This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
You can use findContours on your binary images to define regions of interest and segment out the existing blobs. Before that you might want to remove some small blobs by applying erosion and dilation functionality.
After you retrieved the different contours, you could use those contours to perform object detection. If your banana plant has a specific look from top, then you can create a cascade classifier for the model. Or you could use a machine learning technique like SVM or random forests.
