Identify and Classify Objects with Edge Impulse and ESP32

The system operates by capturing visual data through the ESP32 cam module, which serves as the core processing and vision unit. Power is regulated efficiently across the circuit using the LM117 voltage regulator and stabilized by the 100uF capacitors (C1), ensuring a steady DC supply. Once an image frame is captured, it is evaluated by a machine learning model trained via Edge Impulse, which classifies the object based on its visual features. To display the classification results locally, the system communicates with an SSD1306 OLED display using standard I2C protocols. When a specific classification condition is met, the ESP32 activates a digital output pin connected to a BC547 transistor (T1). This transistor acts as a solid-state electronic switch to drive the 5V relay module safely. A flyback diode (D1) is placed across the relay coil to protect the transistor from inductive voltage spikes when the relay switches off. The FTDI USB programmer is used externally during the initial development phase to flash code compiled in the Arduino IDE onto the ESP32 development board. Together, this combination of embedded machine learning, peripheral display interfacing, and transistor-switched electromechanical control allows the hardware to autonomously make decisions and trigger external mechanical actions based on visual input.