CN-116167850-B - Loan risk assessment method and device based on agricultural remote sensing image

CN116167850BCN 116167850 BCN116167850 BCN 116167850BCN-116167850-B

Abstract

The invention provides a loan risk assessment method and device based on an agricultural remote sensing image, and relates to the technical field of artificial intelligence. The method comprises the steps of obtaining a hyperspectral image dataset of a target area, obtaining crop types corresponding to each pixel in the hyperspectral image dataset according to the hyperspectral image dataset of the target area and a crop identification model, obtaining estimated planting areas of each crop according to the crop types corresponding to each pixel and land areas corresponding to each pixel, predicting agricultural total yield values of the target area according to the estimated planting areas, the estimated unit yield and the estimated unit price of each crop, and obtaining loan risk assessment results according to the agricultural total yield values of the target area and loan amounts of farmers in the target area. The device is used for executing the method. The loan risk assessment method and device based on the agricultural remote sensing image provided by the embodiment of the invention improve the accuracy of risk assessment after loan of the agricultural loan.

Inventors

YANG ZHILING
LI ZHONGXING
LUO XINWEI
ZHANG DADONG

Assignees

中国工商银行股份有限公司

Dates

Publication Date: 20260508
Application Date: 20230306

Claims (8)

1. The loan risk assessment method based on the agricultural remote sensing image is characterized by comprising the following steps of: Acquiring a hyperspectral image dataset of a target area; obtaining crop types corresponding to each pixel in the hyperspectral image dataset of the target area according to the hyperspectral image dataset of the target area and a crop identification model, wherein the crop identification model is obtained based on hyperspectral image sample dataset training; obtaining estimated planting areas of each crop according to the crop types corresponding to each pixel and the land areas corresponding to each pixel, and predicting the agricultural total yield value of the target area according to the estimated planting areas, the estimated unit yield and the estimated unit price of each crop; Obtaining a loan risk assessment result according to the agricultural total yield value of the target area and the loan amount of farmers in the target area; The method for obtaining the crop identification model based on the hyperspectral image sample data set training comprises the following steps of: acquiring an initial training sample set from the hyperspectral image sample data set and performing labeling; training to obtain a pre-training model according to the marked initial training sample set and the original model; supplementing an initial training sample set according to a residual training sample set and the pre-training model to obtain an updated initial training sample set, wherein the residual training sample set is obtained after the initial training sample set is removed from the hyperspectral image sample data set; retraining to obtain an intermediate model according to the labeled updated initial training sample set; If the training ending condition is not met, the intermediate model is used as a pre-training model to supplement the initial training sample set and retrain the intermediate model again; The step of supplementing the initial training sample set according to the remaining training sample set and the pre-training model, and the step of obtaining the updated initial training sample set comprises the following steps: Obtaining a pseudo label corresponding to each sample in the residual training sample set according to the residual training sample set and the pre-training model, wherein the pseudo label corresponding to each sample is the crop type corresponding to each sample; Obtaining a first supplementary training set from the residual training sample set according to pseudo tags corresponding to all samples in the residual training sample set, wherein the samples in the first supplementary training set correspond to the same pseudo tag and the number of the samples in the first supplementary training set is the least of the samples corresponding to all the pseudo tags; if the number of samples corresponding to the first supplementary training set is greater than or equal to a supplementary sample threshold value, updating an initial training sample set according to the first supplementary training set; If the number of samples corresponding to the first supplementary training set is smaller than a supplementary sample threshold, a second supplementary training set is acquired from a secondary residual training sample set based on a sample supplementary rule so that the sum of the number of samples corresponding to the first supplementary training set and the number of samples corresponding to the second supplementary training set is greater than or equal to the supplementary sample threshold; And updating an initial training sample set according to the first supplementary training set and the second supplementary training set.
2. The method of claim 1, wherein the sample replenishment rules comprise: And acquiring a sample with a smaller sample optimal suboptimal class value from the secondary residual training sample set.
3. The method of claim 1, wherein the acquiring an initial training sample set from the hyperspectral image sample dataset comprises: And acquiring samples with preset proportions from the hyperspectral image sample data set to form the initial training sample set.
4.A method according to any one of claims 1 to 3, further comprising: and if the loan risk assessment result is that the overdue risk is high, outputting prompt information for predicting overdue loans.
5. Loan risk assessment device based on agricultural remote sensing image, characterized by comprising: the acquisition module is used for acquiring a hyperspectral image dataset of the target area; The identification module is used for obtaining crop types corresponding to each pixel in the hyperspectral image dataset of the target area according to the hyperspectral image dataset of the target area and a crop identification model, wherein the crop identification model is obtained based on hyperspectral image sample dataset training; The prediction module is used for obtaining the estimated planting area of each crop according to the crop type corresponding to each pixel and the land area corresponding to each pixel, and predicting the agricultural total yield value of the target area according to the estimated planting area, the estimated unit yield and the estimated unit price of each crop; The evaluation module is used for obtaining a loan risk evaluation result according to the agricultural total yield value of the target area and the loan amount of farmers in the target area; Wherein the apparatus further comprises: the sample acquisition module is used for acquiring an initial training sample set from the hyperspectral image sample data set and performing labeling; The training module is used for training to obtain a pre-training model according to the marked initial training sample set and the original model; The supplementing module is used for supplementing an initial training sample set according to a residual training sample set and the pre-training model to obtain an updated initial training sample set, wherein the residual training sample set is obtained after the initial training sample set is removed from the hyperspectral image sample data set; The retraining module is used for retraining to obtain an intermediate model according to the labeled updated initial training sample set; The judging module is used for taking the intermediate model as a pre-training model to supplement the initial training sample set and retrain the intermediate model again if the training ending condition is not met, and taking the intermediate model as a crop recognition model if the training ending condition is met; Wherein, supplementary module includes: the first obtaining unit is used for obtaining a pseudo tag corresponding to each sample in the residual training sample set according to the residual training sample set and the pre-training model, wherein the pseudo tag corresponding to each sample is a crop type corresponding to each sample; The first supplementary training set is used for acquiring a first supplementary training set from the residual training sample set according to the pseudo tags corresponding to the samples in the residual training sample set, wherein the samples in the first supplementary training set correspond to the same pseudo tag, and the number of the samples in the first supplementary training set is the least number of the samples corresponding to all the pseudo tags; The first updating unit is used for updating the initial training sample set according to the first supplementary training set if the number of samples corresponding to the first supplementary training set is greater than or equal to a supplementary sample threshold value; The second supplementary training set is obtained from the second supplementary training set based on a sample supplementary rule so that the sum of the number of samples corresponding to the first supplementary training set and the number of samples corresponding to the second supplementary training set is greater than or equal to a supplementary sample threshold value, wherein the sample supplementary rule is preset; And the second updating unit is used for updating the initial training sample set according to the first supplementary training set and the second supplementary training set.
6. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of any of claims 1 to 4 when executing the computer program.
7. A computer readable storage medium, characterized in that it stores a computer program which, when executed by a processor, implements the method of any one of claims 1 to 4.
8. A computer program product, characterized in that it comprises a computer program which, when executed by a processor, implements the method of any one of claims 1 to 4.

