Project Files committed

Project README.md

@@ -0,0 +1,17 @@
+# Vehicle LICENSE Plate Detection
+
+Detects the LICENSE Plates for variety of different Vehicles.
+
+It has been trained using the YOLOv8 model from **ultraytics** for 15 epochs on NVIDIA T4 GPU
+
+**Roboflow Dataset used: https://universe.roboflow.com/roboflow-universe-projects/license-plate-recognition-rxg4e**
+
+**App: https://huggingface.co/spaces/Harsh72AI/Vehicle-License-Plate-Detection**
+
+YOLOv8 model has achieved:
+* Box Loss - 0.432
+* mAP50 - 0.977
+* mAP50-95 - 0.682
+
+
+

YOLOv8 Training Results.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29bece8ef0a7aedd08a4d699c45f8c5d5e8179c636c7ed1d565407e13a152f58
+size 103926861

YOLOv8_best_weights.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:709142c277aaaffa348f907994ee42f2cd40b85c79701572cad8197762ee593f
+size 52028609

app.py

@@ -0,0 +1,34 @@
+from predict_pipeline import DetectionPipeline
+import streamlit as st
+
+st.title('Automatic Vechile LICENSE Plate detection')
+st.write('Detects the License plate of a car and predicts the digits present in it! \nPowered by YOLOv8 Medium model')
+
+st.write('')
+
+detect_pipeline = DetectionPipeline()
+
+st.info('License Plate Detector MODEL loaded successfully!')
+
+uploaded_file = st.file_uploader("Upload an image", type=["jpg", "png", "jpeg"])
+
+if uploaded_file is not None:
+    
+    with st.container():
+        col1, col2 = st.columns([3, 3])
+        col1.header('Input Image')
+        col1.image(uploaded_file, caption='Uploaded Image', use_column_width=True)
+
+        col1.text('')
+        col1.text('')
+
+        if st.button('Detect!'):
+            preprocessed_img_array = detect_pipeline.preprocess_image(uploaded_file=uploaded_file)
+            detections = detect_pipeline.detectLicensePlates(input_array=preprocessed_img_array)
+            detections_img = detect_pipeline.detections2Image(preprocess_image=preprocessed_img_array, detections=detections)
+
+            col2.header('Detections')
+            col2.image(detections_img, caption='Predictions by model', use_column_width=True)
+
+
+      

predict_pipeline.py

@@ -0,0 +1,46 @@
+from ultralytics import YOLO
+import numpy as np
+from PIL import Image
+import cv2 as cv
+
+
+class DetectionPipeline():
+    def __init__(self) -> None:
+        # Initialize and Load in Custom YOLOv8 model training weights
+        self.license_plate_detector = YOLO('YOLOv8_best_weights.pt')  
+
+
+    def preprocess_image(self, uploaded_file):
+        """
+        Takes a File in the format - '.jpg' / '.png' / '.jpeg' then converts it into a Numpy array and returns it
+
+        Args:
+            uploaded_file: File that needs to be converted to Numpy array
+
+        Returns:
+            img_array: Image in the `np.array` format
+        """
+
+        img = Image.open(uploaded_file).convert('RGB')
+        img_array = np.array(img)
+
+        return img_array
+    
+    def detectLicensePlates(self, input_array):
+        detections = self.license_plate_detector(input_array)[0]
+        license_plate_detections = []
+        for license_plate in detections.boxes.data.tolist():
+            x1, y1, x2, y2, score, class_id = license_plate
+            license_plate_detections.append([int(x1), int(y1), int(x2), int(y2), score])
+
+        return license_plate_detections
+    
+    def detections2Image(self, preprocess_image: np.array, detections:list):
+        img = np.array(preprocess_image, dtype='uint8')
+        for license_plate_info in detections:
+            x1, y1, x2, y2, score = license_plate_info
+            cv.rectangle(img, pt1=(x1, y1), pt2=(x2, y2), color=(0, 255, 0), thickness=2)
+        img_detections = np.array(img)
+        return img_detections
+
+

requirements.txt

@@ -0,0 +1,3 @@
+ultralytics
+torch
+numpy