• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a system (S) for estimating a harvest volume within a vineyard organized in rows (R1, R2, R3) of vine stocks, comprising: - a photographic device (DP) for capturing a photographic image of a set (E) of vines as the system travels between two rows (R1, R2) of the vineyard; means (Dd) for measuring a distance (d) between the vines and the photographic device (DP); calculating means (MC) for detecting bays within the photographic image, determining a number of bays on a cluster, determining a number of bunches on a vine and determining an indicator of the volume according to the number of bays on a cluster, the number of clusters on a vine and the diameter of the bays determined from the photographic image and the distance measurement.
    公开号:FR3036830A1
    申请号:FR1554862
    申请日:2015-05-29
    公开日:2016-12-02
    发明作者:Christian Germain；Barna Keresztes；Gilbert Grenier；Olivier Lavialle；Costa Jean-Pierre Da
    申请人:Ecole Nat Superieure Des Sciences Agronomiques De Bordeaux (bordeaux Sciences Agro)；Centre National de la Recherche Scientifique CNRS；Universite de Bordeaux；Institut Polytechnique de Bordeaux；
    IPC主号:
    专利说明:

    [0001] FIELD OF THE INVENTION The invention relates to the field of viticultural agriculture. More precisely, it concerns an automatic tool for helping to estimate a harvest volume in a vineyard.
    [0002] BACKGROUND OF THE INVENTION In order to best manage a vine, a winemaker must have an estimate of the volume of harvest of grape berries. This estimation can in particular enable him to anticipate on the human and material means to put in place to actually carry out the harvest, to optimize the organization of the harvest. This estimate can be determined by plot of a vineyard, to the extent that the conditions of terroir, micro-climates, cutting strategy and culture, etc. can influence the evolution of the vine in a very local way. In addition, the evolution of this estimate can be used as a parameter to determine the best time to achieve this harvest. It can also detect abnormalities in the development of grape berries. In general, this knowledge is empirical or constructed manually. A winemaker must therefore pass through the vineyards to determine a number of parameters, including the number of clusters per vine and the average weight of a cluster. To the extent that it is humanly impossible to determine these measurements for an entire vine, sampling must be done. This sampling should be carried out as randomly as possible and without human bias. The number of samples depends on the desired precision and variance in the geographical area studied.
    [0003] As a result, this method is complex to implement and also very time consuming. The error rate is also quite high, usually between 20% and 50% and can reach 200% in some cases. The article by P. Clingeleffer, G. Dunn, M. Krstic and S. Martin, "Crop development, crop estimation and crop control and quality assurance and production of major wine grape varieties", can be consulted in this regard: A national approach In addition, in order to obtain a satisfactory estimate, it appears that several samplings per season may be required. This results in the destruction of a significant number of berries. In addition, automatic methods have been devised to improve the estimation process of a harvest volume. For example, the article "Yield Estimate in Vineyards: Experiments with Different Varieties and Calibration Procedure" by Stephen Nuske et al. The IEEE / RSI International Conference 25-30 Sept. 2011 describes such a process.The yield of the crop is estimated on the basis of the number of bays counted automatically by the vehicle and Comparing this number with data from the previous year, this method does not give satisfactory results and its result depends on the values obtained previously.In addition, the described mechanism is based on a shoot in the vines and a step Further digital processing of the obtained images has to be done in a remote center of the vines, and the data capture process is complex and expensive because of the 3D cameras.
