Background Subtraction

Class: BackgroundSubtractionBlockV1

Source: inference.core.workflows.core_steps.classical_cv.background_subtraction.v1.BackgroundSubtractionBlockV1

This block uses background subtraction to detect motion in an image in order to highlight areas of motion. The output of the block can be used to train and infer on motion based models.

Type identifier

Use the following identifier in step "type" field: roboflow_core/background_subtraction@v1to add the block as as step in your workflow.

Properties

Name	Type	Description	Refs
name	str	Enter a unique identifier for this step..	❌
threshold	int	The threshold value for the squared Mahalanobis distance for background subtraction. Smaller values increase sensitivity to motion. Recommended values are 8-32..	✅
history	int	The number of previous frames to use for background subtraction. Larger values make the model less sensitive to quick changes in the background, smaller values allow for more adaptation..	✅

The Refs column marks possibility to parametrise the property with dynamic values available in workflow runtime. See Bindings for more info.

Available Connections

Compatible Blocks

Check what blocks you can connect to Background Subtraction in version v1.

• inputs: Corner Visualization, Color Visualization, Image Slicer, Camera Calibration, Image Blur, QR Code Generator, Line Counter, Image Convert Grayscale, Dot Visualization, Image Threshold, Dynamic Crop, SIFT Comparison, Blur Visualization, Morphological Transformation, Label Visualization, Background Color Visualization, Bounding Box Visualization, SIFT, Classification Label Visualization, Camera Focus, Stability AI Outpainting, Keypoint Visualization, Trace Visualization, Polygon Visualization, Mask Visualization, Pixelate Visualization, Absolute Static Crop, Ellipse Visualization, Model Comparison Visualization, Pixel Color Count, Line Counter, Triangle Visualization, Grid Visualization, Contrast Equalization, Image Preprocessing, Polygon Zone Visualization, Relative Static Crop, Stability AI Inpainting, Halo Visualization, Image Contours, Line Counter Visualization, Stitch Images, Crop Visualization, Stability AI Image Generation, Distance Measurement, Template Matching, Icon Visualization, Circle Visualization, SIFT Comparison, Depth Estimation, Background Subtraction, Camera Focus, Image Slicer, Perspective Correction, Reference Path Visualization
• outputs: Detections Stitch, Seg Preview, Byte Tracker, Qwen3-VL, YOLO-World Model, Google Gemini, Image Convert Grayscale, QR Code Detection, Dynamic Crop, Blur Visualization, SIFT, Stability AI Outpainting, Bounding Box Visualization, Camera Focus, Keypoint Visualization, Trace Visualization, Instance Segmentation Model, Polygon Visualization, Dominant Color, Pixel Color Count, Ellipse Visualization, OpenAI, Model Comparison Visualization, Anthropic Claude, Triangle Visualization, Qwen2.5-VL, SAM 3, Polygon Zone Visualization, Halo Visualization, LMM, Stability AI Image Generation, Time in Zone, VLM as Detector, Florence-2 Model, CLIP Embedding Model, Single-Label Classification Model, Detections Stabilizer, Email Notification, Circle Visualization, Google Vision OCR, Google Gemini, Motion Detection, Clip Comparison, Camera Focus, Anthropic Claude, Object Detection Model, Instance Segmentation Model, Perspective Correction, Perception Encoder Embedding Model, VLM as Classifier, Reference Path Visualization, Corner Visualization, Color Visualization, Twilio SMS/MMS Notification, EasyOCR, Multi-Label Classification Model, Image Slicer, OpenAI, Image Blur, Buffer, Camera Calibration, SmolVLM2, VLM as Detector, SAM 3, Dot Visualization, Image Threshold, Morphological Transformation, Label Visualization, Background Color Visualization, OCR Model, Classification Label Visualization, Roboflow Dataset Upload, Keypoint Detection Model, Mask Visualization, Pixelate Visualization, Absolute Static Crop, Keypoint Detection Model, Moondream2, Image Preprocessing, Contrast Equalization, Google Gemini, Stability AI Inpainting, Barcode Detection, Template Matching, Line Counter Visualization, OpenAI, Image Contours, Crop Visualization, Relative Static Crop, Stitch Images, OpenAI, Llama 3.2 Vision, Icon Visualization, Clip Comparison, SIFT Comparison, Gaze Detection, Depth Estimation, Single-Label Classification Model, VLM as Classifier, Florence-2 Model, Background Subtraction, LMM For Classification, SAM 3, Object Detection Model, Multi-Label Classification Model, Segment Anything 2 Model, CogVLM, Image Slicer, Roboflow Dataset Upload

Input and Output Bindings

The available connections depend on its binding kinds. Check what binding kinds Background Subtraction in version v1 has.

Bindings

• input

  • image (image): The input image for this step..
  • threshold (integer): The threshold value for the squared Mahalanobis distance for background subtraction. Smaller values increase sensitivity to motion. Recommended values are 8-32..
  • history (integer): The number of previous frames to use for background subtraction. Larger values make the model less sensitive to quick changes in the background, smaller values allow for more adaptation..

• output

  • image (image): Image in workflows.

Example JSON definition of step Background Subtraction in version v1

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/background_subtraction@v1",
    "image": "$inputs.image",
    "threshold": 16,
    "history": 30
}