Draw blade number of pixels function distribution plan according to the sum of all pixels of different period flowers and trees blades at last, judge that according to the remarkable minimizing of blade number of pixels blade is in exsiccosis.Here still 800 * 800 big small drops of water GUANYINYE pictures in the step 6 are carried out the statistics of blade area information as research object, listed two groups in the table 3 resting on the result that the kwan-yin that drips in the dry environment carries out the blade area statistics.We may safely draw the conclusion through analyzing: the blade sum of all pixels under the normal condition is between 388500-390500; And the blade sum of all pixels under the exsiccosis is all reduced to below 387500; So can be as criterion, control sprinkling irrigation shower nozzle waters to the flowers and trees of lack of water.

Table 4 plant blade area information and vegetation growth state relation

Eight, be combined in the leaf color distribution, blade shape parameter, blade texture density degree and the blade area that detect in the step 4, five, six, seven and change to judge whether flowers and trees are in exsiccosis, control the sprinkling irrigation shower nozzle through industrial computer it is watered.And, then send to terminal device in time through remote server for the leaf color that has obtained in the above-mentioned steps, texture, shape and area information, realized the Internet of Things remote monitoring that famous flower and tree is cultivated.Hereto, just be through with to the blade testing of a strain famous flower and tree.

Above embodiment is merely the present invention's a kind of embodiment wherein, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to claim of the present invention.Should be pointed out that for the person of ordinary skill of the art under the prerequisite that does not break away from the present invention's design, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with accompanying claims.

Claims (12)

1. based on the vegetation growth state monitoring system of computer vision and Internet of Things; It is characterized in that: the control and executive module that it comprises high-speed dsp computer vision module that plant leaf blade is taken and taken image is gathered and analyzed, plant is watered; And send plant leaf blade information to Internet of Things module that remote terminal supplies staff's reference and operation through remote server, data connection between said high-speed dsp computer vision module, control and executive module and the Internet of Things module.
2. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 1 is characterized in that: said high-speed dsp computer vision module comprises the industrial camera that is positioned at directly over the plant to be monitored, the DSP image processing module that is used to obtain the ccd image acquisition module of the taken image of said industrial camera and includes the high-speed dsp chip that is used for Flame Image Process.
3. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 2; It is characterized in that: said industrial camera is wrapped with one deck waterproof membrane all over the body; Said industrial camera is other to be provided with lighting source, and said lighting source is a led light source.
4. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 2 is characterized in that: said ccd image acquisition module and DSP image processing module all are loaded on the portable dsp board card of encapsulation type.
5. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 1; It is characterized in that: said control and executive module comprises: a plurality of sprinkling irrigation shower nozzles that are positioned at plant to be monitored top and the aqueduct that supplies water for said sprinkling irrigation shower nozzle; Said sprinkling irrigation shower nozzle is controlled by strong power controller, and said strong power controller all is connected industrial computer with said high-speed dsp computer vision module, Internet of Things module.
6. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 2, it is characterized in that: the step of its work is:
1) said industrial camera is regularly taken the coloured image of the blade of plant to be monitored;
2) said ccd image acquisition module sends said DSP image processing module to after obtaining said coloured image;
3) said DSP image processing module detects the plant leaf blade in the said coloured image, if find plant hydropenia then notify said control and executive module that plant is watered.
7. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 6; It is characterized in that: in the said step 3), said DSP image processing module carries out in the following detection one or more to the plant leaf blade in the said coloured image: leaf color detects, the blade texture detects and blade shape detects.
8. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 7 is characterized in that: the step that said leaf color detects is:
1) plant leaf blade in the said coloured image being removed petiole handles;
2) with said coloured image by the RGB color space conversion to the hsv color space, set H component and S component variation scope and the two divided equally;
3) create the two-dimensional histogram that each dimension is divided equally according to the data statistics on H plane and S plane;
4) utilize former said coloured image said two-dimensional histogram to be converted to the color 2 D histogram of the rgb space of blade;
5) setting threshold through the flowers and trees blade in the said color 2 D histogram is analyzed, determines plant and whether belongs to exsiccosis.
9. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 7 is characterized in that: the step that said blade texture detects is:
1) with said coloured image gray processing and adopt mean filter;
2) set dynamic threshold to different types of flowers and trees blade, the edge in algorithm detection of utilization dual threshold and the connection layout picture;
3) the texture quantity of blade unit area is added up, judge through the density degree of texture whether plant belongs to exsiccosis.
10. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 7; It is characterized in that: said blade shape detects and comprises that blade shape parameter detecting and blade area detect, and the step of wherein said blade shape parameter detecting is:
1) plant leaf blade in the said coloured image being removed petiole handles;
2), and extract the profile of plant leaf blade with the gray level image binaryzation of the plant leaf blade in the said coloured image;
3) mean value of the pixel coordinate of the profile through calculate extracting plant leaf blade obtains blade centre of form point coordinate, and is that maximum circumradius of blade and minimum inscribed circle radius are asked in the center of circle with the centroid point, with the two ratio as the blade shape parameter;
4) compare through the blade shape parameter that calculates and the blade shape parameter under the normal condition, judge whether plant belongs to exsiccosis.
11. the vegetation growth state monitoring system based on computer vision and Internet of Things according to claim 10 is characterized in that: the step that said blade area detects is:
1) with the gray level image binaryzation of the plant leaf blade in the said coloured image, obtains binary image;
2) all gray-scale values of the said binary image of traversal statistics are the sum of all pixels of 0, with its blade area as plant leaf blade;
3) historical variations of the blade area of this plant leaf blade relatively judges that if blade area diminishes this plant leaf blade is for belonging to exsiccosis.
12. according to Claim 8 or 10 one of them described vegetation growth state monitoring system, it is characterized in that: wherein the plant leaf blade in the said coloured image is removed the petiole processed steps and comprise based on computer vision and Internet of Things:
1) at first will comprise the leaf image binaryzation of petiole, and with its rotate to be blade upwards, petiole position down, to obtain bianry image;
2) said bianry image is carried out transversal scanning from bottom to top, to confirm the breadth extreme of petiole part;
3) through to the continuously transversal scanning of black pixel point and number statistics in the bianry image, with less than step 2) in the part of definite petiole part breadth extreme be petiole, thereby determine the definite position of petiole;
4) with the petiole image inverse that obtains, itself and former said coloured image are carried out or operate, obtained rejecting the blade coloured image of petiole.
CN 201110451798 2011-12-30 2011-12-30 Plant growth condition monitoring system based on compute vision and internet of things CN102564593B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110451798 CN102564593B (en) 2011-12-30 2011-12-30 Plant growth condition monitoring system based on compute vision and internet of things