    [0004] SUMMARY OF THE INVENTION The object of the present invention is to provide a system and method at least partially overcoming the aforementioned drawbacks. To this end, the present invention provides a method for estimating a harvest volume within a vineyard organized in a row of vines, comprising steps of: - capture of a photographic image of a set of vines, by photographic device integral with an automatic system circulating between two rows of vines; - detection of the berries within said photographic image; Estimating a number of berries on a cluster; estimating a number of clusters on a vine of said set; estimating said volume from said number of berries on a cluster, said number of clusters on a vine and the diameter of said bays determined from said photographic image and a measurement of distance between said set of vine stocks and said automatic system. According to preferred embodiments, the invention comprises one or more of the following features which can be used separately or in partial combination with one another or in total combination with one another: - the detection of bays comprises the determination of circular forms within said photographic image; The determination of a number of berries on a cluster comprises the use of an experimental model linking a number of berries visible on a photographic image to said number of berries on a cluster; 5 - the number of clusters on a vine is determined by an abacus depending on the variety of said vine and vine holding technique implemented by the operator of said vineyard; - the harvest volume is dependent on a geographical location.
    [0005] A second object of the invention relates to a system for estimating a harvest volume within a vineyard organized in rows of vines, comprising: a photographic device for capturing a photographic image a set of vines, when said system circulates between two rows of said vineyard; means for measuring a distance between said set of vine stocks and said photographic device; calculation means for detecting bays within said photographic image, determining a number of bays on a cluster, determining a number of bunches on a group of said set and determining an indicator of said volume as a function of said number of bays on a cluster, said number of clusters on a vine and the diameter of said bays determined from said photographic image and said distance measurement.
    [0006] According to preferred embodiments, the system according to the invention comprises one or more of the following characteristics which can be used separately or in partial combination with one another or in total combination with one another: the calculation means are designed to detect berries by determining circular shapes within said photographic image; the calculation means are provided for determining a number of bays on a cluster by the use of an experimental model linking a number of bays visible on a photographic image to said number of bays on a cluster; - The calculation means are provided to determine the number of clusters on a vine according to an abacus depend on the variety of said vine 10 and vine held by the operator of said vineyard. - The system further comprises a location device to make said harvest volume dependent on a geographical location.
    [0007] Another aspect of the invention relates to a vehicle having a system as previously defined. Other features and advantages of the invention will appear on reading the following description of a preferred embodiment of the invention, given by way of example and with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 schematically represents an exemplary implementation of the invention FIG. 2 schematically represents an example of mapping a number of berries visible on a photographic image to a number of bays actually counted on a cluster.
    [0008] FIG. 3 diagrammatically represents an exemplary embodiment of the invention by means of a straddle.
    [0009] DETAILED DESCRIPTION OF THE INVENTION Typically, a vineyard is organized in rows of 5 vines, so as to allow the operator to move between each row, including using a vehicle. As shown in Figure 1, the system S according to the invention is intended to flow between two rows R1, R2 of the vineyard. This system can be worn by a member of the farm staff. Preferably, it can be made integral with a vehicle that can be towed by a staff member or be a motor vehicle. According to one embodiment, this motor vehicle can be a robot. The vehicle can take the form of a tractor, a quad, a suit, etc. Figure 3 illustrates the particular case of a straddle vehicle V. This figure shows a vine in a cross section. Three rows R1, R2, R3 are shown. The straddle is a vehicle having two legs, J1, J2 terminated by wheels and adapted to roll in the grooves formed by two successive rows. Each leg can roll in a groove distinct from the other, so that the body of the straddle can be located above the vines. This body may include a passenger compartment in which takes place a driver. The photographic device and the distance measuring means, which will be described below, can be located on one of the legs of the straddle, at the height of the grape bunches of the vine. In general, the system S has a DP photographic device arranged so as to capture photographic images of the vines. This photographic device is suitable for capturing photographic images. It may therefore be a digital photographic apparatus, but also a video camera, or any other suitable device. Preferably, it can be arranged so that the axis of view is perpendicular to the axis of the rows of vines and, therefore, the axis of displacement of the system. This photographic device DP is arranged in the system S so that its elevation relative to the ground corresponds substantially to that of the bunches of grapes. This elevation can therefore be an adjustable