FPGA-Based Embedded Motion Estimation Sensor

Abstract

Accurate real-time motion estimation is very critical to many computer vision tasks. However, because of its computational power and processing speed requirements, it is rarely used for real-time applications, especially for micro unmanned vehicles. In our previous work, a FPGA system was built to process optical flow vectors of 64 frames of640×480image per second. Compared to software-based algorithms, this system achieved much higher frame rate but marginal accuracy. In this paper, a more accurate optical flow algorithm is proposed. Temporal smoothing is incorporated in the hardware structure which significantly improves the algorithm accuracy. To accommodate temporal smoothing, the hardware structure is composed of two parts: the derivative (DER) module produces intermediate results and the optical flow computation (OFC) module calculates the final optical flow vectors. Software running on a built-in processor on the FPGA chip is used in the design to direct the data flow and manage hardware components. This new design has been implemented on a compact, low power, high performance hardware platform for micro UV applications. It is able to process 15 frames of640×480image per second and with much improved accuracy. Higher frame rate can be achieved with further optimization and additional memory space.

Bio-inspired robust optical flow processor system for VLSI implementation

“…A complete accuracy analysis has been accomplished, it being possible to examine the quality of the improved results with respect to previous works [2,3] under the common Barron [4] error metric. Regarding accuracy, this design delivers better results (Table 1) than the other systems [5][6][7], even with 100% density. Regarding performance, it is possible to have throughput higher than 5 Mpixel/s keeping the limitation stage from the final datapath in the modulus and phase stage.…”

“…It deals efficiently with many challenges, such as illumination, static patterns, contrast invariance, different kinds of noise, robustness against fails, is very useful in camouflage tasks, justification of some optical illusions, detection of second-order motion, while avoiding operations such as matrix inversion or iterative methods that are not biologically justified. This platform has the advantage to give information where other standard gradient models [5][6][7] fail.…”

Hardware-Friendly Vision Algorithms for Embedded Obstacle Detection Applications

“…The optical flow algorithm used in the proposed obstacle detection sensor is based on our previous development [16], [25], [26]. It is a tensor-based algorithm that is suitable for hardware implementation because of its simple and repetitive computations.…”

FPGA-Based Implementation of Real Time Optical Flow Algorithm and Its Applications for Digital Image Stabilization

“…Hardware components: (a) CMOS imager, (b) Virtex-II-1000-Board. Table 5 shows a comparison between our algorithm and algorithms presented in [3], [18] and [19]. The comparison was performed with respect to the amount of resources necessary to implement the algorithm, power consumption as well as the accuracy using error measure (12).…”