Cross-category fruit data set automatic labeling system and method

Abstract

The invention provides a cross-category fruit data set automatic labeling method, which comprises the steps of S1, acquiring a source domain fruit image and a target domain fruit image, labeling the source domain fruit image, and constructing a labeled source domain fruit data set; and S2, based on the labeled source domain fruit data set, using a generative adversarial network and a fruit detection network to realize detection of cross-category fruit images, and obtaining target fruit detection box information in target domain fruit image data. The invention further provides a cross-category fruit data set automatic labeling system. According to the method, the function of automatically labeling different types of unlabeled fruit data sets can be realized, and the method can be applied to labeling work of different types of fruit data sets, so that the labor cost is saved, and the working efficiency is improved.

Description

technical field [0001] The invention belongs to the fields of computer vision and intelligent agriculture, and in particular relates to an automatic labeling system and method for cross-category fruit data sets. Background technique [0002] In recent years, with the rapid development of modern information technology, the demand for the application of computer technology in the agricultural field is also increasing, and smart agriculture is being valued and developed. Among them, in modern smart orchards, intelligent cultivation and machine operation are effective ways for sustainable fruit industry development, which is conducive to reducing economic and environmental costs and improving orchard production efficiency, and is the core part of smart agriculture. In the natural environment of orchards, high-performance fruit visual inspection technology is an important technical basis for fruit positioning, fruit yield estimation, and fruit automatic picking. Therefore, wheth...

AI Technical Summary

Problems solved by technology

Since most of the target fruits in the image are small and dense, labeling is difficult, and a lot of manual labor is required for data labeling. The whole process is time-consuming, labor-intensive and inefficient.

Images