parameter depending on the type of vine (grape variety) and on the way the operator prunes and uses his vine. Thus, as the system travels between two rows, the DP photographic device can capture photographic images corresponding to an entire row. Each photographic image, however, corresponds to a set E, depending on the distance d between the 15 vines and the photographic device DP as well as the opening angle thereof. This set E corresponds to a set of vines, which can possibly be reduced to a single vine stock. The system S can thus be provided to capture series of photographic images covering each of the adjoining sets, so that all the photographic images for one row cover all the vines of a row. According to another embodiment, the system S captures photographic images according to a determined rhythm, generally periodic, which may be dependent on the speed of movement of the system. It may further have calculating means MC designed to determine overlaps between photographic images in order to avoid taking into account several times the same photographed object (cluster, vine, bay ...). According to one embodiment of the invention, a sampling may be set up so as to capture photographic images only from one part of a row. This sampling can be controlled by a random generator 3036830 8 to avoid any bias. Preferably, this sampling is performed on a significant fraction of the row. It can also be provided that the system S has a lighting device, in particular to allow the capture of photographic images 5 whatever the natural lighting conditions, but also to standardize these conditions of a photographic image to another. The lighting device may in particular be chosen from various possible lighting such as a halogen lamp, LED lighting, a flash, strobe or continuous lamp, etc.
    [0010] The photographic device can be adapted to capture photographic images of different natures. In particular, the photographic images captured can be monospectral, multi spectral, hyperspectral ... Spectroimaging, also called "hyperspectral" imaging as opposed to "multispectral" or "superspectral" imaging, is a technology allowing the representation of a scene following a large number of spectral bands (generally more than one hundred), narrow (<10 nm) and contiguous. On the other hand, the photographic device can operate in different wavelength ranges. It can work in visible light, but also in ultraviolet or infra-red. Preferably, the photographic device can operate in the visible range and in 3 bands: red, green, blue, in order to generate a color photographic image.
    [0011] In the remainder of the description, it will generally be of interest to process a given photographic image corresponding to a set E of vine stocks. The shift from estimating a crop volume corresponding to an image to a given area of a wine farm can be done in various ways, depending on how the photographic images were captured. According to one embodiment of the invention, the system can have two photographic devices in order to capture in parallel photographic images of the two opposite rows. Thus, a first DP photographic device may be located to the left of the system S (in FIG. 1) and capture photographic images of the row R1, while a second device, not shown in the figure, is located on the right and captures 10 photographic images of the row R2. Of course, other implementations are possible, for example having more than two photographic devices. The photographic images provided by the photographic device (s) DP are digital images allowing their processing by the calculation means MC, also embedded in the system S. These calculation means are designed to: detect bays within the photographic image representing the set E, - estimating a number of berries on a cluster, - estimating a number of clusters on a vine of the set E and - estimating the harvest volume, from or in function, this number of clusters per cep and the diameter of the racks determined from the photographic image and a measurement of the distance d. This estimation of the volume can be made by a statistical approximation starting from these elements. The system S may further comprise means Dd for measuring a distance d between said set of vine stocks and the photographic device DP. These means can be a rangefinder for example.
    [0012] As will be seen later, this distance makes it possible to scale the captured image and transform a measurement in number of pixels into a size in centimeters. According to some embodiments, the distance measuring means Dd (rangefinders) can be integrated within the DP photographic devices. The detection of grape berries in a digital image can be carried out in different ways.
    [0013] According to one embodiment of the invention, the computing means MC determine the circular shapes within the digital image by detecting arcs of a circle. Indeed, a bay may be partially obscured by another bay located in a plane closer to the camera, so that only an arc is detectable. From the arcs of circles detected, it is possible to deduce the visible berries. The calculation means MC can then estimate the number of these bays (visible totally or partially) in the photographic image, as well as the diameter of the bays. An average diameter can be determined at this stage, or the diameters of each bay can be stored to allow more complex statistical calculations. The calculation means MC are then provided to determine a number of berries on a cluster. Indeed, a photographic image only provides a two-dimensional view of a set of clusters. However, a cluster is a three-dimensional object, part of which is masked because it is opposed to the photographic device. Various stereological techniques are possible to perform this passage in three dimensions in order to estimate a number of berries on a cluster from a previously visible number of berries visible.
