Bošnak Matevž, Klančar Gregor
Faculty of Electrical Engineering, University of Ljubljana, 1000 Ljubljana, Slovenia.
Sensors (Basel). 2020 Jun 23;20(12):3552. doi: 10.3390/s20123552.
Simultaneous determination of linear and angular positions of rotating objects is a challenging task for traditional sensor applications and a very limited set of solutions is available. The paper presents a novel approach of replacing a set of traditional linear and rotational sensors by a small set of image sensors. While the camera's angle of view can be a limiting factor in the tracking of multiple objects, the presented approach allows for a network of image sensors to extend the covered area. Furthermore, rich image data allows for the application of different data processing algorithms to effectively and accurately determine the object's position. The proposed solution thus provides a set of smart visual encoders emulated by an image sensor or a network of image sensors for more demanding spatially distributed tasks. As a proof of concept, we present the results of the experiment in the target application, where a 1.6 MP image sensor was used to obtain sub-degree angular resolution at 600 rpm and thus exceeding the design parameters and requirements. The solution allows for a compact, cost-effective, and robust integration into the final product.
同时测定旋转物体的线性和角位置对传统传感器应用来说是一项具有挑战性的任务，并且可用的解决方案非常有限。本文提出了一种新颖的方法，即用一小部分图像传感器取代一组传统的线性和旋转传感器。虽然相机的视角可能是跟踪多个物体时的一个限制因素，但本文提出的方法允许图像传感器网络扩展覆盖区域。此外，丰富的图像数据允许应用不同的数据处理算法来有效且准确地确定物体的位置。因此，所提出的解决方案提供了一组由图像传感器或图像传感器网络模拟的智能视觉编码器，用于要求更高的空间分布式任务。作为概念验证，我们展示了在目标应用中的实验结果，其中使用了一个160万像素的图像传感器在600转/分钟的转速下获得亚度角分辨率，从而超过了设计参数和要求。该解决方案允许紧凑、经济高效且稳健地集成到最终产品中。