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110451798 CN102564593B (en) 2011-12-30 2011-12-30 Plant growth condition monitoring system based on compute vision and internet of things

Publications (2)

Publication Number Publication Date
CN102564593A true CN102564593A (en) 2012-07-11
CN102564593B CN102564593B (en) 2013-10-02

Family

ID=46410619

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110451798 CN102564593B (en) 2011-12-30 2011-12-30 Plant growth condition monitoring system based on compute vision and internet of things

Country Status (1)

Country Link
CN (1) CN102564593B (en)

Cited By (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103075982A (en) * 2012-12-13 2013-05-01 中国农业大学 Three-dimensional reconstruction and measurement device and method of greenhouse strawberry canopy
CN103309310A (en) * 2013-05-21 2013-09-18 江苏大学 Method for monitoring operation of plug seedling transplanting robot based on laser scanning
CN103591887A (en) * 2013-09-30 2014-02-19 北京林业大学 Method for detecting regional phenotype of Arabidopsis
CN103749246A (en) * 2014-01-24 2014-04-30 成都万先自动化科技有限责任公司 Lawn watering machine without dead corner
WO2014166081A1 (en) * 2013-04-10 2014-10-16 中国科学院自动化研究所 Plant characteristic data measuring and storing system based on internet of things and cloud platform
CN104705080A (en) * 2015-03-20 2015-06-17 温弘成 Bottle culture plant planting control method
CN105137932A (en) * 2015-08-12 2015-12-09 成都易思科科技有限公司 Monitoring information service cloud platform comprehensive management system based on IOT (Internet of Things) application
CN105277495A (en) * 2015-10-26 2016-01-27 北京农业信息技术研究中心 Method for monitoring and acquiring evaporation characteristics of fog droplets on leaf surfaces in time sequence mode
CN105427279A (en) * 2015-11-02 2016-03-23 中国农业大学 Grassland drought status monitoring system based on and machine vision and Internet of things, grassland drought status monitoring method
CN105472338A (en) * 2015-12-11 2016-04-06 卢志文 Plant wall pattern monitoring system
CN105516663A (en) * 2015-12-11 2016-04-20 卢志文 Plant wall growth situation monitoring system
CN105516665A (en) * 2015-12-11 2016-04-20 卢志文 Plant wall monitoring system
CN105493840A (en) * 2015-12-11 2016-04-20 卢志文 System for monitoring growth condition of plant wall
CN105554448A (en) * 2015-12-11 2016-05-04 卢志文 Image analysis-based plant wall monitoring system
CN105574897A (en) * 2015-12-07 2016-05-11 中国科学院合肥物质科学研究院 Crop growth situation monitoring Internet of Things system based on visual inspection
CN105628206A (en) * 2015-12-30 2016-06-01 中国农业科学院茶叶研究所 Method for measuring colors of tea leaves at different positions
CN105806999A (en) * 2016-03-15 2016-07-27 宁德师范学院 Plant growth testing system and method
CN106054844A (en) * 2016-07-12 2016-10-26 河海大学 Intelligent remote agricultural management system
CN106682570A (en) * 2016-11-04 2017-05-17 东莞市隆声智能科技有限公司 Method and device for monitoring growing situations of plants
CN106767564A (en) * 2016-11-03 2017-05-31 广东工业大学 A kind of detection method for being applied to phone housing surface roughness
CN106768081A (en) * 2017-02-28 2017-05-31 河源弘稼农业科技有限公司 A kind of method and system for judging fruits and vegetables growth conditions
CN106873678A (en) * 2017-03-02 2017-06-20 湖南省烟草公司长沙市公司宁乡县分公司 A kind of self-adaptation control method of tobacco seedlings greenhouse and tobacco seedlings greenhouse
CN106895883A (en) * 2017-04-28 2017-06-27 无锡北斗星通信息科技有限公司 Intelligent environment is detected and maintaining method
CN107609078A (en) * 2017-09-04 2018-01-19 北京农业信息技术研究中心 Growing state survey model update method, sensor, server and system
CN108074236A (en) * 2017-12-27 2018-05-25 广东欧珀移动通信有限公司 Irrigating plant based reminding method, device, equipment and storage medium
WO2018123630A1 (en) * 2016-12-28 2018-07-05 Honda Motor Co.,Ltd. Information processing device, water-supply system, information processing system and program
CN108921835A (en) * 2018-06-28 2018-11-30 深圳市诚品鲜智能科技股份有限公司 Crop control method and relevant apparatus and storage medium based on machine vision
CN109145785A (en) * 2018-08-03 2019-01-04 百度在线网络技术(北京)有限公司 The determination method and apparatus of plant-nursing mode
CN109168724A (en) * 2018-10-15 2019-01-11 李杨 A kind of seedling system and method
CN109470179A (en) * 2018-12-14 2019-03-15 武汉大学 A kind of extensive water ploughs vegetables growing way detection system and method
CN110648069A (en) * 2019-09-26 2020-01-03 张信信 Vegetable quality analysis system based on transport vehicle

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070289207A1 (en) * 2005-12-21 2007-12-20 May George A Expert system for controlling plant growth in a contained environment
CN201298902Y (en) * 2008-11-04 2009-08-26 北京中科嘉和科技发展有限公司 Video terminal with environmental monitoring data
WO2010031780A1 (en) * 2008-09-16 2010-03-25 Basf Plant Science Gmbh Method for improved plant breeding
CN102156923A (en) * 2011-04-22 2011-08-17 华建武 Comprehensive plant production management system and method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070289207A1 (en) * 2005-12-21 2007-12-20 May George A Expert system for controlling plant growth in a contained environment
WO2010031780A1 (en) * 2008-09-16 2010-03-25 Basf Plant Science Gmbh Method for improved plant breeding
CN201298902Y (en) * 2008-11-04 2009-08-26 北京中科嘉和科技发展有限公司 Video terminal with environmental monitoring data
CN102156923A (en) * 2011-04-22 2011-08-17 华建武 Comprehensive plant production management system and method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
孙燕等: "基于叶片面积的温室植物水分监测系统的设计", 《传感器与微系统》 *

Cited By (45)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103075982B (en) * 2012-12-13 2015-05-13 中国农业大学 Three-dimensional reconstruction and measurement device and method of greenhouse strawberry canopy
CN103075982A (en) * 2012-12-13 2013-05-01 中国农业大学 Three-dimensional reconstruction and measurement device and method of greenhouse strawberry canopy
WO2014166081A1 (en) * 2013-04-10 2014-10-16 中国科学院自动化研究所 Plant characteristic data measuring and storing system based on internet of things and cloud platform
CN103309310A (en) * 2013-05-21 2013-09-18 江苏大学 Method for monitoring operation of plug seedling transplanting robot based on laser scanning
CN103309310B (en) * 2013-05-21 2015-03-25 江苏大学 Method for monitoring operation of plug seedling transplanting robot based on laser scanning
CN103591887A (en) * 2013-09-30 2014-02-19 北京林业大学 Method for detecting regional phenotype of Arabidopsis
CN103591887B (en) * 2013-09-30 2016-06-15 北京林业大学 A kind of detection method of arabidopsis region phenotype
CN103749246A (en) * 2014-01-24 2014-04-30 成都万先自动化科技有限责任公司 Lawn watering machine without dead corner
CN104705080A (en) * 2015-03-20 2015-06-17 温弘成 Bottle culture plant planting control method
CN104705080B (en) * 2015-03-20 2016-11-30 温弘成 Plant plantation control method supported by bottle
CN105137932A (en) * 2015-08-12 2015-12-09 成都易思科科技有限公司 Monitoring information service cloud platform comprehensive management system based on IOT (Internet of Things) application
CN105277495A (en) * 2015-10-26 2016-01-27 北京农业信息技术研究中心 Method for monitoring and acquiring evaporation characteristics of fog droplets on leaf surfaces in time sequence mode
CN105277495B (en) * 2015-10-26 2019-05-28 北京农业信息技术研究中心 A kind of time series pattern droplet blade face evaporation characteristic monitors acquisition method
CN105427279B (en) * 2015-11-02 2018-06-26 中国农业大学 A kind of grassland Drought Information Monitoring System and method based on computer vision and Internet of Things
CN105427279A (en) * 2015-11-02 2016-03-23 中国农业大学 Grassland drought status monitoring system based on and machine vision and Internet of things, grassland drought status monitoring method
CN105574897A (en) * 2015-12-07 2016-05-11 中国科学院合肥物质科学研究院 Crop growth situation monitoring Internet of Things system based on visual inspection
CN105493840A (en) * 2015-12-11 2016-04-20 卢志文 System for monitoring growth condition of plant wall
CN105554448A (en) * 2015-12-11 2016-05-04 卢志文 Image analysis-based plant wall monitoring system
CN105472338A (en) * 2015-12-11 2016-04-06 卢志文 Plant wall pattern monitoring system
CN105516663A (en) * 2015-12-11 2016-04-20 卢志文 Plant wall growth situation monitoring system
CN105516665A (en) * 2015-12-11 2016-04-20 卢志文 Plant wall monitoring system
CN105628206A (en) * 2015-12-30 2016-06-01 中国农业科学院茶叶研究所 Method for measuring colors of tea leaves at different positions
CN105628206B (en) * 2015-12-30 2018-03-20 中国农业科学院茶叶研究所 A kind of method for the color for determining diverse location tea leaf
CN105806999A (en) * 2016-03-15 2016-07-27 宁德师范学院 Plant growth testing system and method
CN106054844A (en) * 2016-07-12 2016-10-26 河海大学 Intelligent remote agricultural management system
CN106054844B (en) * 2016-07-12 2019-04-30 河海大学 A kind of agricultural intelligent remote management system
CN106767564A (en) * 2016-11-03 2017-05-31 广东工业大学 A kind of detection method for being applied to phone housing surface roughness
CN106682570A (en) * 2016-11-04 2017-05-17 东莞市隆声智能科技有限公司 Method and device for monitoring growing situations of plants
WO2018123630A1 (en) * 2016-12-28 2018-07-05 Honda Motor Co.,Ltd. Information processing device, water-supply system, information processing system and program
CN106768081A (en) * 2017-02-28 2017-05-31 河源弘稼农业科技有限公司 A kind of method and system for judging fruits and vegetables growth conditions
CN106768081B (en) * 2017-02-28 2018-02-02 河源弘稼农业科技有限公司 A kind of method and system for judging fruits and vegetables growth conditions
CN106873678A (en) * 2017-03-02 2017-06-20 湖南省烟草公司长沙市公司宁乡县分公司 A kind of self-adaptation control method of tobacco seedlings greenhouse and tobacco seedlings greenhouse
CN106873678B (en) * 2017-03-02 2018-10-12 湖南省烟草公司长沙市公司宁乡县分公司 A kind of self-adaptation control method of tobacco seedlings greenhouse
CN106895883B (en) * 2017-04-28 2019-03-26 北京慧辰智慧生态环境科技有限公司 Intelligent environment detection and maintaining method
CN106895883A (en) * 2017-04-28 2017-06-27 无锡北斗星通信息科技有限公司 Intelligent environment is detected and maintaining method
CN107609078A (en) * 2017-09-04 2018-01-19 北京农业信息技术研究中心 Growing state survey model update method, sensor, server and system
CN108074236B (en) * 2017-12-27 2020-05-19 Oppo广东移动通信有限公司 Plant watering reminding method, device, equipment and storage medium
CN108074236A (en) * 2017-12-27 2018-05-25 广东欧珀移动通信有限公司 Irrigating plant based reminding method, device, equipment and storage medium
CN108921835A (en) * 2018-06-28 2018-11-30 深圳市诚品鲜智能科技股份有限公司 Crop control method and relevant apparatus and storage medium based on machine vision
CN109145785A (en) * 2018-08-03 2019-01-04 百度在线网络技术(北京)有限公司 The determination method and apparatus of plant-nursing mode
CN109168724A (en) * 2018-10-15 2019-01-11 李杨 A kind of seedling system and method
CN109470179A (en) * 2018-12-14 2019-03-15 武汉大学 A kind of extensive water ploughs vegetables growing way detection system and method
CN109470179B (en) * 2018-12-14 2020-10-13 武汉大学 Large-scale hydroponic vegetable growth detection system and method
CN110648069A (en) * 2019-09-26 2020-01-03 张信信 Vegetable quality analysis system based on transport vehicle
CN110648069B (en) * 2019-09-26 2020-08-14 桐乡市常新农机专业合作社 Vegetable quality analysis system based on transport vehicle

