System and method for analyzing time series growth of crops based on receptacle analysis and tracking

1 . A receptacle analysis and tracking-based time series crop growth analysis method performed by a computer, the method comprising:

acquiring a crop image representing an image of a crop taken by a camera;

recognizing, from the crop image, predetermined unit objects included in the crop;

clustering the predetermined unit objects to form a plurality of clusters;

reconstructing the plurality of clusters on the basis of a receptacle unit; and

generating linkage information between the predetermined unit objects and clusters or between the plurality of clusters in images having different temporal information,

wherein generating the linkage information comprises:

acquiring location information of a unit object among the predetermined unit objects identified from the crop image taken at a first time point;

acquiring location information of the plurality of clusters identified from the crop image taken at a second time point subsequent to the first time point;

extracting a cluster having a closest distance from the plurality of clusters and having a predetermined threshold distance from the location information of the unit object; and

generating the linkage information for the extracted cluster and the unit object.

2 . The method of claim 1 , wherein the recognizing comprises:

recognizing the predetermined unit objects included in the crop from the crop image; and

recognizing the predetermined unit objects as unit objects identified by growth stages.

3 . The method of claim 1 , wherein the clustering comprises:

calculating information about a distance between a first unit object and a second unit object among the predetermined unit objects; and

constructing the first and second unit objects into a cluster in response to the distance information being within a predetermined threshold distance.

4 . The method of claim 1 , wherein the reconstructing comprises:

reconstructing the plurality of clusters on the basis of the receptacle unit in response to the number of unit objects included in the cluster being greater than or equal to the minimum number of unit objects.

5 . The method of claim 1 , wherein the reconstructing comprises:

detecting a stem for any one unit object in the cluster;

detecting a first junction of points where the detected stem extends; and

reconstructing the consecutive unit objects at the first junction on the basis of a single receptacle unit.

6 . The method of claim 5 , wherein reconstructing the consecutive unit objects comprises:

reconstructing the consecutive unit objects at the first junction on the basis of the single receptacle unit in response to the number of unit objects satisfying at least the minimum number of unit objects to constitute the receptacle unit.

7 . The method of claim 5 , wherein detecting the stem comprises:

detecting the stem for the unit object by detecting a straight or curved stem in an upward direction relative to a boundary of the unit object by tracking an energy-maximized path relative to the boundary of the unit object.

8 . The method of claim 1 , further comprising:

separately storing information of the recognized unit object and information of the reconstructed cluster

wherein the information of the unit object includes a unit object index, a class, a location on the crop image, and a crop image acquisition time, and

wherein the information of the cluster includes a cluster index, a list of unit objects in the cluster, information of a location of a cluster center point on the crop image, information of a cluster size on the crop image, and a crop image acquisition time.

9 . The method of claim 1 , wherein extracting the cluster having the closest distance comprises:

comparing a growth stage of the unit object at the first time point with a maximum-growth stage in the cluster having the closest distance; and

extracting the cluster located at the next closest distance in response to the cluster having the closest distance having a maximum-growth stage that is earlier than the growth stage of the unit object.

10 . A receptacle analysis and tracking-based time series crop growth analysis method performed by a computer, the method comprising:

acquiring a crop image representing an image of a crop taken by a camera;

recognizing, from the crop image, predetermined unit objects included in the crop;

clustering the predetermined unit objects to form a plurality of clusters;

reconstructing the plurality of clusters on the basis of a receptacle unit; and

generating linkage information between the predetermined unit objects and clusters or between the plurality of clusters in images having different temporal information,

wherein the generating comprises:

acquiring location information of respective unit objects in a cluster among a plurality of clusters identified from the crop image taken at a first time point;

acquiring location information of a plurality of clusters identified from the crop image taken at a second time point subsequent to the first time point;

extracting a cluster among the plurality of clusters at the second time point having a closest distance and falling within a predetermined threshold distance for the respective unit objects at the first time point; and

generating the linkage information between the extracted cluster and the respective unit objects.

11 . The method of claim 10 , wherein extracting the cluster at the second time point comprises:

comparing a maximum-growth stage of the respective unit objects at the first time point and a maximum-growth stage of the cluster having the closest distance; and

extracting a cluster located at the next closest distance in response to the cluster having the closest distance having a maximum-growth stage earlier than the maximum-growth stage of the unit object.

12 . A receptacle analysis and tracking-based time series crop growth analysis system comprising:

a communication module configured to receive a crop image representing an image of a crop taken by a camera;

a memory storing instructions; and

a processor configured to execute the instructions to:

recognize predetermined unit objects included in the crop from the crop image,

cluster the predetermined unit objects to form a plurality of clusters,

reconstruct the plurality of clusters on the basis of a receptacle unit, and

generate linkage information between the unit objects and the clusters or between the plurality of clusters in images having different time information,

wherein to generate the linkage information, the processor is further configured to execute the instructions to:

acquire location information of a unit object among the predetermined unit objects identified from the crop image taken at a first time point;

acquire location information of the plurality of clusters identified from the crop image taken at a second time point subsequent to the first time point;

extract a cluster having a closest distance from the plurality of clusters and having a predetermined threshold distance from the location information of the unit object; and

generate the linkage information for the extracted cluster and the unit object.