Technology

Built to see, think, and sort.

Every component in Recycla was chosen for a reason. From the neural network architecture to the servo motors, here is the full technical breakdown of a system designed to classify waste in under 200 milliseconds on a $35 computer.

Classification Engine

MobileNetV2 with Transfer Learning

Pre-trained on 1.2 million images by Google, then fine-tuned on our custom waste dataset. The model classifies six material categories: cardboard, glass, metal, paper, plastic, and general trash.

MobileNetV2 was chosen for its depthwise separable convolutions, which reduce computation by 8-9x compared to standard convolutions while maintaining classification accuracy. This makes it ideal for inference on resource-constrained hardware.

Architecture MobileNetV2
Input Size 224 x 224px
Categories 6 classes
Format TFLite INT8
Inference <200ms
Training Transfer learning
Hardware

Purpose-built from the ground up.

Compute

Raspberry Pi 4

4GB RAM, quad-core ARM Cortex-A72. Runs TFLite inference, controls GPIO for sensors and actuators, and manages the full classification pipeline. Chosen for its balance of processing power, GPIO accessibility, and cost-effectiveness.

Vision

Arducam 8MP V2.3

Sony IMX219 sensor. Captures sharp images across varying lighting conditions for reliable classification of glossy, matte, and transparent surfaces. The CSI interface provides low-latency image transfer directly to the Pi's GPU.

Sensing

HC-SR04 Ultrasonic

Detects object presence within 30cm to trigger the classification pipeline. Low power consumption allows continuous standby monitoring. Dual transducer design provides reliable distance measurement with 3mm precision.

Actuation

Dual SG90 Servos

Independent servo control for recycling and garbage compartment lids. 180-degree rotation, 1.8kg/cm torque. Responsive enough for real-time sorting, reliable enough for thousands of cycles.

Training Pipeline

Built for real-world conditions.

Our training pipeline uses aggressive data augmentation to ensure the model performs outside of clean lab conditions.

Brightness & Contrast

Random shifts simulate fluorescent, natural, and low-light conditions found across campus buildings.

Geometric Transforms

Rotation, flips, and zoom account for varied item orientations as waste is dropped into the bin.

Color Temperature

Channel shifts handle the difference between warm indoor lighting and cool outdoor daylight.

Surface Handling

Specular highlight training addresses glossy packaging surfaces that can confuse standard classifiers.

The dataset is designed to grow. Adding new training data is as simple as dropping photos into the dataset folder and retraining.