Special Issue on Machine Vision for Outdoor Environments

Context

Recent developments in machine vision technologies have enabled opportunities for automatic outdoor scene analysis to provide useful and advanced capabilities in a number of important sectors, including transport, construction and agriculture. For example, in the latter case, new precision agricultural techniques can particularly benefit from information generated in real-time by machine vision in the field. Here two- and three-dimensional texture and shape analysis can now be employed for improved segmentation in automated weeding, thereby increasing efficiency in weed eradication, leading to improved crop yield and so reducing the environmental impact of herbicides and operational costs for farmers. Similar practical benefits can potentially be obtained in other sectors, and so outdoor imaging research is receiving growing attention from researchers and practitioners within academia and industry, as well as from government organisations seeking to support new and exciting research and development via collaborative project grants. In contrast to the laboratory or the relatively structured setting of a conventional indoor industrial process, outdoor machine vision requires rugged hardware and novel algorithms to address the formidable and challenging issues associated with withstanding the elements and variations in the environment. Common themes when applying machine vision solutions to real-world outdoor problems or in complex settings with only a limited or a complete absence of any form of environmental structuring, often include a need to tolerate harsh environmental conditions (for example heat, vibration, water and dust), an ability to cope with and adapt to uncertainty and change (for example in lighting and in the nature and position of objects) and a capacity to handle and interpret unprecedented quantities of noisy or incomplete data. Solutions may therefore call upon cutting edge aspects of novel hardware design, together with state-of-the-art developments in innovative imaging techniques, including 3D and multispectral, and in data processing, including those techniques that have been rapidly gaining traction in computer vision, such as the internet of things, big data analysis and deep learning.

Aim and scope

Reliable and robust operation of machine vision systems in unstructured and outdoor environments remains a significant and highly topical challenge. This special issue aims to report on the theoretical foundations, novel science and engineering solutions required for the application of machine vision to outdoor or highly unstructured industrial applications. We welcome original, high quality and unpublished manuscripts from academia and industry concerning recent advances in different aspects of outdoor vision research and its application. We expect proposed solutions to be innovative with a particular focus on new and exciting developments within the aims and scope of the journal.

Topics of interest include, but are not limited to:

• Agri-technology – including forestry and timber; crop monitoring / inspection; automated harvesting; weed control; animal monitoring e.g. cattle, pigs and poultry.
• Security and surveillance – including traffic monitoring; facial / demographic recognition; biometrics and directed advertising.
• Transport – including autonomous / self-drive vehicles (on and off road) and drones / UAVs; applications in mass transport passenger assistance / experience; road inspection.
• Scene recognition / interpretation – including object detection / recognition / tracking; outdoor navigation and localisation / positioning.
• Remote sensing and aerial imaging – including for building construction / architecture; applications in the construction industry and construction equipment.
• Assistive technology – including outdoor augmented reality applications; wearable technology; human-computer interaction; embedded computer vision outside the factory; aids for visually impaired.

Submission and Review

Only submissions that are within the scope of the Special Issue and which indicate a clear contribution to the field will be accepted. The full papers will be subjected to a blind review, possibly in two rounds depending on the reviewers’ comments. Submission of full papers will be done through EVISE at https://www.evise.com/profile/#/COMIND/login  Please select Machine Vision for Outdoors article type name when submit contributions to this special issue.

Please follow the instructions for authors found on the Computers in Industry website: https://www.journals.elsevier.com/computers-in-industry

Contributions must present sound applications or validations in a real-world application. Pure theoretical papers will not be accepted.

Time schedule:

• Submission of full paper: 31 August 2017
• Completion of first round reviews: 31 October 2017
• Notifications of first round decisions: 10 November 2017
• Re-submission deadline (if necessary): 20 December 2017
• Completion of second round reviews: 10 February 2018
• Notifications of second round decisions: 20 February 2018
• Camera ready version: 20 March 2018
• Publication of the Special Issue: before 20 April 2018

Special Issue Guest Editors and Contact Information:

Professor Melvyn Smith, Professor of Machine Vision, Bristol Robotics Laboratory, UK,
e-mail: melvyn.smith@brl.ac.uk

Professor Lyndon Smith, Professor in Computer Simulation and Machine Vision Bristol Robotics Laboratory, UK,
e-mail: lyndon.smith@uwe.ac.uk