US 12,608,934 B2Granted

System and method for estimating dynamic soil parameters based on multispectral or hyperspectral image

AssigneeSmartcloudfarming GmbH

InventorsMichele Bandecchi, Nenad Antonic, Marko Pavlovic, Gavrilo Bosakov, Dubravko Culibrk

Life of this patent

Estimating the term…

Loading patent…

Abstract

The invention proposes a new system and method for estimating dynamic soil parameters based on multispectral or hyperspectral images. The system 105 has two-stage estimation architecture and a fully connected layer 108 that connects the two estimators. The system estimates dynamic parameters such as organic carbon in the soil based on multispectral or hyperspectral images 102 and also uses geocoordinates 100 and soil elevation 101 as inputs. Multispectral or hyperspectral images 102 are further processed in the pre-processing module 103 and the data 104 is transmitted to the training neural network 107 of the 1st degree estimation module 106. It further processes them in submodule 108 for the extraction of latent features, estimates the low-dynamic parameters of the soil and forwards them to the output. Also, the output data of submodule 108 is the input data of module 2 for the 2nd degree estimation which estimates the highly dynamic parameters. The low-dynamic parameters 112 are e.g. soil texture, and highly dynamic parameters 113 are e.g. the concentration of organic carbon in the soil. Submodule 108 connects said modules 106 and 110.

Claims20

Dependent claims

A system for estimating dynamic soil parameters, where said system comprises:

a) a pre-processing module;
b) an estimation system comprising:
i) a first stage estimator module comprising a neural network, a reduction submodule, and a submodule;
and
ii) a second stage estimator module;
wherein the system is configured for:
(i) receiving input data from a database;
(ii) pre-processing data in the pre-processing module to produce preprocessed data;
and
(iii) forwarding data to an estimation system, which processes the data and forwards them to an output, wherein a) the reduction submodule is configured for extracting one or more latent features, b) the submodule is configured for estimating one or more low-dynamic soil parameters;
c) the neural network is configured to be trained with data received from the pre-processing module and in the next to the last layer of the neural network d) the submodule is configured for reducing the dimensionality of the neural network and for extracting the one or more latent features and for passing these on to the submodule and also passed on to the second stage module estimator, e) the submodule is configured for estimating low-dynamic soil parameters, which are then forwarded to the output;
f) the second stage module estimator is configured to be trained with the latent features and input data and for estimating highly dynamic parameters for forwarding the highly dynamic parameters to the output;
wherein said submodule configured for reduction and extraction of latent features is a fully connected layer which connects the first stage estimator module and the second stage estimator module.

10 more elements

Dependencies are inferred from the claim text.