Hybrid Object Detection Using Domain-Specific Datasets

Abstract

This paper describes a combination of color determination and object detection. It describes the creation of a hybrid system that would increase production and streamline the process of crop harvesting. The system aims to delineate all potential crops by determining color. If the potential crops are of the sufficient size then object detection is performed using YOLO technology which determines the confidence of strawberry prediction. The main part is the analysis and the implementation of this hybrid system in Python. The last part of the paper is devoted to the evaluation and verification of the created system.