In this article, we present a new spherical image representation, called uniform spherical mapping of omnidirectional images (UniphorM), and show its strong potential in robotic vision. UniphorM provides an accurate and distortion-free representation of a 360-degree image, by relying on multiple subdivisions of an icosahedron and its associated Voronoi diagrams. The geometric mapping procedure is described in detail, and the tradeoff between pixel accuracy and computational complexity is investigated. To demonstrate the benefits of UniphorM in real-world problems, we applied it to direct visual attitude estimation and visual place recognition (VPR), by considering dual-fisheye images captured by a camera mounted on multiple robotic platforms. In the experiments, we measured the impact of the number of subdivision levels of the icosahedron on the attitude estimation error, time efficiency, and size of convergence domain of an existing visual gyroscope, using UniphorM and three competing mapping algorithms. A similar evaluation procedure was carried out for VPR. Finally, two new omnidirectional image datasets, one recorded with a hexacopter, calledSVMIS+, the other based on theMapillaryplatform, have been created and released for the entire research community.

At small scales, automating robotic tasks such as assembly, force/displacement characterization, positioning, etc., appear to be particularly limited. This is due to the lack of sufficiently performing and easy-to-implement multi-degrees-of-freedom measurement systems able to measure the relative pose between micro-parts. In order to address this issue, a measurement method based on High-Precision fiducial markers (named HP code) is proposed. This measurement method combines a periodic pattern (providing high resolution by phase-based computation) with more regular QR codes (bringing versatile implementations and a quick detection). The design and method to efficiently locate these HP codes are presented in this paper. Experimental investigations demonstrate ultra-high resolution: 2 nm and$5 ~\mu $rad along$X,Y$and$\Theta $respectively (i.e. one thousandth of a pixel typically). The method is designed to be scalable as well as self-calibrated and to provide high robustness and high versatility. Two typical challenging applications in the field of microrobotics are automated to demonstrate these disruptive performances and the easy-to-implement capability of the method: (1) the automated assembly of two micro-fluidic chips through visual servoing with an achieved positioning accuracy below 50 nm, and (2) the automated micromechanical characterization of single fibers achieved by the integration of HP codes into a compliant structure enabling simultaneous micro-force and displacement sensing capabilities. These achievements highlight the versatility of the method and open the door to the rapid automation of high-quality robotic tasks at the micro scale. Note to Practitioners—The motivation for this work/study is based on the fact that many application areas are extensively orienting towards microrobotic systems to perform precise tasks with versatility. However, at the micro scale, many disturbances such as the effects of climate change strongly affect this precision. This problem is amplified by the fact that sensors cannot be easily integrated, either by lack of space or by the lack of measurement systems available. Vision-based approaches are widespread at this scale and appear very promising to measure the relative pose between micro-parts. Nevertheless, existing vision-based approaches like digital image correlation are both scale and texture dependent. Due to the lack of space, they are also difficult to use in practice at small scales for high resolution measurement. The main contribution of this paper lies in the capability to achieve ultra-high resolution measurements. For that, a structure based on High-Precision fiducial markers (named HP codes) is proposed and requires few and simple settings while achieving very high resolution both in position and orientation, typically down to one thousandth of a pixel and a few micro radians, respectively. It provides an off-the-shelf solution, versatile, easy to implement and achieves high resolution measurements in the plane (XY$\Theta $). HP codes are applicable to a wide range of applications such as tracking of a component/part of a mobile or deformable system, visual servoing of microrobots, positioning of samples, assembly of components or even mechanical characterization. A free distribution of the library is available online athttps://projects.femto-st.fr/vernier/.