Title:
CROP YIELD ESTIMATION METHOD BASED ON DEEP TEMPORAL AND SPATIAL FEATURE COMBINED LEARNING
Document Type and Number:
WIPO Patent Application WO/2021/098472
Kind Code:
A1
Abstract:
A crop yield estimation method based on deep temporal and spatial feature combined learning. The method comprises: obtaining historical crop yield data and meteorologic data of a region, and preprocessing the meteorologic data and the yield data to respectively obtain meteorologic parameters and a detrended yield as input and output data of a subsequent crop unit yield temporal and spatial feature deep learning model; constructing the crop unit yield temporal and spatial feature deep learning model, and optimizing a hyperparameter of the model; and forming a training set sample by taking the meteorologic parameters as an input and the detrended yield as an output so as to train the crop unit yield temporal and spatial feature deep learning model to obtain parameters of the model, inputting meteorologic parameters of a crop yield to be estimated into the trained model, and outputting an estimation result, thereby obtaining a crop yield estimation result. According to the method, temporal feature learning and spatial feature learning are combined, and in a research region having large and complex spatial difference, the crop yield estimation precision is higher and the stability is better.
Inventors:
LIN TAO (CN)
ZHONG RENHAI (CN)
XU JINFAN (CN)
JIANG HAO (CN)
YING YIBIN (CN)
DING GUANZHONG (CN)
ZHONG RENHAI (CN)
XU JINFAN (CN)
JIANG HAO (CN)
YING YIBIN (CN)
DING GUANZHONG (CN)
Application Number:
PCT/CN2020/124870
Publication Date:
May 27, 2021
Filing Date:
October 29, 2020
Export Citation:
Assignee:
UNIV ZHEJIANG (CN)
International Classes:
G06Q10/04
Foreign References:
| CN111027752A | 2020-04-17 | |||
| CN109886496A | 2019-06-14 | |||
| CN110443420A | 2019-11-12 |
Attorney, Agent or Firm:
HANGZHOU QIUSHI PATENT OFFICE CO., LTD. (CN)
Download PDF: