User-orientated approach to developing modular video analysis systems for engineering research applications

This paper introduces an enhanced method for developing video analysis tools, in the context of engineering research applications. It aims to expand the usual boundaries for design of image analysis systems to include explicitly ease of use, maintenance and upgradeability by researchers who have little or no background in image analysis. This challenge moves the emphasis away from the design of optimally efficient (but esoteric) tools for the processing of video- and image-based data, and shifts it towards the development of tools that perform satisfactorily, while being easy to understand, maintain and develop. In the proposed design methodology, this balance is struck using popular commercial software that has a thoroughly tried and tested user interface and yet supports modular expansion through the use of 'plug-ins'. This approach eliminates the need to deal with the vast majority of image analysis infrastructure, while focusing attention on the development of the code at the core of the required image analysis functionality. The development of this core code is then simplified through the selection of a plug-in creator tool that opts for visual programming 'in lieu' of procedural programming as far as possible. In order to evaluate the effectiveness of this proposed new system design approach, a case study is investigated that involves developing a solution to a real image analysis problem faced by research colleagues at the University of Brighton. This concerns a study of the air-side fouling phenomenon of compact heat exchangers by large particles. The implemented system delivers close monitoring of a radiator's surface so as to track the clogging mechanism during the execution of a series of wind-tunnel experiments, which attempt to replicate the fouling phenomenon realistically. The data extracted are presented in the form of several graphs, which describe the progress of these experiments, and have made a definite contribution to the work of the primary researchers. Primary researchers report that they have found the system easy to use, and have adapted it to suit their needs better.