duplicate-image-classifier

This module implements a vision service that detects duplicate images from the previous frame. It takes the average pixel-wise difference between the current frame and the previous frame to determine if they are under a certain threshold of similarity.

NOTE: This module will return a classification of "different" if the images are different, and nothing if the images are similar.

API and Namespace Triplet

• API: rdk:service:vision
• Organization Namespace: viam
• Model: viam:duplicate-classifier:duplicate-classifier

Attributes

The following attributes are available for this model:

Name	Type	Inclusion	Description
camera_name	string	Required	The name of the camera configured on your robot.
average_pixel_difference_threshold	float	Optional	The threshold above which an image would be considered different. The range is 0.0 to 255.0. The default value is 5.0.

Configuration

Start by configuring a camera on your robot. Remember the name you give to the camera, it will be important later.

Note

Before configuring your camera or vision service, you must create a robot.

Navigate to the Config tab of your robot’s page in the Viam app. Click on the Services subtab and click Create service. Select the vision type, then select the duplicate-image-classifier model. Enter a name for your service and click Create.

On the new component panel, copy and paste the following attribute template into your base’s Attributes box.

{
    "camera_name": "myCam",
    "average_pixel_difference_threshold": 5.0
}

Example Configuration

{
  "modules": [
    {
      "type": "registry",
      "name": "viam-duplicate-image-classifier",
      "module_id": "viam:duplicate-image-classifier",
      "version": "0.0.1"
    }
  ],
  "services": [
    {
      "name": "myDuplicateImageClassifier",
      "type": "vision",
      "namespace": "rdk",
      "model": "viam:duplicate-image-classifier:duplicate-image-classifier",
      "attributes": {
        "camera_name": "myCam",
        "average_pixel_difference_threshold": 5.0
      }
    }
  ]
}

Usage

The following methods are implemented from the Vision Service API:

• CaptureAllResult()
• GetClassificationsFromCamera()
• GetClassifications()
• GetProperties()

When the module returns classifications, the class_name will always be "different" and the confidence will always be 1.0 if the images are different. Nothing will be returned if the images are similar.