Object Detection

Object detection based on deep learning involves localizing object classes and identifying them by placing a bounding box around them. Deep learning-based object detection enables the detection or counting of objects that are next to each other or overlapping.

In HALCON, users have the option to adapt these bounding boxes according to the orientation of the objects with each other. This leads to more accurate detection. To detect objects, labeled data must be provided in the form of bounding box coordinates. The trained model can confidently detect various instances of objects of different types along with their positions in the image and return a corresponding predicted bounding box and class for each instance.