Development of an Automated Image Processing Framework for Segmenting Floor Plans for Optimal Placement of Internet Access Points

本提案的目的是开发一种由图像分析技术组成的软件工具，以自动将基于像素的数字化平面图（FP）转换为结构化数据。我们提出的软件将自动提取存在于FP中的信息，并将其表示在墙壁，窗户和门等结构中。该信息可用于创建由FP表示的建筑物的数字模型（DMB）。随着物联网的发展，无线设备的数量趋于增加。通过正确放置无线接入点（AP）可以实现可靠的互联网连接。最佳AP分布可以通过计算机模拟无线电波在建筑物内的传播来获得。为了生成模拟所需的模型，可以使用建筑物的FP。爱立信目前正在开发一种软件来规划建筑物内的AP分布，但FP转换为DMB的性能很差，生成的模型缺少功能和错误识别。在这个项目中，我们将确定结合基于机器学习的最先进的图像分割方法的可行性，以解决将FP转换为DMB的一些关键技术挑战。与我们的工业合作伙伴爱立信加拿大公司合作，我们打算通过开发新的图像分割方法来提高模型创建的准确性和速度。这些方法将确保提高FP转换为DMB的准确性和速度，使爱立信能够在这个竞争激烈的领域处于优势地位。

The objective of this proposal is to develop a software tool consisting of image analysis techniques to automatically convert pixel based digitalized Floor Plans (FPs) into a structured data. Our proposed software will automatically extract the information present into the FPs and represent it in structures like walls, windows, and doors. This information can be used to create a digital model of the building (DMB) represented by the FP. With the grow of the Internet of Things the number of wireless devices tends to increase. A reliable Internet connection can be achieved by correctly placing the wireless access points (APs). The optimal APs distribution can be obtained by computer simulations of the radio waves propagation within a building. To generate the model necessary to the simulation, the FP of the building can be used. Ericsson is currently developing a software to plan the APs distribution within a building, but the FP conversion into a DMB is performing poorly, generating a model with missing features andmisidentifications. In this project, we will determine the feasibility of incorporating state-of-the-art image segmentation methods based on machine learning to address some key technical challenges in converting FPs into DMBs. In conjunction with our industrial partner, Ericsson Canada, we intend to increase the accuracy and speed of this model creation by developing new image segmentation methods. These methods will ensure improved accuracy and speed in conversion of FPs into DMBs that would enable Ericson to be in the edge in this highly competitive field.