0

I'm developing and application for a learning company. I need to image analysis of Test sheet (scanned) and analysis the answer of students. Answer sheets are like this picture : test answer sheet

there are 4 options for every test. I think there is components ready for that. I searched a lot and find different image processing stuffs but nothing about this one. Where can I find tutorial or information or component for this? I mean any good website or source to start learning the solutions? I'm using Delphi but I can change it to C# if it is Necessary.

4

1 回答 1

0

第 1 步:使用像 Canny 这样的边缘检测器来检测边缘...使用 findContour 选择最大的块,如下图绿色所示...使感兴趣的区域或 ROI 以供将来处理..绿边

第 2 步:取两个正方形模板(学生应该在其中打勾或变暗)..一个代表打勾/变暗另一个只是一个空白方块...在 ROI 中运行这两个模板的模板匹配以获得可能的位置的答案。

第 3 步:由于答卷的布局应该是不变的,因此您可以在之前将答卷网格化,并将其用作先验知识,将模板匹配位置映射到实际答案。在此处输入图像描述

于 2013-03-10T12:32:22.033 回答