TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

We Built an Urban Intelligence Platform

2 点作者 dfine将近 10 年前

1 comment

mswen将近 10 年前
I looked around the site a bit. Looks like pretty interesting data collection using simple IP cameras and computer vision algorithms to identify unique foot and vehicle traffic and counting and time tagging each one. The business use case that I read - narrowing 4 potential new retail locations down to one due to some logic about foot traffic and their target market seemed plausible. Of course we don&#x27;t really know what would have happened if the store had opened in the higher traffic location but with lower foot traffic and the desired times. Bricks &amp; mortar locations are not A&#x2F;B testable in quite the same cost effective manner as a digital store.<p>Still there is often remarkable strength in doing something as simple as counting - when you are able to do it more accurately and afford to do it continuously over extended periods of time.<p>Question: can the computer vision capture behavior such as speed of gait, stopping and turning to look at the store-front, dwell time before turning away and continuing to walk? Distinguish and code for type of clothing to distinguish between someone on their way to work and someone such as a tourist? Could you distinguish between someone arriving at a store doorway purposefully - like they already knew to come here versus someone who has stumbled upon the store while walking?