Also Published As

Publication number Publication date
CN102564593B (en) 2013-10-02

Similar Documents

Publication Publication Date Title
Chen et al. Counting apples and oranges with deep learning: A data-driven approach
Hamuda et al. A survey of image processing techniques for plant extraction and segmentation in the field
CN104238602B (en) Greenhouse intelligent control management system based on information gathering
US10192185B2 (en) Farmland management system and farmland management method
CN105173085B (en) Unmanned plane variable farm chemical applying automatic control system and method
US10568316B2 (en) Apparatus and methods for in-field data collection and sampling
Bai et al. Crop segmentation from images by morphology modeling in the CIE L* a* b* color space
Perez et al. Colour and shape analysis techniques for weed detection in cereal fields
CN101356877B (en) Cucumber picking robot system and picking method in greenhouse
CA2740503C (en) Variable rate sprayer system and method of variably applying agrochemicals
McCarthy et al. Applied machine vision of plants: a review with implications for field deployment in automated farming operations
US10028426B2 (en) Agronomic systems, methods and apparatuses
US9538714B2 (en) Managing resource prescriptions of botanical plants
ES2611212T3 (en) Plant growth kinetics captured by motion tracking
Narvaez et al. A survey of ranging and imaging techniques for precision agriculture phenotyping
CN103593962B (en) Organic vegetable quality remote network real-time monitoring method based on sensing communication
Zhu et al. In-field automatic observation of wheat heading stage using computer vision
Vaesen et al. Ground-measured spectral signatures as indicators of ground cover and leaf area index: the case of paddy rice
WO2016009752A1 (en) Information processing device, method for generating control signal, information processing system, and program
US20100268391A1 (en) Resource Use Management
CN102954816B (en) Crop growth monitoring method
Tang et al. Weed detection using image processing under different illumination for site-specific areas spraying
Smith Review of precision irrigation technologies and their applications
Yu et al. Automatic image-based detection technology for two critical growth stages of maize: Emergence and three-leaf stage
US20180014452A1 (en) Agronomic systems, methods and apparatuses

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
C14 Grant of patent or utility model