the system consists of a Convolutional neural network model that classify the fruit and its condition eg rotten oranges or fresh apples. the model is then deployed using TensorFlow lite on raspberry pi which controls a servo that removes the rotten fruit
the model is convolutional neural network trained on 10901 image of 3 fruits,Apples, Oranges, and Bananas achieving accuracy of 98% on validation data
model summary
Layer (type) | Output Shape | Param |
---|---|---|
conv2d (Conv2D) | (None, 298, 298, 16) | 448 |
max_pooling2d (MaxPooling2D ) | (None, 149, 149, 16) | 0 |
conv2d_1 (Conv2D) | (None, 147, 147, 32) | 4640 |
max_pooling2d_1 (MaxPooling 2D) | (None, 73, 73, 32) | 0 |
conv2d_2 (Conv2D) | (None, 71, 71, 64) | 18496 |
max_pooling2d_2 (MaxPooling 2D) | (None, 35, 35, 64) | 0 |
conv2d_3 (Conv2D) | (None, 33, 33, 64) | 36928 |
max_pooling2d_3 (MaxPooling 2D) | (None, 16, 16, 64) | 0 |
conv2d_4 (Conv2D) | (None, 14, 14, 64) | 36928 |
max_pooling2d_4 (MaxPooling 2D) | (None, 7, 7, 64) | 0 |
flatten (Flatten) | (None, 3136) | 0 |
dense (Dense) | (None, 512) | 1606144 |
dense_1 (Dense) | (None, 6) | 3078 |
Total params: 1,706,662