Method and system for dividing plant organ point cloud

Patent metadata

Field	Value
Publication number	US-9547793-B2
Application number	US-201314765806-A
Country	US
Kind code	B2
Filing date	Sep 13, 2013
Priority date	Apr 19, 2013
Publication date	Jan 17, 2017
Grant date	Jan 17, 2017

Abstract

Disclosed are a method and system for dividing a plant organ point cloud. According to information about lamina point cloud blocks and stem point cloud blocks in a previous adjacent point cloud frame of a point cloud frame to be divided and a flatness feature of a data point of the point cloud frame to be divided, a point cloud of the point cloud frame to be divided is divided into lamina point cloud blocks and stem point cloud blocks; taking the adjacent point cloud frame as a reference to divide the point cloud frame facilitates the divided organ in maintaining the consistency of time and space between adjacent point cloud frames, and improves the accuracy of plant organ division. When a new lamina point cloud block or a stem point cloud block is discovered, the point cloud in the point cloud frame before the point cloud block to be divided is redivided through tracking back; when redividing, according to information about the lamina point cloud blocks and the stem point cloud blocks in the next adjacent point cloud frame, the point cloud frame to be redivided is divided; and the redivision during tracking back facilitates extracting, from the point cloud frame, a new lamina point cloud block or a stem point cloud block which was not discovered in the previous division process, thereby further improving the accuracy of plant organ division.

First claim

What is claimed is:

1. A method for dividing a plant organ point cloud, comprising the following steps S 120 -S 160 :
S 120 : acquiring point cloud frames corresponding to multiple plant growth times of a monitored plant, and denoting the acquired multiple point cloud frames as F 1 , . . . , F t , . . . , F n according to a sequence of the corresponding plant growth times;
S 140 : dividing a point cloud in the first point cloud frame F 1 into a lamina point cloud block and a stem point cloud block according to a flatness feature of a data point in the first point cloud frame F 1 ; and
S 160 : dividing the point cloud frames F 2 to F n after the first point cloud frame into lamina point cloud blocks and stem point cloud blocks according to a sequence of corresponding plant growth times, wherein the dividing a point cloud in a certain point cloud frame F t (2≦t≦n) into a lamina point cloud block and a stem point cloud block comprises the following steps S 220 -S 280 :
S 220 : according to information about a lamina point cloud block and a stem point cloud block in a previous point cloud frame F t−1 of F t and a flatness feature of a data point of the F t , dividing the point cloud in the F t into a lamina point cloud block and a stem point cloud block;
S 240 : judging whether the number of the stem point cloud blocks contained in F t is larger than that of the stem point cloud blocks contained in F t−1 or whether the number of the lamina point cloud blocks contained in F t is larger than that of the lamina point cloud blocks contained in F t−1 , if no, performing step S 260 , and if yes, performing step S 280 ;
S 260 : if t<n, entering the step of dividing the point cloud in the next point cloud frame F t+1 of F t into lamina point cloud blocks and stem point cloud blocks, if t=n, ending; and
S 280 : tracking back to successively redivide the point cloud in the previous point cloud frame of the F t until the number of the lamina point cloud blocks and the stem point cloud blocks of the redivided point cloud frame does not increase, and then stopping tracking back and entering the step of dividing the point cloud in the next point cloud frame F t+1 of F t into lamina point cloud blocks and stem point cloud blocks, and the step of redividing the point cloud of a certain point cloud frame F j (1≦j≦t−1) before F t comprises:
S 282 : according to the information about the lamina point cloud blocks and the stem point cloud blocks in the next point cloud frame F j+1 of F j and the flatness feature of the data points of F j , dividing the point cloud in F j into the lamina point cloud blocks and the stem point cloud blocks.

2. The method for dividing a plant organ point cloud according to claim 1 , wherein, if the F t in step S 220 is denoted as a point cloud frame to be divided, the F t−1 is denoted as an adjacent point cloud frame of the point cloud frame to be divided, the F j in step S 282 is denoted as a point cloud frame to be divided, and the F j+1 is denoted as an adjacent point cloud frame of the point cloud frame to be divided,
the step of, according to information about lamina point cloud blocks and stem point cloud blocks in an adjacent point cloud frame of a point cloud frame to be divided and a flatness feature of a data point of the point cloud frame to be divided, dividing a point cloud of the point cloud frame to be divided into lamina point cloud blocks and stem point cloud blocks comprises the following steps S 310 -S 360 :
S 310 : according to the information about lamina point cloud blocks and stem point cloud blocks in the adjacent point cloud frame of the point cloud frame to be divided and the flatness feature of the data point in the point cloud frame to be divided, classifying the data point in the point cloud frame to be divided as a lamina point or a stem point;
S 320 : aggregating lamina points, which are connected with each other, in the point cloud frame to be divided to make up a lamina point cloud block;
S 330 : establishing a minimum spanning tree of stem points in the point cloud frame to be divided, dividing the minimum spanning tree into a separate branch that do not comprise forked stems, and mapping the stem points in the point cloud frame onto respective separate branches, to form respective stem point cloud blocks, wherein one separate branch corresponds to one stem point cloud block;
S 340 : finding out point cloud blocks, which have most overlapping data points, corresponding to each lamina point cloud block and each stem point cloud block in the point cloud frame to be divided and the adjacent point cloud frame, marking the corresponding point cloud blocks, and extracting the lamina point cloud blocks and the stem point cloud blocks, which are marked less than twice, in the point cloud frame to be divided and the adjacent point cloud frame;
S 350 : classifying the lamina points in the point cloud frame to be divided to categories corresponding to the lamina point cloud blocks extracted, and making the lamina points within the same category form one lamina point cloud block in the point cloud frame to be divided; and
S 360 : classifying the stem points in the point cloud frame to be divided to categories corresponding to the lamina point cloud blocks extracted, and making the stem points within the same category form one stem point cloud block in the point cloud frame to be divided.

3. The method for dividing a plant organ point cloud according to claim 2 , wherein step S 310 comprises the following steps:
S 410 : solving to cause a target energy function to reach the least classification function, the target energy function being a function about a classification function, of which the function value denotes classification cost, and the classification function being a classification function that classifies data points of the point cloud frame to be divided into lamina points and stem points;
wherein a data point set in the point cloud frame to be divided is denoted as P t ={p t } and a plant organ data point category set is denoted as {L,S}, wherein L represents a lamina point category, S represents a stem point set, ƒ B is a classification function that maps P t to the set {L,S} and the target energy function is as follows:

E
⁡

(

f
B

)

=

∑

p
t

∈

p
t

⁢

⁢

D

p
t

⁡

(

f
B

⁡

(

p
t

)

)

+

Assignees

• Shenzhen Inst Of Adv Tech Cas

Inventors

• Li Yangyan
• Chen Baoquan
• Fan Xiaochen

Classifications

• G06T7/0079

  Physics · mapped topic

• G06T7/0087

  Physics · mapped topic

• G06K9/00201

  Physics · mapped topic

• G06K9/00657 Primary

  Physics · mapped topic

• G06T7/0093

  Physics · mapped topic