Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{170225,
        author = {Srinivasa Rao Perumalla and Ganesh Racha},
        title = {FPGA Implementation of CNN Binary Kernel for Agricultural Applications},
        journal = {International Journal of Innovative Research in Technology},
        year = {2024},
        volume = {11},
        number = {6},
        pages = {3627-3631},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=170225},
        abstract = {Honeybees play a crucial role in agriculture through their essential contribution to pollination. Examining the pollen resources accessible to honeybees not only aids in comprehending their foraging behavior but also offers valuable insights into the well-being of their habitats. However, conventional methods for monitoring pollen sources can be both time-consuming and costly.
This research project focuses on the development of a Convolutional Neural Network (CNN) binary kernel designed for insect identification using Xilinx Vivado and subsequently deployed on the Xilinx Spartan-3 FPGA development board. The application of this CNN on a Field-Programmable Gate Array (FPGA) hardware platform is explored for the purpose of detecting pollen upon the bees' entrance to the hive.
The proposed CNN architecture comprises four convolutional layers followed by a fully connected layer, optimized for FPGA implementation to achieve optimal throughput and minimal latency. Training the CNN involves a dataset of images capturing pollen grains at the beehive entrance, obtained through a cost-effective, portable imaging system. The dataset is categorized based on the identified types of pollen grains. A comparative evaluation between the FPGA and software-based implementations, the latter executed on a computer using a graphics processing unit (GPU), indicates a significant speedup with the FPGA implementation.
Advantages of the proposed system over traditional pollen detection methods include automation, eliminating the need for manual labor, and non-invasiveness, ensuring that bees remain undisturbed during pollen detection. The system holds potential applications in honeybee research, enabling the monitoring of pollen sources, studying foraging behavior, and assessing habitat health. Additionally, it can contribute to understanding the impact of pesticides on bee populations and optimizing beehive placement for efficient honey production. The system's cost-effectiveness, employing low-cost hardware that can be easily scaled up, further enhances its appeal.
In summary, the utilization of CNNs for pollen detection has the potential to revolutionize honeybee research by providing a faster and more efficient means of monitoring pollen sources. The FPGA implementation enhances speed and efficiency, making real-time monitoring feasible. With its numerous advantages over traditional methods, this proposed system stands as a valuable tool for honeybee research and the enhancement of honey production processes.},
        keywords = {Honeybees, Convolution Neural Network (CNN), Xilinx Vivado, FPGA, Xilinx Spartan 3, Graphic Processing Unit (GPU), Pollen sources.},
        month = {December},
        }