Nature inspired image processing software based on Maxwell Equations

计算机视觉是自动化和机器人学的重要组成部分，因为它允许人们分析图像并提取有用信息，用于复杂环境中的决策、三维操作、复杂零件的现代制造和自主三维导航。*近年来，在视频监控和汽车行业中快速增长的使用方面，CV也非常存在于我们的日常生活中。**这项研究的目的是开发一个演示软件，向业界展示我们新开发的抓取算法的出色性能，以及对仍在开发中的图像分割算法的优化、改进和实现。这些算法是自然启发的，因为它们是基于麦克斯韦电磁方程的拓扑性质的使用，通过将像素转换为带电粒子或偶极子。**电磁图像分析的优势在于它可以实时使用，可以分析其他技术可能无法识别的特征。这些特征包括轮廓概率的计算、远距离轮廓交互以及独立于图像的大小、分辨率和旋转的特征检测。最后，提出的方法不需要训练数据，这在期望机器人工作的高度变化的环境中应该被证明是一个优势。****

Computer vision (CV) is an important part of automation and robotics, as it allows one to analyze an image and extract useful information for decision-making in complex environments, 3-D manipulations, modern manufacturing of complex parts and autonomous 3D Navigation. **In the recent year CV also became very present in our daily lives in terms of video surveillance and through its fast growing use in the automotive industry. **The objective of this research proposition is to develop a demonstration software to showcase for the industry the outstanding performances of our newly developed grasping algorithm as well as the optimization, improvement and implementation of an image segmentation algorithm still in development. These algorithms are nature inspired, as they are based on the use of the topological properties of Maxwell's equations of electromagnetism by converting pixels into charged particles or dipoles. **The advantage of electromagnetic image analysis is that it can be used in real time and can analyze characteristics that might not be recognized by other techniques. These characteristics include the computation of contour probabilities, long-distance contour interactions, and feature detection independent of the size, resolution and rotation of the image. Finally, the proposed method functions without the need of training data, which should prove to be an advantage in highly changing environments where robots are expected to work. ****