Mobile Image Processing to Enhance Visual Accessbility for Pineapple Maturity Indices

Abstract

The primary objective is to employ advanced image processing techniques capable of accurately distinguishing between the various stages of pineapple maturation. To achieve this, the study integrates cloud computing into the process, enabling seamless synchronization of annotated data and expediting the training of a cutting-edge object detection model. The methodology introduced in this research involves the incorporation of an updated model of object detection tools, aiming to elevate the precision and efficiency of maturity index differentiation. Leveraging the power of cloud computing infrastructure ensures that the model is continually refined and improved through a cyclical process of annotation, training, and deployment. Furthermore, the study extends its impact beyond the research domain by developing a userfriendly mobile application. This application serves as a practical interface, allowing users, including farmers and agricultural stakeholders, to conveniently assess pineapple maturity on-site. The highest accuracy model is deployed within the mobile, which is the YOLOv8 model with mAP50 value of 0.741 ensuring that end-users benefit from the cutting-edge technology without the need for specialized expertise in image processing or data analytics. Pre -trained model is implemented during app development process by recalling the model parameter. In conclusion, this interdisciplinary approach, combining mobile image processing, cloud computing, and user-centric application development, holds substantial promise for revolutionizing pineapple production and, by extension, contributing to the advancement of precision agriculture.