一般情况

视觉功能为像素阵列（ROI）的每个连续区域解释一个字。因此，每个区域内都应该有一个不带空格的字。对齐（读取方向）由参数 SearchAngle 定义，适用于所有字词。结果排序为从左到右、从上到下（水平、顺时针方向）。

关于 ROI 的说明

根据指定的 ROI 和方向（SearchAngle），图像内部会自动定义矩形区域（每个单词 1 个），并在其中进行字符识别。为此，ROI 中每个连续区域的最小周边矩形将按照指定的搜索角方向进行计算。因此，这些字符读取区域可能与实际的 ROI 集不同。当 TestExecute = 1 时，将返回这些区域，以便在执行时进行可视化。

关于置信度的说明

此过程变量的值（小或大）并不一定表示字符识别的正确性。此外，该值也不能说明字符的打印效果是好是坏。

特殊特性

如果常数 MaxStringSize 的值大于 27，运行时间会明显延长。另一方面，可以读取字符数相应较多的单词。不过，对图像质量的要求也会大大提高。字体和背景必须具有均匀的灰度值和较高的对比度。

此外，字符应排列在同一方向（阅读方向），且不能彼此直接相邻。否则，结果可能不正确。此外，点印字符也很容易出错。如果单个字符识别效果不佳，也可能是因为 MaxStringSize 值太小。

General

The vision function interprets one word for each contiguous area of the pixel array (ROI). Accordingly, there should be exactly 1 word without spaces in each of these areas. The alignment (reading direction) is defined by the parameter SearchAngle and applies to all words. The results are sorted so that the words are output from left to right and from top to bottom (horizontal, clockwise font).

Notes on the ROI

Based on the specified ROI and orientation (SearchAngle), rectangular areas in the image are automatically defined internally (1 per word) in which character recognition is performed. For this purpose, the smallest surrounding rectangle of each contiguous area of the ROI is calculated in the direction of the specified SearchAngle. Accordingly, these character reading areas may differ from the actual ROI set. The areas are returned for visualization for TestExecute = 1 from an execution.

Notes on Confidence

The value of this process variable (small or large) does not necessarily give an indication of the correctness of the character recognition. In addition, the value does not provide an indication of whether the character is printed well or poorly.

Special characteristics

If the value of the constant MaxStringSize is greater than 27, a significantly longer runtime results. On the other hand, words with a correspondingly large number of characters can be read. However, the requirements for the image quality are significantly higher. The font as well as the background must have homogeneous gray values and a high contrast to each other

In addition, the characters should be arranged in the same orientation (reading direction) and not directly adjacent to each other. Otherwise, the result may be incorrect. In addition, it is very easy for errors to occur in the case of characters with dot print. If the character recognition does not work well enough for individual characters, this may also be due to the MaxStringSize value being too small.

关于深度 OCR 的评论