Description

Loan risk assessment method and device based on agricultural remote sensing image Technical Field The invention relates to the technical field of artificial intelligence, in particular to a loan risk assessment method and device based on an agricultural remote sensing image. Background Agricultural loans are loans issued by financial institutions such as banks to farmers, who typically make agricultural products, and who pay back the loans after selling agricultural products. After the agricultural loan is issued, a post-loan risk assessment is periodically performed. In the prior art, the post-loan risk assessment of the agricultural loan can be performed in a manual regular investigation mode, but the post-loan risk assessment of the agricultural loan is not accurate enough because the actual agricultural production condition of farmers is difficult to master. Disclosure of Invention Aiming at the problems in the prior art, the embodiment of the invention provides a loan risk assessment method and a loan risk assessment device based on an agricultural remote sensing image, which can at least partially solve the problems in the prior art. In a first aspect, the present invention provides a loan risk assessment method based on an agricultural remote sensing image, including: Acquiring a hyperspectral image dataset of a target area; obtaining crop types corresponding to each pixel in the hyperspectral image dataset of the target area according to the hyperspectral image dataset of the target area and a crop identification model, wherein the crop identification model is obtained based on hyperspectral image sample dataset training; obtaining estimated planting areas of each crop according to the crop types corresponding to each pixel and the land areas corresponding to each pixel, and predicting the agricultural total yield value of the target area according to the estimated planting areas, the estimated unit yield and the estimated unit price of each crop; and obtaining a loan risk assessment result according to the agricultural total yield value of the target area and the loan amount of the farmer in the target area. In a second aspect, the present invention provides a loan risk assessment device based on an agricultural remote sensing image, including: the acquisition module is used for acquiring a hyperspectral image dataset of the target area; The identification module is used for obtaining crop types corresponding to each pixel in the hyperspectral image dataset of the target area according to the hyperspectral image dataset of the target area and a crop identification model, wherein the crop identification model is obtained based on hyperspectral image sample dataset training; The prediction module is used for obtaining the estimated planting area of each crop according to the crop type corresponding to each pixel and the land area corresponding to each pixel, and predicting the agricultural total yield value of the target area according to the estimated planting area, the estimated unit yield and the estimated unit price of each crop; And the evaluation module is used for obtaining a loan risk evaluation result according to the agricultural total yield value of the target area and the loan amount of the farmers in the target area. In a third aspect, the present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the loan risk assessment method based on an agricultural remote sensing image according to any of the embodiments above when executing the program. In a fourth aspect, the present invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the loan risk assessment method based on an agricultural remote sensing image as described in any of the above embodiments. In a fifth aspect, the present invention provides a computer program product, where the computer program product includes a computer program, and when executed by a processor, implements the loan risk assessment method based on an agricultural remote sensing image according to any of the above embodiments. According to the loan risk assessment method and device based on the agricultural remote sensing image, a hyperspectral image dataset of a target area can be obtained, the crop type corresponding to each pixel in the hyperspectral image dataset of the target area is obtained according to the hyperspectral image dataset of the target area and the crop identification model, the estimated planting area of each crop is obtained according to the crop type corresponding to each pixel and the land area corresponding to each pixel, the agricultural total yield of the target area is predicted according to the estimated planting area, the estimated unit yield and the estimated unit price of each crop, the loan risk assessment result is ob