Real-Time Deployment of MobileNetV3 Model in Edge Computing Devices Using RGB Color Images for Varietal Classification of Chickpea

Chickpeas are one of the most widely consumed pulses globally because their high protein content. The morphological features chickpea seeds, such as colour and texture, observable play a major role in classifying different varieties. This process is often carried out by human experts, time-consuming, inaccurate, expensive. objective study was to design an automated classifier using RGB-colour-image-based model for considering seed. As part data acquisition process, five hundred fifty images were collected per variety four varieties (CDC-Alma, CDC-Consul, CDC-Cory, CDC-Orion) industrial RGB camera mobile phone camera. Three CNN-based models NasNet-A (mobile), MobileNetV3 (small), EfficientNetB0 evaluated transfer-learning-based approach. classification accuracy 97%, 99%, 98% models, respectively. used further deployment on Android Raspberry Pi 4 devices based its higher light-weight architecture. 100% while deployed both platforms.