FPGA implementation of an image segmentation algorithm using logarithmic arithmetic

Development and Assessment of a Field-Programmable Gate Array (FPGA)-Based Image Processing (FIP) System for Agricultural Field Monitoring Applications

“…The proposed FIP system was able to acquire higher-resolution imagery (800 × 600 px) than the previous literature during the lab and outdoor evaluations of the system [11,13,15,17,18]. In addition, the integration of image acquisition and processing on a lightweight FPGA platform for PA crop monitoring purposes responded adequately to the current needs of real-time farm management decision support systems [19].…”

“…To address these limitations, on-board processing systems based on field programmable gate arrays (FPGAs) are reconfigurable, lightweight, and may be used for real-time object identification and tracking [11][12][13][14][15]. FPGAs are flexible and may be integrated into a wide range of devices, and they reduce the amount of hardware and increase the costeffectiveness of systems [14].…”

“…Furthermore, FPGAs are also described as reconfigurable hardware because the same device may be reused by downloading a new program for different purposes. FPGAs only require hours to carry out the design-implement-test-debug cycle; in comparison, application-specific integrated circuit (ASIC) designs require days to complete all the processing [13,14]. Lastly, their ability to implement parallel processing ensures that multiple local operations may be carried out simultaneously [14] and applied to image processing techniques, as described by Price et al [15].…”

“…FPGA implementations do not have widespread acceptance in the vision community yet [14], although a few studies have included vehicle identification systems [16], real-time license plate localization systems [17], and human presence detection systems [16]. In addition, they have been tested across a wide range of low-resolution dimensions; for example, 180 × 180 px to 600 × 600 px [15], 128 × 128 px [11], 256 × 256 px [13], and 640 × 480 px [17,18]; however, there is the potential for FPGAs to be applicable to higherresolution imagery compared to VGA (640 × 480). Hence, the development of real-time image acquisition and processing systems based on FPGA hardware for crop monitoring purposes is the major focus of this research.…”

Multiprocessing systems development for implementation of applications

“…where the FPGA is only used to speed-up specific mathematic functions. In [124] an image segmentation approach using logarithmic arithmetic is again implemented on an FPGA device, but the performance results are inconclusive because of lack of available FPGA area for the complete system. Yamaoka et al [125] present a pattern matching implementation architecture on an FPGA for input images of 80 x 60 pixels.…”