問卷輸入系統最佳攝距參數自動調整功能研究

Study of the Automatically Adjusting Function on the Optimal Parameter of Shooting Range for Questionnaire Input System

戴辰倫(Chen-Lun Tai)

指導教授 ： 蒲永仁

長榮大學/健康科學學院/職業安全與衛生學系(所)/碩士(2015年)

https://doi.org/10.6833/CJCU.2015.00164

摘要

問卷調查是現今各類科學研究蒐集資料的方法之一，在職業安全衛生領域中扮演著相當重要的角色。近期許多研究使用網路問卷方式進行資料的蒐集，雖然此種方式能夠很方便迅速的在受測者填答完成後立即統計，但也遭遇到許多問題，如受測對象、受測者方便性等。傳統的紙本問卷填答方式，依然為目前研究者最常使用的方法，為求分析的數據能夠有足夠的準確性及代表性，往往有相當龐大的問卷數據量，若要進行資料輸入是相當大的考驗，因此本研究利用機器視覺問卷輸入影像系統進行問卷輸入辨識，改善其必須手動調整攝距最佳參數的缺點，使此系統能夠發揮更大之功效。
　　本論文問卷系統最佳攝距參數自動調整功能，改善了原系統只將攝距固定在特定長度，可能會產生框心誤差之缺點，能夠尋找出定位標距離與最佳攝距參數之關係，利用迴歸分析產生方程式，最後進行程式設計，整合入原系統以改善其缺點。此外，程式設計並新增辨識比例調整功能，辨識實驗結果顯示，自動攝距調整與辨識比例二種功能，能有效降低ROI框心與答案框心間之距離誤差像素值，使系統能夠大幅增加問卷辨識之準確性，並使系統在使用上更加便利。

關鍵字

問卷辨識 ； 機器視覺 ； 安全衛生 ； 影像分析 ； 影像量測

並列摘要

Questionnaire Survey is one of the commonly used methods to collect data in various scientific researches. It plays important roles especially in the research field of occupational safety and health. Recently, many studies use internet questionnaires to do the work. Although this approach can easily and quickly perform statistics immediately after the respondents finish their answers, it might encounter some problems including the appropriateness of subjects and the inconvenience of the respondents, etc. Currently, the traditional questionnaire surveys with paper are still frequently used by researchers. For the sake of data accuracy and representativeness, people often administer a huge amount of questionnaires. When they are collected, the data input becomes a realistic burden. Therefore, this study continues developing the automatic questionnaire input system by machine vision to improve its weakness that the shooting range parameter has to be preset by hand, so that the system is able to demonstrate a greater effect.
　　This thesis developed an automatically adjusting function for the optimal parameter of the shooting range for the existing questionnaire input system. It improved the original system from inevitable errors in locating checkbox caused by the fixed shooting range parameter. In order to find the relation between the distances of the locating marks and the optimal shooting range parameters, linear and higher order equations were generated by regression. Subsequently, the optimal regression equation was determined and integrated into the system. Moreover, a function of recognition ratio was introduced to the system to cope with the varying shooting range as well. The experiment results showed that both automatic shooting range and recognition ratio adjustments are able to reduce the errors between the ROI centers and the checkbox centers. Therefore, the modified system will not only increase greatly the accuracy of the questionnaire input, but also provide convenience to software usage.

並列關鍵字

Questionnaire ； Machine Vision ； Safety and Health ； Image Process ； Image Measurement

參考文獻

[18] 張玉其， 機器視覺問卷輸入影像辨識研究，長榮大學職業安全與衛生學系碩士論文，中華民國九十九年一月，2010。

[10] 董思廷，機器視覺問卷輸入影像系統開發，長榮大學職業安全與衛生學系碩士論文，中華民國一百零一年六月，2012。

[15] 李俊瑋，垂直移動式單CCD光學座標量測系統之量測精度研究，大同大學機械工程研究所碩士論文，中華民國九十九年七月，2010。

[2] David D. S. Poor, Image capture and storage techniques in association with optical mark reading. United States Patent 1995; 5,452,379.

[1] Abraham N. Oppenheim, Questionnaire Design, Interviewing and Attitude Measurement, Lundon: Continuum, 2001.