Fruit Quality Detection Using Convolutional Neural Networks

Fruit ripening stage is a major determinant of fruit quality. Factors such as size, shape, and color are typically used to assess quality during the selection process. This work addresses the detection of fruit quality by comparing images from the dataset to real-time images. In the context of detecting fruit quality, real-time images refer to images that are captured and processed immediately as they are acquired, without any significant delay. These images are taken in the present moment, reflecting the current state of the fruits being examined. The term real-time implies that there is little to no time lag between capturing the image and its analysis. When comparing images from a dataset to real-time images, the idea is to assess the quality of the fruits based on certain criteria, such as size, shape, and color. The lightweight DenseNet model is proposed for the detection of fruit quality. The process of training and testing the proposed lightweight model involves a combination of sophisticated algorithms, datasets, and validation techniques. We first apply data preprocessing and then split the data into its training and testing datasets. The training of the CNN model will take place in its three layers, which are the convolutional layer, the maxpool layer, and the dropout layer. We then apply the ReLU activation function to start the testing of the model. The results show that the efficacy of the proposed lightweight DenseNet model has achieved an accuracy of 99.48% over an 80:20 ratio on an open, freely available dataset (Fruit Dataset. 2023. Last Accessed: 10 May 2023. Available at: https://storage.googleapis.com/cvstock-932a9-h58gl/kc81r5iu8qsmomlfs5ug6e.zip ). In addition, the proposed model is also a promising approach for fruit quality detection and has the potential to be used in a variety of applications.