Deep learning classification system for coconut maturity levels based on acoustic signals

June Anne Caladcad,Eduardo Jr Piedad

2024-08-27

Abstract:The advancement of computer image processing, pattern recognition, signal processing, and other technologies has gradually replaced the manual methods of classifying fruit with computer and mechanical methods. In the field of agriculture, the intelligent classification of post-harvested fruit has enabled the use of smart devices that creates a direct impact on farmers, especially on export products. For coconut classification, it remains to be traditional in process. This study presents a classification of the coconut dataset based on acoustic signals. To address the imbalanced dataset, a data augmentation technique was conducted through audiomentation and procedural audio generation methods. Audio signals under premature, mature, and overmature now have 4,050, 4,050, and 5,850 audio signals, respectively. To address the updation of the classification system and the classification accuracy performance, deep learning models were utilized for classifying the generated audio signals from data generation. Specifically, RNN and LSTM models were trained and tested, and their performances were compared with each other and the machine learning methods used by Caladcad et al. (2020). The two DL models showed impressive performance with both having an accuracy of 97.42% and neither of them outperformed the other since there are no significant differences in their classification performance.

Sound,Audio and Speech Processing,Signal Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is some deficiencies in the traditional methods of coconut maturity classification, especially the inconsistency and inefficiency of the manual classification method. Specifically:

1. **Limitations of Traditional Classification Methods**: Traditional coconut maturity classification mainly relies on manually knocking on coconuts and making judgments based on the sound. This method is not only time - consuming and highly subjective but also error - prone. Especially when dealing with a large number of coconuts, it is difficult to ensure efficiency and accuracy.

2. **Dataset Imbalance Problem**: In previous studies, the datasets used to train models had the problem of unbalanced sample distribution, resulting in poor classification performance of the model on certain categories. For example, in Caladcad et al.'s study, the number of coconut samples at different maturity levels varies greatly, which affects the overall performance of the model.

3. **Improving Classification Accuracy and Performance**: By introducing deep - learning models (such as RNN and LSTM), the aim is to improve the accuracy and performance of coconut maturity classification. Especially when dealing with large - scale coconut exports and industrial processing, it can significantly reduce the waste caused by classification errors.

To solve these problems, this study proposes a coconut maturity classification system based on acoustic signals and makes improvements through the following methods:

- **Data Augmentation**: Use audio augmentation and procedural audio generation methods to increase the sample size of the dataset, solving the problem of sample imbalance in the original dataset.

- **Application of Deep - Learning Models**: Utilize RNN and LSTM, two time - series deep - learning algorithms, to classify the generated audio signals and compare them with previous machine - learning methods, verifying the superiority of deep - learning models in classification tasks.

Finally, this study demonstrates the high efficiency and accuracy of deep - learning models in coconut maturity classification, providing new ideas and technical support for intelligent classification in the agricultural field.