FRUIT QUALITY CLASSIFICATION USING IMAGE PROCESSING TECHNIQUES.

ELECTRONICS AND COMMUNICATION ENGINEERING

Digital image processing is widely used in the classification and evaluation of fruit quality. This essay explores the process of determining whether a bunch of fruits is good or flawed. This helps the customer avoid purchasing faulty fruits and also helps the consumer prevent any unintended health problems that can result from consuming faulty fruits. Nowadays, it is imperative to check fruit products for flaws before releasing them onto the market. This addendum helps the business understand the client's contribution to determining the fruits' quality. Given the most recent developments in digital image processing technology, using these techniques to evaluate the quality of food products, especially fruits have become very significant. These techniques allow one to categorize fruits and identify which ones are rotten and which are good. Regarding the current disposition, it is critical to accurately identify food items and evaluate their quality. These automated technologies can reduce a great deal of manual labor and accelerate the food processing industry. Regarding this, the latest developments in Deep Learning-based architectures have brought forth a multitude of options that provide exceptional results in various categorization problems. Accurate recognition of food items along with quality assessment is of paramount importance in the agricultural industry. Such automated systems can speed up the wheel of the food processing sector and save tons of manual labor. In this connection, the recent advancement of Deep learning-based architectures has introduced a wide variety of solutions offering remarkable performance in several classification tasks. In this work, we have exploited the concept of CNN, DenseNet121 and NasNetLarge for fruit quality assessment. These have been applied on the images chosen from Fruitnet Database. The feature propagation towards the deeper layers has enabled the network to tackle the vanishing gradient problems and ensured the reuse of features to learn meaningful insights. Evaluating on a dataset of 19,526 images containing six fruits having three quality grades for each, the proposed pipeline achieved an accuracy of 86%. The robustness of the model was further tested for fruit classification and quality assessment tasks where the model produced a similar performance, which makes it suitable for real-life applications. The DenseNet121 outperforms NasNetLarge in terms of the performance metrics used such as F1 score, accuracy, etc. Hence, DenseNet121 is our preferred architecture for classification of fruits into good and bad ones.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.