Corner Tracker

Corner Tracker example resides in L3/examples/cornertracker directory.

This benchmark tests the performance of cornertracker function with a sequence of 2 images. This example illustrates how to detect and track the characteristic feature points in a set of successive frames of video. A Harris corner detector is used as the feature detector, and a modified version of Lucas Kanade optical flow is used for tracking. The core part of the algorithm takes in current and next frame as the inputs and outputs the list of tracked corners. The current image is the first frame in the set, then corner detection is performed to detect the features to track.

The tutorial provides a step-by-step guide that covers commands for building and running kernel.

Executable Usage

• Work Directory(Step 1)

The steps for library download and environment setup can be found in README of L3 folder. For getting the design,

cd L3/examples/cornertracker

• Build kernel(Step 2)

Run the following make command to build your XCLBIN and host binary targeting a specific device. Please be noticed that this process will take a long time, maybe couple of hours.

export OPENCV_INCLUDE=< path-to-opencv-include-folder >
export OPENCV_LIB=< path-to-opencv-lib-folder >
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:< path-to-opencv-lib-folder >
export DEVICE=< path-to-platform-directory >/< platform >.xpfm
make host xclbin TARGET=hw

• Run kernel(Step 3)

To get the benchmark results, please run the following command.

make run TARGET=hw

• Example output(Step 4)

-----------Corner Tracker Design---------------
Found Platform
Platform Name: Xilinx
XCLBIN File Name: krnl_cornertracker
INFO: Importing vision/L3/tests/cornertracker/cornertrack/Xilinx_Cornertrack_L3_Test_vitis_hw_u200/build_dir.hw.xilinx_u200_xdma_201830_2/krnl_cornertracker.xclbin
Loading: 'vision/L3/tests/cornertracker/cornertrack/Xilinx_Cornertrack_L3_Test_vitis_hw_u200/build_dir.hw.xilinx_u200_xdma_201830_2/krnl_cornertracker.xclbin'
***************************************************
Test Case no: 1

 Harris Execution

 Harris: Buffers created

 Harris: data copied to device

 Harris: args set

 Harris kernel called


 Harris Done

 Pyrdown Execution

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 144
 pyr in wd = 256
 pyr out ht = 72
 pyr out wd = 128

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 72
 pyr in wd = 128
 pyr out ht = 36
 pyr out wd = 64

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 36
 pyr in wd = 64
 pyr out ht = 18
 pyr out wd = 32

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 18
 pyr in wd = 32
 pyr out ht = 9
 pyr out wd = 16

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 **********Optical Flow Computation*******************

 Buffers created

  *********OF Computation Level =4*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done4

 Buffers created

  *********OF Computation Level =3*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =9flow_in_cols =16

 level3calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done4

 Buffers created

  *********OF Computation Level =2*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =18flow_in_cols =32

 level2calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done4

 Buffers created

  *********OF Computation Level =1*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =36flow_in_cols =64

 level1calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done4

 Buffers created

  *********OF Computation Level =0*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =72flow_in_cols =128

 level0calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done4

 OF done

 **********Corner Update Computation*******************

 kernel args set

 flow_rows =144flow_cols=256num of corners=47harris_flag=1

 Corner Update called

 Corner Update done

 OF done

 Num of corners = 47
***************************************************
***************************************************
Test Case no: 2

 Harris Execution

 Harris: Buffers created

 Harris: data copied to device

 Harris: args set

 Harris kernel called


 Harris Done

 Pyrdown Execution

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 144
 pyr in wd = 256
 pyr out ht = 72
 pyr out wd = 128

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 72
 pyr in wd = 128
 pyr out ht = 36
 pyr out wd = 64

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 36
 pyr in wd = 64
 pyr out ht = 18
 pyr out wd = 32

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 Pyrdown: Buffers created

 Pyrdown: data copied to device

 Pyrdown Args set

 pyr in ht = 18
 pyr in wd = 32
 pyr out ht = 9
 pyr out wd = 16

 Pyrdown kernel called


 Pyrdown data copied to host

 Pyrdown Execution done

 **********Optical Flow Computation*******************

 Buffers created

  *********OF Computation Level =4*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =9flow_cols=16flow_in_rows =9flow_in_cols =16

 level4calls done4

 Buffers created

  *********OF Computation Level =3*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =9flow_in_cols =16

 level3calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =18flow_cols=32flow_in_rows =18flow_in_cols =32

 level3calls done4

 Buffers created

  *********OF Computation Level =2*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =18flow_in_cols =32

 level2calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =36flow_cols=64flow_in_rows =36flow_in_cols =64

 level2calls done4

 Buffers created

  *********OF Computation Level =1*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =36flow_in_cols =64

 level1calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =72flow_cols=128flow_in_rows =72flow_in_cols =128

 level1calls done4

 Buffers created

  *********OF Computation Level =0*********

  *********OF Computation iteration =0*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =72flow_in_cols =128

 level0calls done0

  *********OF Computation iteration =1*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done1

  *********OF Computation iteration =2*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done2

  *********OF Computation iteration =3*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done3

  *********OF Computation iteration =4*********

 Data copied from host to device

 kernel args set

 flow_rows =144flow_cols=256flow_in_rows =144flow_in_cols =256

 level0calls done4

 OF done

 **********Corner Update Computation*******************

 kernel args set

 flow_rows =144flow_cols=256num of corners=47harris_flag=0

 Corner Update called

 Corner Update done

 OF done

 Num of corners = 47
***************************************************
------------------------------------------------------------

Profiling

The corner tracker design is validated on Alveo U200 board at 300 MHz frequency. The hardware resource utilizations are listed in the following table.

Table 1 Hardware resources for Corner Tracker

Name	Dataset	LUT	BRAM	FF	DSP
cornertracker	2x4K images, 8 NPPC	37547	225	35089	115
cornertracker	2xFull HD images, 8 NPPC	34624	129	34198	115

The performance is shown below

Table 2 Performance numbers in terms of FPS (Frames Per Second) for 2 consecutive frames for Corner Tracker

Dataset	FPS(CPU)	FPS(FPGA)
4k (3840x2160)	0.53	3
Full HD(1920x1080)	2	11