    [0014] One possible implementation is to use an experimental model linking a number of berries visible on a photographic image to said number of berries on a cluster. This model can be constructed by matching the results provided by the previous step on the number of berries visible for a cluster to a manual measurement of the total number of berries on that same cluster. Figure 2 illustrates such mapping. The x-axis indicates the actual number of bays per cluster (ie manually counted). The ordinate axis indicates the number of 10 visible bays automatically determined by the computing means MC from a digital image representing the same cluster. Each cluster is represented by a point (corresponding to the legend "series 1"). A very strong correlation is observed, making it possible to determine a reliable polynomial model with R2 = 0.92. For each image, this model makes it possible to automatically determine the true number of berries per cluster in 47% of cases, with an average error of less than 2.5%. It is possible to adapt this model according to the type of variety, for each vineyard, etc. It is also possible to adapt this model to the parameters of the previous step. Indeed, depending on the degree of taking into account an arc to generate the detection of a bay, the number of berries visible per cluster will be different, but it is possible to adapt the polynomial model to determine the real number of bays, so that this actual number is independent of certain parameters of the berry detection algorithm. A next step of the method according to the invention, implemented by the calculation means MC, is to estimate a number of clusters for a vine stock.
    [0015] This step is based on counting the number of clusters visible on the processed photographic image for a given vine stock. Then, an abacus can be used to deduce the actual number of clusters on this vine stock. The abacus is also determined experimentally and allows to take into account clusters masked by the vine leaves. This abacus may depend on the type of grape variety but also on the techniques of keeping the vine: depending on how the farmer prunes his vineyard, a different chart may be used. From the number of berries on a cluster, the number of clusters on a vine and the diameter of the bays, the calculating means MC can estimate the harvest volume. The diameter of the berries is an important parameter since it allows to directly supply a volume of the berry and therefore a weight, and, consequently, a quantity of grape juice that can be extracted. In addition, the measurement of this diameter makes it possible to obtain an indication of the degree of maturity of a vine and to better determine both the right moment for a crop and also any anomalies in the development of the vine. Consequently, the harvest volume estimated by the steps of the invention makes it possible to have a good estimate of the yield of a vine 20 and possibly to determine a moment optimizing this yield. Thus, the method according to the invention makes it possible to achieve a very good estimate, at a lower cost than the methods of the prior art known. In addition, it is non-destructive, and can thus be used repeatedly, if necessary, if it is desired to refresh the estimate over time (for example, at regular intervals, or at the same time). following an inclement weather, etc.). According to one embodiment of the invention, the system S comprises a DL geolocation device. This device can be a simple GPS device ("Global Positioning System") or GNSS ("Global 3036830 13 Navigation Satellite System"). It allows to associate each photographic image to a geographical position. Thus, it is possible to determine location-dependent volume indicators. For example, a function linking a crop volume indicator to the location can be determined. This geolocation of the indicator makes it possible to refine the knowledge that the operator can have of his wine exploitation. In particular, it makes it possible to adapt the response to be made more accurately and to detect local anomalies. If an anomaly is detected, the geolocation also makes it possible to intervene on the corresponding site. Of course, the present invention is not limited to the examples and to the embodiment described and shown, but it is capable of numerous variants accessible to those skilled in the art. 15
    权利要求:
    Claims (11)
    [0001]
    REVENDICATIONS1. Method for estimating a harvest volume within a vineyard organized in a vine stock row, comprising steps of capturing a photographic image of a set of vines, by a solidary photographic device an automatic system circulating between two rows of vines; detecting berries within said photographic image; estimating a number of berries on a cluster; estimating a number of clusters on a vine of said set estimating said volume from said number of bays on a cluster, said number of clusters on a vine and the diameter of said bays determined from said photographic image and a measurement of distance between said set of vine stocks and said automatic system.
