Building a Self-Controlled Car through AI inferences & IoT

Install Python on Windows, validate the setup in an IDE with simple code, and plan for IoT integration with Arduino, Jupyter Notebook, and Anaconda.

Create and activate a Conda virtual environment to isolate project dependencies, install libraries such as numpy, OpenCV, and TensorFlow, and launch a Jupyter Notebook within that environment.

Learn how to install OpenCV in a Python environment for image and video processing, compare open source libraries, and verify the setup for image datasets, preprocessing, and CNN-based ML workflows.

Install TensorFlow in a virtual environment, and understand tensors, forward and backward propagation with gradient descent for AI model training.

Install PyTorch using the official PyTorch site, choose Windows and CPU options, verify Python version, and run pip install torch torchvision to prepare for model inferences.

Learn the basics of digital images and computer vision, including color and grayscale images, dimensions, and essential image processing, then extract features for classification, detection, segmentation, and tracking with AI.

Explore image basics, from digitized pictures to pixel data, and how rgb color channels with 0–255 intensities and width and height are represented in NumPy arrays.

Read an image with cv2.imread, choose grayscale or color, and visualize it with cv2.imshow in a Jupyter notebook session, wait for a key, and destroy all windows.

Learn video basics: frame by frame analysis, frames per second, and reading videos with OpenCV for later object detection and tracking in computer vision.

Learn to read a video with OpenCV, capture frames, and extract frames per second, width, height, and frame count while running code in a virtual environment and Jupyter notebook.

Read a video with cv2 and display frame by frame, resizing frames for clear visualization. Handle read errors, optionally use a webcam, and release resources when finished.

Learn to annotate video frames by drawing circles, rectangles, and lines using OpenCV, with practical guidance on coordinates, color, thickness, and frame-by-frame processing.

Learn to overlay text on video frames using cv2.putText, including text, location, font, size, color, and thickness for real-time computer vision display, applicable to ai inferences and iot projects.

Explore ai, deep learning, and convolutional neural networks to process image and video data, extract patterns, and apply yolo-based object detection and image classification on Arduino and IoT cars.

Compare deep learning with classical machine learning, through neural networks, forward and backward propagation, training epochs, and bidirectional learning that updates weights for better predictions.

Explore the core deep learning components, including neural networks, activation and loss functions, forward and backward propagation, optimization, and architectures like input, hidden, and output layers.

Explore how convolutional neural networks extract image features using filters, strides, and padding, building a feature pyramid for image classification, detection, and OCR.

Develop an object detection model for autonomous cars using YOLO, exploring image classification, segmentation, and detection, and preparing data sets and pre-trained models.

learn to develop the yolo algorithm with a custom dataset through transfer learning from pre-trained weights, download the models and weights, and add custom classes using ultralytics.

Annotate images with bounding boxes and labels using tools like Label Studio or Roboflow, export annotations, and create train/val/test splits plus a data.yaml for YOLO with 80 classes.

Prepare and annotate datasets by uploading images, creating bounding-box annotations, configuring preprocessing and augmentation, and exporting a train/val/test-split YOLOv7 dataset for training.

Train the model with a prepared dataset and pre-trained weights via YOLO and Ultralytics, tuning epochs, batch size, and image size in a Jupyter workflow.

Build an Arduino-driven car by integrating IoT devices, Python AI, and computer vision, assembling hardware from a four-wheel driver kit with RoboScare parts.

Assemble a complete autonomous car using a wifi-enabled Arduino, servo motors, ultrasonic and UV sensors, and a camera, controlled by Python and AI inference with YOLO.

Use a gpu-enabled laptop with an anaconda environment to run trained yolo live inferences on an esp32 camera stream, guiding an arduino car via uv sensors.

Test ai algorithms and iot devices by running a self-controlled car equipped with real-time object detection, detecting people, water bottles, and devices, and adjusting motion with ai-driven sensing.

Conclude with software setup, real-time testing, and hardware integration for a self-driven car, covering computer vision, YOLO detection, and deep learning basics with IoT devices like Arduino and ESP32 camera.