Scholarlink

This study presents an implementation of rice crop yield classification for Telangana State, employing an Artificial Neural Network (ANN) algorithm executed on both a Central Processing Unit (CPU) and the PYNQ-Z2 processor. The objective was to evaluate the performance improvement achieved by executing the same model on the PYNQ-Z2 processor. A comprehensive dataset specific to Telangana State, comprising diverse rice crop features and corresponding yield data, was utilized for training the ANN algorithm. The trained ANN algorithm was subsequently deployed on the PYNQ-Z2 processor, capitalizing on its FPGA-based hardware acceleration capabilities. Experimental evaluations showed a significant 40% speed improvement when running the code on the PYNQ-Z2 processor. This highlights the inherent performance benefits of FPGA-based acceleration. By leveraging the improved execution speed, timely insights into rice crop yield can be obtained, facilitating informed decision-making regarding crop management, resource allocation, and yield optimization. The utilization of the PYNQ-Z2 processor offers a cost-effective solution for real-time rice crop yield classification, enhancing agricultural productivity.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Efficient Rice Yield Classification: Accelerating ANN Processing on PYNQ-Z2 Processor

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • S. Pavani,
  • P. Augusta Sophy Beulet

摘要

This study presents an implementation of rice crop yield classification for Telangana State, employing an Artificial Neural Network (ANN) algorithm executed on both a Central Processing Unit (CPU) and the PYNQ-Z2 processor. The objective was to evaluate the performance improvement achieved by executing the same model on the PYNQ-Z2 processor. A comprehensive dataset specific to Telangana State, comprising diverse rice crop features and corresponding yield data, was utilized for training the ANN algorithm. The trained ANN algorithm was subsequently deployed on the PYNQ-Z2 processor, capitalizing on its FPGA-based hardware acceleration capabilities. Experimental evaluations showed a significant 40% speed improvement when running the code on the PYNQ-Z2 processor. This highlights the inherent performance benefits of FPGA-based acceleration. By leveraging the improved execution speed, timely insights into rice crop yield can be obtained, facilitating informed decision-making regarding crop management, resource allocation, and yield optimization. The utilization of the PYNQ-Z2 processor offers a cost-effective solution for real-time rice crop yield classification, enhancing agricultural productivity.