HikRobot SC2000 Series User Manual page 56

This tool relies on colors as a template for object classification and recognition. When different types
of objects that have obvious color differences, this tool can be used to accurately classify objects,
and output related classification information.
Steps
1. Click to create the color recognition model.
2. Click Select Current or Select Others to load images.
3. Click to add a label.
4. Click different tool icons to create masks.
5. Click Add to Label to add the sample to Label List.
One type of objects can be placed in one label, and the sample can be moved to the correct label
list when the sample is incorrectly marked. You can set following parameters according to actual
demands.
Sensitivity has 3 modes, including low sensitivity mode, medium sensitivity mode, and high
sensitivity mode. It is recommended to select high sensitivity mode when the image is sensitive to
external environment.
Feature Type includes feature histogram and feature spectrum, and feature histogram is more
sensitive.
Brightness refers to the effect of light on the image. If you need to keep the recognition result more
stable under the change of light, it is recommended to disable this parameter.
The operating parameters of color recognition tool includes K Value and KNN Distance.
K Value means that the category with the largest quantity in the first K samples is selected as the
best recognition result.
KNN Distance includes Euclidean Distance, Manhattan Distance and Intersect Distance. There are
slight differences among these distances, and you can select according to actual conditions, and it
is recommended to use the default value.
SC2000 Series Vision Sensor User Manual
Figure 7-28 Create Model
47