船只目标分类、图像识别研究场景船只图片数据

该数据为二次加工数据，通过船只图片数据，标注得到结构化的图片数据。数据可用于船只目标分类、图像识别研究等场景。1.通过人工点击和半自动标注工具得到船只位置信息。具体做法是：由人工点击采样图片中船舶所在区域的中心点位，由训练好的标注工具对采样图片进行分析，结合人工给出的船区域中心点位信息，给出船所在区域的包围框信息。2.将采样图片和包围框信息进行存储，其中采样图片进行base64编码存储，包围框信息则以相对于图片“左-上-宽-高”的相对位置数据存储。This is a secondary processed dataset, generated from vessel image data through annotation to obtain structured image data. It can be applied to scenarios including vessel target classification and image recognition research. 1. Vessel position information is acquired via manual clicks and semi-automatic annotation tools. The detailed procedure is: manually click the center point of the vessel's region in the sampled images, then use the pre-trained annotation tool to analyze the sampled images, and combine the center point information of the vessel region provided by humans to generate the bounding box information of the vessel's region. 2. The sampled images and bounding box information are stored. The sampled images are stored in base64 encoding, while the bounding box information is stored as relative position data following the "left-top-width-height" format relative to the image.

搜集汇总

数据集介绍

特点

该数据集包含201条船只图片数据，每条数据包含图片名、图像和标注信息，适用于船只目标分类和图像识别研究。