Overview The TrashClassificator class is the primary interface for processing video frames to detect, classify, and visualize trash objects. It orchestrates the segmentation model and drawing components to provide a complete trash detection pipeline.
This class automatically initializes both segmentation and drawing components, making it ready to use immediately after instantiation.
Class Definition class TrashClassificator : def __init__ ( self ): self .segmentation: SegmentationModelInterface = SegmentationModel() self .draw_detections: DrawingInterface = Drawing()
Constructor __init__()Initializes the TrashClassificator with segmentation and drawing components.
Attributes Initialized: segmentation
SegmentationModelInterface
Instance of SegmentationModel for trash detection and tracking
Instance of Drawing for visualizing detections on frames
Example: import numpy as np from trash_classificator.processor import TrashClassificator # Initialize the classificator classificator = TrashClassificator()
Methods frame_processing()Processes a single video frame to detect trash objects and draw visualizations.
def frame_processing ( self , image : np.ndarray)
Parameters Input image as a NumPy array in BGR format (OpenCV format). The image should be a valid 3-channel color image.
Returns Processed image with trash detections drawn (if trash detected) or original image (if no trash detected)
Detection status message:
"No trash detected" - No trash objects found in the frame"Trash detected" - Trash objects successfully detected and drawn The method returns early with "No trash detected" status if no tracking IDs are assigned to detected objects.
Implementation Details The method follows a two-step pipeline:
Trash Segmentation : Uses the segmentation model to detect and track trash objectsVisualization : Draws bounding boxes, masks, and tracking lines on detected trash
Source Code Reference: def frame_processing ( self , image : np.ndarray): # step 1: trash segmentation trash_image = image.copy() trash_track, trash_classes, device = self .segmentation.inference(trash_image) for trash in trash_track: if trash.boxes.id is None : return image, 'No trash detected' # step 2: draw detections image_draw = image.copy() image_draw = self .draw_detections.draw(image_draw, trash, trash_classes, device) return image_draw, 'Trash detected' return image, 'Trash detected'
Usage Example Complete example showing how to use TrashClassificator with video processing:
import cv2 import numpy as np from trash_classificator.processor import TrashClassificator # Initialize the classificator classificator = TrashClassificator() # Open video file or camera cap = cv2.VideoCapture( 'video.mp4' ) while cap.isOpened(): ret, frame = cap.read() if not ret: break # Process the frame processed_frame, status = classificator.frame_processing(frame) # Display the result cv2.imshow( 'Trash Detection' , processed_frame) print ( f "Status: { status } " ) if cv2.waitKey( 1 ) & 0x FF == ord ( 'q' ): break cap.release() cv2.destroyAllWindows()
Real-time Camera Processing import cv2 from trash_classificator.processor import TrashClassificator classificator = TrashClassificator() cap = cv2.VideoCapture( 0 ) # Use default camera while True : ret, frame = cap.read() if not ret: break result_frame, status = classificator.frame_processing(frame) # Add status text to frame cv2.putText(result_frame, status, ( 10 , 30 ), cv2. FONT_HERSHEY_SIMPLEX , 1 , ( 0 , 255 , 0 ), 2 ) cv2.imshow( 'Real-time Trash Detection' , result_frame) if cv2.waitKey( 1 ) & 0x FF == ord ( 'q' ): break cap.release() cv2.destroyAllWindows()
Integration The TrashClassificator requires:
numpy for array operationsSegmentationModel from trash_classificator.segmentation.mainDrawing from trash_classificator.drawing.mainOpenCV (cv2) for image processing in user applications
Show Performance Considerations
Image copies are created at each processing step to preserve the original input
The segmentation model uses GPU acceleration when available (managed by DeviceManager)
Tracking persistence is maintained across frames for consistent object IDs
Stream processing mode is enabled for efficient video frame handling