Anki Vector Robot Dataset Object Detection Dataset - v5 (Resize + Grayscale + Rotate + Blur + BB optimizations)

**Background**

The Anki Vector robot (assets currently owned by Digital Dream Labs LLC which bought Anki assets in 2019) was first introduced in 2018. In my opinion, the Vector robot has been the cheapest fully functional autonomous robot that has ever been built. The Vector robot can be trained to recognize people; however Vector does not have the ability to recognize another Vector. This dataset has been designed to allow one to train a model which can detect a Vector robot in the camera feed of another Vector robot.

**Details**
Pictures were taken with Vector’s camera with another Vector facing it and had this other Vector could move freely. This allowed pictures to be captured from different angles. These pictures were then labeled by marking the rectangular regions around Vector in all the images with the help of a free Linux utility called labelImg. Different backgrounds and lighting conditions were used to take the pictures. There is also a collection of pictures without Vector.

**Example**
An example use case is available in my Google Colab notebook, a version of which can be found in my [Git](https://github.com/hiamitabha/course/blob/master/notebooks/TrainVectorWithYOLOv5.ipynb).

**More**
More details are available in [this ](https://programmablerobots.substack.com/p/teaching-the-vector-robot-to-detect-another-vector-robot-a04bf88603c9) article on my [blog](https://programmablerobots.substack.com/).
If you are new to Computer Vision/ Deep Learning/ AI, you can consider my course on '[Learn AI with a Robot](https://robotics.thinkific.com/courses/aiwitharobot)' which attempts to teach AI based on the AI4K12.org curriculum. There are more details available in [this](https://medium.com/programming-robots/learn-ai-with-a-robot-8de7c5779972) post.

背景
Anki Vector机器人（目前由Digital Dream Labs LLC拥有，该公司于2019年收购了Anki的资产）于2018年首次推出。据我之见，Vector机器人是迄今为止所构建的最经济的功能完备的自主机器人。Vector机器人可以被训练以识别人类；然而，Vector不具有识别另一台Vector的能力。本数据集旨在允许训练一个模型，该模型能够检测到另一台Vector机器人摄像头中的Vector机器人。
详情
使用Vector机器人的摄像头拍摄了图片，另一台Vector面向它，且该Vector可以自由移动。这允许从不同的角度捕捉图片。随后，利用名为labelImg的免费Linux实用工具在所有图像中标注了围绕Vector的矩形区域以对这些图片进行标记。使用了不同的背景和光照条件来拍摄图片。此外，还收集了一组没有Vector的图片。
示例
我的Google Colab笔记本中提供了一个示例用例，其版本可在我的[Git](https://github.com/hiamitabha/course/blob/master/notebooks/TrainVectorWithYOLOv5.ipynb)上找到。
更多
有关此[文章](https://programmablerobots.substack.com/p/teaching-the-vector-robot-to-detect-another-vector-robot-a04bf88603c9)的更多细节可在我的[博客](https://programmablerobots.substack.com/)上找到。如果您是计算机视觉/深度学习/人工智能的新手，可以考虑我的“[通过机器人学习人工智能](https://robotics.thinkific.com/courses/aiwitharobot)”课程，该课程尝试基于AI4K12.org课程体系教授人工智能。更多详情可在[这篇](https://medium.com/programming-robots/learn-ai-with-a-robot-8de7c5779972)帖子中找到。