This report presents the conceptual design of a new European research infrastructure EuPRAXIA. The concept has been established over the last four years in a unique collaboration of 41 laboratories within a Horizon 2020 design study funded by the European Union. EuPRAXIA is the first European project that develops a dedicated particle accelerator research infrastructure based on novel plasma acceleration concepts and laser technology. It focuses on the development of electron accelerators and underlying technologies, their user communities, and the exploitation of existing accelerator infrastructures in Europe. EuPRAXIA has involved, amongst others, the international laser community and industry to build links and bridges with accelerator science — through realising synergies, identifying disruptive ideas, innovating, and fostering knowledge exchange. The Eu-PRAXIA project aims at the construction of an innovative electron accelerator using laser- and electron-beam-driven plasma wakefield acceleration that offers a significant reduction in size and possible savings in cost over current state-of-the-art radiofrequency-based accelerators. The foreseen electron energy range of one to five gigaelectronvolts (GeV) and its performance goals will enable versatile applications in various domains, e.g. as a compact free-electron laser (FEL), compact sources for medical imaging and positron generation, table-top test beams for particle detectors, as well as deeply penetrating X-ray and gamma-ray sources for material testing. EuPRAXIA is designed to be the required stepping stone to possible future plasma-based facilities, such as linear colliders at the high-energy physics (HEP) energy frontier. Consistent with a high-confidence approach, the project includes measures to retire risk by establishing scaled technology demonstrators. This report includes preliminary models for project implementation, cost and schedule that would allow operation of the full Eu-PRAXIA facility within 8—10 years.

The laser-plasma accelerator (LPA) presently provides electron beams with a typical current of a few kA, a bunch length of a few fs, energy in the few hundred MeV to several GeV range, a divergence of typically 1 mrad, an energy spread of the order of 1%, and a normalized emittance of the order of π.mm.mrad. One of the first applications could be to use these beams for the production of radiation: undulator emission has been observed but the rather large energy spread (1%) and divergence (1 mrad) prevent straightforward free-electron laser (FEL) amplification. An adequate beam manipulation through the transport to the undulator is then required. The key concept proposed here relies on an innovative electron beam longitudinal and transverse manipulation in the transport towards an undulator: a 'demixing' chicane sorts the electrons according to their energy and reduces the spread from 1% to one slice of a few ‰ and the effective transverse size is maintained constant along the undulator (supermatching) by a proper synchronization of the electron beam focusing with the progress of the optical wave. A test experiment for the demonstration of FEL amplification with an LPA is under preparation. Electron beam transport follows different steps with strong focusing with permanent magnet quadrupoles of variable strength, a demixing chicane with conventional dipoles, and a second set of quadrupoles for further focusing in the undulator. The FEL simulations and the progress of the preparation of the experiment are presented.

No abstract

Rapid and automated identification of blight disease in potato will help farmers to apply timely remedies to protect their produce. Manual detection of blight disease can be cumbersome and may require trained experts. To overcome these issues, we present an automated system using the Mask Region-based convolutional neural network (Mask R-CNN) architecture, with residual network as the backbone network for detecting blight disease patches on potato leaves in field conditions. The approach uses transfer learning, which can generate good results even with small datasets. The model was trained on a dataset of 1423 images of potato leaves obtained from fields in different geographical locations and at different times of the day. The images were manually annotated to create over 6200 labeled patches covering diseased and healthy portions of the leaf. The Mask R-CNN model was able to correctly differentiate between the diseased patch on the potato leaf and the similar-looking background soil patches, which can confound the outcome of binary classification. To improve the detection performance, the original RGB dataset was then converted to HSL, HSV, LAB, XYZ, and YCrCb color spaces. A separate model was created for each color space and tested on 417 field-based test images. This yielded 81.4% mean average precision on the LAB model and 56.9% mean average recall on the HSL model, slightly outperforming the original RGB color space model. Manual analysis of the detection performance indicates an overall precision of 98% on leaf images in a field environment containing complex backgrounds.