CODEWITHSUNDEEP
image-processingopencv
Jacek
Jacek

Reputation: 1914

OpenCV - best approach to detect multiple similar, but different objects

My objective is to find multiple barcodes from the pictures with OpenCV. I have tried using SURF to find them, but the method is ineffective in that case (features not unique enough). I was also considering HAAR, but it is not a rotation invariant method.

What do you think is the best approach to handle this problem?

(Full size image)

enter image description here

Upvotes: 3

Views: 1906

Answers (1)

LovaBill
LovaBill

Reputation: 5139

Train a Support Vector Machine. Create a dataset with many barcodes in many orientations. When testing an image, build a scale-pyramid and apply a sliding window technique. It's called "Object detection". Also "multiple similar but different" is called intra-class variation.

Edit: Or try this.

Upvotes: 1

Related Questions