    [0002]
    2. Method according to the preceding claim, wherein the detection of berries comprises the determination of circular shapes within said photographic image.
    [0003]
    3. Method according to one of the preceding claims, wherein the determination of a number of berries on a cluster comprises the use of an experimental model linking a number of berries visible on a photographic image said number of berries on a cluster .
    [0004]
    4. Method according to one of the preceding claims, wherein the number of clusters on a vine is determined by an abacus depending on the grape variety of said vine and vine holding technique implemented by the operator of said vineyard. 3036830 15
    [0005]
    5. Method according to one of the preceding claims, wherein said harvest volume is dependent on a geographical location.
    [0006]
    6. System (S) for estimating a harvest volume within a vineyard organized in rows (R1, R2, R3) of vines, comprising: - a photographic device (DP) for the capturing a photographic image of a set (E) of vines, when said system circulates between two rows (R1, R2) of said vineyard operation; means (Dd) for measuring a distance (d) between said set of vines and said photographic device (DP); calculating means (MC) for detecting arrays within said photographic image, determining a number of arrays on a cluster, determining a number of clusters on a cluster of said set and determining an indicator of said volume according to said number of arrays on a cluster, said number of clusters on a vine and the diameter of said bays determined from said photographic image and said distance measurement. 20
    [0007]
    7. System according to the preceding claim, wherein the calculation means (MC) are provided for detecting the berries by the determination of circular shapes within said photographic image. 25
    [0008]
    8. System according to one of claims 6 or 7, wherein the calculation means (MC) are provided to determine a number of berries on a cluster by the use of an experimental model linking a number of berries visible on a photographic image at number of berries on a cluster. 30
    [0009]
    9. System according to one of claims 6 to 8, wherein said calculating means (MC) are provided for determining the number of clusters on a 3036830 16 cep according to an abacus depend on the variety of said vine and technique vine held by the operator of said winery. 5
    [0010]
    10. System according to one of claims 6 to 9, further comprising a locating device (DL) for rendering said harvest volume dependent on a geographical location.
    [0011]
    11. Vehicle comprising a system according to one of claims 6 to 10.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3304416A1|2018-04-11|System and method for estimating a harvest volume in a vineyard operation
    Wang et al.2013|Automated crop yield estimation for apple orchards
    van der Heijden et al.2012|SPICY: towards automated phenotyping of large pepper plants in the greenhouse
    EP3226672B1|2018-11-28|System and method for the automatic adjustment of the height of an agricultural implement using 3d reconstruction
    FR2977122A1|2013-01-04|METHOD OF CORRECTING TIME SHIFT FOR AGRICULTURAL YIELD MEASUREMENT
    FR3020915A1|2015-11-20|METHOD FOR IMPLEMENTING A SOIL TREATMENT ROBOT AND CORRESPONDING SYSTEM
    FR2944111A1|2010-10-08|METHOD AND SYSTEM FOR ENTERING THE ENVIRONMENT
    FR2974218A1|2012-10-19|ANALYSIS OF THE DIGITAL IMAGE OF THE SURFACE OF A TIRE - TREATMENT OF NON-MEASUREMENT POINTS
    FR3003377A1|2014-09-19|PROCESS FOR ANALYZING A CULTIVATED AGRICULTURAL PLANT
    EP2043044A1|2009-04-01|Method and device for automobile parking assistance
    Nielsen et al.2012|Orchard and tree mapping and description using stereo vision and lidar
