Click Train New Characters Only if you want to add only new characters to the font.Įnter the String that matches the first line of text containing the characters you wish to train. To Font button Polarity cannot be changed once character models have beenīy default, all characters will be added to the font, even if an instance of one or more of the characters already exists in the font. You set the correct Polarity for your characters before clicking the Add The Training tab allows characters to be extracted from an image to Image Reference: Defines which image the tool will utilize to perform its inspection. Selection Mode to reference an Image structure in the active Live once the desired image has been acquired.ĭialog, which allows you to load an image from an In-Sight sensor, Live: Places the sensor into Live Acquisition Mode to display a Otherwise, you can obtain a valid image using any of the following methods: If the desired image is already loaded as the active image, proceed to the next section. The Image tab allows you to obtain images that you will use to train the font. List and press the OK button or double-click the function name. Select the TrainFont function from the function In the left side of the Insert Function dialog,Įxpand the Vision Tools list of tool categories. Right-click an empty cell and select Insert Multiple-byte characters can be trained, but the character model name given to the character model must be a single byte name. Multiple-byte character model names (e.g.Set the user interface to the correct language for the characters contained in the label and retry the operation or use English characters in the labels. When the user interface Language is set to a value other than Chinese or Korean, OCV/OCR operations (including Delete and Relabel) will not function on character instances if the label contains Chinese or Korean characters.Of character instances needed increases as greater tolerance to variations May require only one instance of each character to be extracted. With little variation in character attributes from one image to the next When either of these limits is surpassed, the In-Sight sensor will ignore any additional character models and they will not be added. Up to 255 character models, as well as 8 instances of a particular character each model is assigned a single-character label.
Tolerance of the corresponding character model in the font. By adding more character instances, you increase the variation Instances of each character may be extracted to train each character model The OCRMax function has enhanced performance capabilities. Cognex recommends that the OCRMax function be used instead of the ReadText, TrainFont and VerifyText functions, which are legacy functions.For a complete list of models and supported firmware versions, see Firmware Versions. The TrainFont, ReadText and VerifyText functions are only available on In-Sight vision systems using In-Sight firmware 4.x.x, and are not available on In-Sight vision systems running In-Sight firmware 5.1.0 and later.
0 kommentar(er)
