Persimmon image dataset

This persimmon image dataset consists of 800 training set and 200 validation set. This dataset had been used to train CNN for a low-cost machine vision system designed to flip the persimmons automatically in the real persimmon industry.The size of the training images : 416 x 416The size of the validation images : 3024 x 4032 (non-resizing images)

本柿子图像数据集包含800张训练集图像与200张验证集图像。本数据集曾被用于训练卷积神经网络（Convolutional Neural Network，CNN），以搭建面向实际柿子产业的低成本机器视觉系统，实现柿子的自动化翻面作业。训练图像分辨率为416×416；验证图像分辨率为3024×4032（未经过尺寸调整）。