updated model and added logs

conf/application.conf

@@ -11,4 +11,4 @@ host=localhost
 port=8200
 
 [PATH]
-model_path = scripts/utility/model/yolov8l_r1_r2_lp_best.pt
+model_path = scripts/utility/model/r1_r2_lpr_yolov8s.pt

requirements.txt

@@ -55,4 +55,5 @@ uvicorn==0.19.0
 ultralytics
 paddleocr==2.7.0.3
 paddlepaddle==2.5.2
+#loguru==0.7.2
 

scripts/core/handler/Text_Extraction_Handler.py

@@ -3,28 +3,44 @@ from ultralytics import YOLO
 import torch
 import cv2
 from scripts.config.app_configurations import *
+from scripts.logging.application_logging import logger
 
 device = 'cuda' if torch.cuda.is_available() else 'cpu'
-# print(f'Using device: {device}')
+
 def load_ocr_model():
+
+    """
+    Load paddle OCR model
+
+    """
     ocr = PaddleOCR(use_angle_cls=True, lang='en')
+    logger.info("Loaded OCR model")
     return ocr
 
 def load_yolo_model():
+    """
+    Load license plate detection model
+
+    """
     model = YOLO(MODEL_PATH).to(device)
+    logger.info("Loaded Yolo model")
     return model
 
 
 def predict(img_path, ocr, yolo_model):
+    try:
+
         img = cv2.imread(img_path)
-    results = yolo_model.predict(img, iou=0.7)
+
+        results = yolo_model.predict(img, iou=0.7) # predict r1, r2 and license plate using yolo model r1 - single letter region in license plate, r2 - four or five letter region in license plate
+
         if results:
             r1_txt = None
             r2_txt = None
 
             for r in results:
 
-            # annotator = Annotator(img)
+
 
                 boxes = r.boxes
                 for box in boxes:
@@ -37,53 +53,45 @@ def predict(img_path, ocr, yolo_model):
                     w = abs(x2 - x1)
                     h = abs(y2 - y1)
 
-                cv2.rectangle(img, (x1, y1), (x2, y2), (255, 0, 0), 1)
 
-                # lp = img[max(y1 - 30, 0):min(y1+h + 30, img.shape[0]), max(x1 - 30, 0):min(x1+w + 30, img.shape[1])]
 
                     lp  = img[y1:y1 + h, x1:x1 + w]
 
                     c = box.cls
 
-                # print("class:   ", c)
-                # annotator.box_label(b, model.names[int(c)])
 
-                # img = annotator.result()
-                # cv2.imshow('YOLO V8 Detection', cv2.resize(img, (900, 600)))
                     if int(c) == 0:
+                        #extracting text from r1
+                        # cv2.imwrite("LP.png", lp)
                         result = ocr.ocr(lp, cls=True)
                         for idx in range(len(result)):
                             res = result[idx]
-                        text_x = 10
-                        text_y = 30
+
                             if res:
 
                                 for ind, line in enumerate(res):
-                                # print("line")
-                                # print(line)
+
 
                                     text_tuple = line[1]
                                     if r1_txt is None:
                                         r1_txt = text_tuple[0]
-                                    print("R1  TEXT:   ", r1_txt)
+                                        logger.info(f"r1 text : {r1_txt}")
                                     text_conf = text_tuple[1]
                     elif int(c) == 1:
-
+                        #extracting text from r2
                         result = ocr.ocr(lp, cls=True)
                         for idx in range(len(result)):
                             res = result[idx]
-                        text_x = 10
-                        text_y = 30
+
                             if res:
 
                                 for ind, line in enumerate(res):
-                                # print("line")
-                                # print(line)
+
 
                                     text_tuple = line[1]
                                     if r2_txt is None:
                                         r2_txt = text_tuple[0]
-                                    print("R2   TEXT:   ", r2_txt)
+                                        logger.info(f"r2 text : {r2_txt}")
                                     text_conf = text_tuple[1]
                                 # cv2.putText(lp, txt, (text_x , text_y + ind*50), cv2.FONT_HERSHEY_SIMPLEX , 1, (255, 0, 0), 2, cv2.LINE_AA)
 
@@ -94,12 +102,14 @@ def predict(img_path, ocr, yolo_model):
             if r1_txt is not None and r2_txt is not None:
                 LP = f"{r1_txt}  {r2_txt}"
 
-            # print("LP:  ", LP)
+                logger.info(f"License plate reading : {LP}")
                 return LP
 
         return "Image Not Clear"
-    # result = ocr.ocr(img_path, cls=False,det=True,rec=True)[0]
-    # txts = [line[1][0] for line in result]
+    except Exception as e:
+        logger.info(f"Error while reading license plate : {e}")
+        return f"Error while reading license plate : {e}"
+
 
 
 

scripts/core/services/Text_Extraction_Service.py

@@ -8,6 +8,7 @@ from scripts.core.handler.Text_Extraction_Handler import load_ocr_model,load_yol
 from scripts.logging.application_logging import logger
 from scripts.config.app_constants import TEXT_EXTRACT
 
+
 ocr = load_ocr_model()
 yolo_model = load_yolo_model()
 router = APIRouter()