本研究是建立一個基於FPGA平台的系統，以用於移動取像裝置之實時標記移動目標物。從取樣頻率為26Hz的相機獲取影像後，使用脈動陣列架構對連續兩幀影像進行區塊比對，以完成影像配準。接著利用隨機取樣一致算法找出最適合的背景位移仿射變換模型。最後，我們利用已經完成背景位移補償的兩張影像進行幀間差法，以確定移動的物體並完成對移動物體的標記。整個系統的流程包括影像的獲取、區塊比對、仿射變換模型推導以及最終之移動物體標記。研究結果顯示，相較於傳統之序列式運算，以脈動陣列架構進行之區塊比對有利於電路對平行運算之實現，同時達到節省運算元件的效果。

This research is to build a system based on an FPGA platform for real-time marking of moving objects in portable imaging devices. After capturing images from a camera with a sampling frequency of 26Hz, a systolic array architecture is used to perform block matching on two consecutive frames of images to access image registration. And using RANdom SAmple Consensus(RANSAC) to find the most suitable background displacement affine transformation model. Finally, using the frame differencing method on the two images with background displacement compensation to determine the moving object and complete the marking of the moving object. The entire research includes image acquisition, block matching, derivation of affine transformation models, and marking of moving objects. The results show that compared to traditional sequential computations and the systolic array architecture for block matching. The systolic array is more suitable for parallel computation in the circuit design and reduces computational elements.

[1] Dairath et al. Computer vision-based prototype robotic picking cum grading system for fruits. Smart Agricultural Technology, 4:100210, 2023.
[2] D J Kang. Dynamic programming-based method for extraction of license plate numbers of speeding vehicles on the highway. International Journal of Automotive Technology, 10:205–210, 2009.
[3] Kulkarni et al. Traffic light detection and recognition for self driving cars using deep learning. In 2018 Fourth International Conference on Computing Communication Control and Automation (ICCUBEA), pages 1–4, 2018.
[4] Stjepanovic et al. Automated vehicle and pedestrian detection. 2021.
[5] Chowdhury et al. Drones for disaster response and relief operations: A continuous approximation model. International Journal of Production Economics, 188:167–184, 2017.
[6] 方睿ex al. 卷神网的fpga 并行加速方案. 算机工程与用, (8):32–36, 2015.
[7] Wyawahare et al. Image registration techniques: an overview. International Journal of Signal Processing, Image Processing and Pattern Recognition, 2(3):11–28, 2009.
[8] Barbara Zitova and Jan Flusser. Image registration methods a survey. Image and vision computing, 21(11):977–1000, 2003.
[9] Alba et al. Phase correlation based image alignment with subpixel accuracy. In Advances in Artificial Intelligence: 11th Mexican International Conference on Artificial Intelligence, MICAI 2012, San Luis Potos´ı, Mexico, October 27–November 4, 2012. Revised Selected Papers, Part I 11, pages 171–182. Springer, 2013.
[10] Russakoff et al. Image similarity using mutual information of regions. In Computer Vision-ECCV 2004: 8th European Conference on Computer Vision, Prague, Czech Republic, May 11-14, 2004. Proceedings, Part III 8, pages 596–607. Springer, 2004.
[11] Aroh Barjatya. Block matching algorithms for motion estimation. IEEE Transactions Evolution Computation, 8(3):225–239, 2004.
[12] Yang et al. A simple camera calibration method based on sub-pixel corner extraction of the chessboard image. volume 3, pages 688 – 692, 2010.
[13] Tang et al. Fpga implementation of ransac algorithm for real-time image geometry estimation. pages 290 – 294, Putrajaya, Malaysia, 2013.
[14] Wong et al. A sum of absolute differences implementation in fpga hardware. In Proceedings. 28th Euromicro Conference, pages 183–188. IEEE, 2002.
[15] Hisham et al. Template matching using sum of squared difference and normalized cross correlation. In 2015 IEEE student conference on research and development (SCOReD), pages 100–104. IEEE, 2015.
[16] Hwang-Seok Oh et al. Adaptive bit-reduced mean absolute difference criterion for block-matching algorithm and its vlsi design. 37(12):3272–3281, 1998.
[17] Kung et al. Systolic arrays (for vlsi). In Sparse Matrix Proceedings 1978, volume 1, pages 256–282. Society for industrial and applied mathematics Philadelphia, PA, USA, 1979.
[18] Paolo Gorlani and Christian Plessl. High level synthesis implementation of a three-dimensional systolic array architecture for matrix multiplications on intel stratix 10 fpgas. 2021.
[19] Nishu Singla. Motion detection based on frame difference method. International Journal of Information & Computation Technology, 4(15):1559–1565, 2014.
[20] R.C. Gonzalez and R.E. Woods. Digital Image Processing. Upper Saddle River, 2008.
[21] LUC DE VOS and MICHAEL STEGHERR. Parameterizable vlsi architectures for the full- search block- matching algorithm. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS, 36(10), 1989.
[22] Sheng-Hsiang Hung. Fpga-based real-time moving target detection with a moving camera. Master’s thesis, National Cheng Kung University, 2021.
[23] Komarek et al. Array architectures for block matching algorithms. IEEE Transactions on Circuits and Systems, 36(10):1301–1308, 1989.
[24] TRDB-D5M Hardware specification V0.2, 2009.
[25] PMI KM33-05A-400N0-0331. http://www.nsk-cs.com.tw/?f=km-monostage.
[26] DE2-115 User Manual, 2017.