    Wang et al.2012|Design of crop yield estimation system for apple orchards using computer vision
    EP2986981B1|2018-10-31|System for characterising the physiological state of plants and corresponding method
    EP3555562A1|2019-10-23|Crop scanner
    WO2012117210A1|2012-09-07|Method and system for estimating a similarity between two binary images
    FR3025683A1|2016-03-11|METHOD AND DEVICE FOR ESTIMATING DIGITAL IMAGE EQUALIZATION PARAMETERS
    EP0559594B1|1999-06-09|Process for creating the signature of an object represented on a digital image, from the type consisting in defining at least one caracteristic dimensional calibre of that object and corresponding process for verifying the signature of an object
    EP2877979B1|2019-09-18|Monocamera method of determining a direction of a solid
    Andújar et al.2015|Using depth cameras for biomass estimation–a multi-angle approach
    EP3918904A1|2021-12-08|Automated method for selecting and harvesting fruit and mechanical equipment implementing the method
    CN107529726A|2018-01-02|Check device, inspection method and program
    FR2985028A1|2013-06-28|Device for measuring directional transmission of light in vegetation cover for determining index surface of cover, has sensor whose collecting solid angle covers zenith angle and azimuth angle ranging between specific degree
    EP3114459A1|2017-01-11|Method and system for characterising a state of adhesion of particles such as cells
    Enciso et al.2016|Unmanned Aerial System | for Precision Agriculture and Management Decisions
    FR3077721A1|2019-08-16|METHOD FOR CONTROLLING THE OPERATION OF A SOIL TREATMENT ROBOT, EQUIPMENT AND SYSTEM FOR IMPLEMENTING SAID METHOD
    同族专利:
    公开号 | 公开日
    US10672138B2|2020-06-02|
    US20180158207A1|2018-06-07|
    WO2016193602A1|2016-12-08|
    FR3036830B1|2017-06-09|
    EP3304416A1|2018-04-11|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    WO2016004026A1|2014-06-30|2016-01-07|Carnegie Mellon University|Methods and system for detecting curved fruit with flash and camera and automated image analysis with invariance to scale and partial occlusions|CN108171693B|2017-12-27|2020-09-11|合肥市雅视智能科技有限公司|Method for automatically detecting inferior mushrooms|
    US10891482B2|2018-07-10|2021-01-12|Adroit Robotics|Systems, devices, and methods for in-field diagnosis of growth stage and crop yield estimation in a plant area|
    WO2020022215A1|2018-07-27|2020-01-30|キヤノン株式会社|Information processing device, information processing method, and program|
    JP6744898B2|2018-10-12|2020-08-19|ソニーセミコンダクタソリューションズ株式会社|Measuring device, measuring method, program|
    法律状态:
    2016-03-23| PLFP| Fee payment|Year of fee payment: 2 |
    2016-12-02| PLSC| Publication of the preliminary search report|Effective date: 20161202 |
    2017-03-20| PLFP| Fee payment|Year of fee payment: 3 |
    2018-03-21| PLFP| Fee payment|Year of fee payment: 4 |
    2019-05-29| PLFP| Fee payment|Year of fee payment: 5 |
    2020-05-27| PLFP| Fee payment|Year of fee payment: 6 |
    2022-02-11| ST| Notification of lapse|Effective date: 20220105 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1554862A|FR3036830B1|2015-05-29|2015-05-29|SYSTEM AND METHOD FOR ESTIMATING A HARVEST VOLUME IN A WINE FARMING|FR1554862A| FR3036830B1|2015-05-29|2015-05-29|SYSTEM AND METHOD FOR ESTIMATING A HARVEST VOLUME IN A WINE FARMING|
    US15/578,062| US10672138B2|2015-05-29|2016-05-30|System and method for estimating a harvest volume in a vineyard operation|
    PCT/FR2016/051279| WO2016193602A1|2015-05-29|2016-05-30|System and method for estimating a harvest volume in a vineyard operation|
    EP16733646.0A| EP3304416A1|2015-05-29|2016-05-30|System and method for estimating a harvest volume in a vineyard operation|
    [返回顶部]