Fuzzy pollen particle picture processing method and device and electronic equipment

Abstract

The invention provides a fuzzy pollen particle picture processing method and device and electronic equipment, and the method comprises the steps: inputting a to-be-processed fuzzy pollen particle picture into a generative adversarial network model, and outputting a clear pollen particle picture, wherein the generative adversarial network model is obtained after unsupervised training by taking a fuzzy pollen particle picture and a clear pollen particle picture corresponding to the fuzzy pollen particle picture as sample data, and the clear pollen particle picture corresponding to the blurred pollen particle picture is the clear pollen particle picture with the maximum feature similarity with the blurred pollen particle picture in different types of clear pollen particle pictures. The methodand device are advantaged in that construction cost of the real paired data is saved, a problem that the paired data is difficult to acquire is solved, data distribution inconsistency between the synthetic data and the real data is effectively avoided, and a deblurring effect of the model on the real blurred data is improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method, device and electronic equipment for processing fuzzy pollen grain images. Background technique [0002] In recent years, the expansion of plant planting areas has become an indispensable part of urbanization in the process of urbanization. However, with the expansion of the planting area, the pollen of a large number of allergenic plants floating in the air will greatly increase the incidence of pollen allergy patients, which will seriously affect the daily life of allergy patients. However, effective pollen concentration forecasting can greatly reduce morbidity in allergic patients. However, an important prerequisite for effective pollen concentration forecasting is the ability to accurately identify pollen grains, and clear pollen grain pictures are the basis for pollen grain identification tasks. Therefore, the deblurring task of pollen grain images has be...

AI Technical Summary

Problems solved by technology

[0003] Pollen particle image deblurring is a kind of image restoration. At present, most image deblurring algorithms require paired data to train the model, but the acquisition cost of paired data is relatively high, and in many cases, such a component does not even exist. For data, so the common solution for this type of algorithm is to use the method of synthesizing data pairs to construct paired data for model training

[0004] However, although deblurring algorithms using synthetic data pairs can solve the problem of unavailability of paired data to a certain extent, the performance of such algorithms on real data is not satisfactory.

The reason is that there is a real data distribution difference between the synthetic data and the real data. The existence of this difference leads to the final training result of the model only fitting the data distribution of the synthetic data better, but the fitting effect of the real data is not good. Poor, resulting in poor deblurring effect of the model on real blurred data

Images