Development of a Questionnaire Input Software by Machine Vision

Questionnaire survey is one of the important research tools to conduct quantitative information collections. The number of completed questionnaires is usually quite large and sometimes can reach thousands; therefore the data collections become one of the key steps during research. This study set out to find an automatic and vast input method by machine vision, so that the future designs for questionnaires are free to paper materials and question formats. Coded with a visual programming language, the developed software is integrated with a CCD device, wherein the questionnaire images are continuously captured, digitized and processed in real time. The information of all checkboxes is then saves as a text file which can be provided later for further statistic analysis. The experiments determine the thresholds for the chosen and canceled marks in each checkbox, and the rates of recognizing accuracy. Finally, the average processing time for each checkbox is found to be around 8.5 ms.