您好! 我可以根据 网站 内容通过默认配置文件和数据集成功运行 DMAPAC 的立体视差应用。 ,我将该"ar差_max"值设置为2 μ s、则 会导致以下误差:
493157.467914 s: VX_ZONE_ERROR:[ownContextSendCmd:814] Command ack message returned failure cmd_status: -1 493157.467933 s: VX_ZONE_ERROR:[ownContextSendCmd:850] tivxEventWait() failed. 493157.467962 s: VX_ZONE_ERROR:[ownNodeKernelInit:538] Target kernel, TIVX_CMD_NODE_CREATE failed for node Stereo_Processing 493157.467981 s: VX_ZONE_ERROR:[ownNodeKernelInit:539] Please be sure the target callbacks have been registered for this core 493157.467999 s: VX_ZONE_ERROR:[ownNodeKernelInit:540] If the target callbacks have been registered, please ensure no errors are occurring within the create callback of this kernel 493157.468018 s: VX_ZONE_ERROR:[ownGraphNodeKernelInit:583] kernel init for node 0, kernel com.ti.hwa.dmpac_sde ... failed !!! 493157.468041 s: VX_ZONE_ERROR:[vxVerifyGraph:2055] Node kernel init failed 493157.468060 s: VX_ZONE_ERROR:[vxVerifyGraph:2109] Graph verify failed
"config/app.cfg "文件的内容如下:
# NOTE: You have to define the APP_STEREO_DATA_PATH environment variable,as location to stereo_test_data.
# For example:
# export APP_STEREO_DATA_PATH=<path-to-test-data>/test_data/psdkra/stereo_test_data/
# where,"<path-to-test-data>" refers to the parent directory to which you extracted 'psdk_rtos_ti_data_set_xx_xx_xx.tar.gz'
# You can even add the above line to .bashrc
# location of input image (*.bmp) files
# input image path for left camera images is construced as ${APP_STEREO_DATA_PATH}/<left_img_file_path>
# input image path for right camera images is construced as ${APP_STEREO_DATA_PATH}/<right_img_file_path>
left_img_file_path Stereo0001/left
right_img_file_path Stereo0001/right
# location for output *.bmp files
# output image path is construced as ${APP_STEREO_DATA_PATH}/<output_file_path>
output_file_path Stereo0001/output
# filename prefix, output file name is constructed as
# <filename prefix>_disparity_ddddd.bmp, <filename prefix>_conf_ddddd.bmp
# e.g. tisde_disparity_00000.bmp, tisde_conf_00000.bmp
output_file_prefix out
# first file number to pick as input
start_seq 1
# last file number to pick as input
end_seq 119
# input data width in pixels
width 1280
# input data height in lines
height 720
# confidence value for visualization <0 ~ 7>
# the disparity whose confidence is < vis_confidence will be removed
visualize_confidence 1
# <0 or 1>, disable/enable histogram output
hist_output 0
# <0 or 1>, disable/enable .sde output
sde_output 0
# enable display mode (set to 0 for PC)
display_option 0
# number of iterations (This can be set to more than one to observe PC execution time, if desired)
num_iterations 1
# enable interactive mode (set to 0 for PC)
is_interactive 0
# SDE params
# <0 or 1>, disable/enable post-processing 5x5 median filter
median_filter_enable 1
# <0 or 1>, disable/enable reduced range search on pixels near right margin
reduced_range_search_enable 0
# minimum disparity (0 -> 0 pixels, 1 -> -3 pixels)
disparity_min 0
# maximum disparity (0 -> 63 pixels, 1 -> 127 pixels, 2 -> 191 pixels)
disparity_max 2
# <0 - 255>, left-right consistency check threshold in pixels
threshold_left_right 3
# <0 or 1>, disable/enable texture based filtering
texture_filter_enable 0
# <0 - 255>, scaled threshold for texture based filtering
threshold_texture 100
# <0 - 127>, Optimization penalty constant for smalle disparity change
aggregation_penalty_p1 32
# <0 - 192>, Optimization penalty constant for large disparity change
aggregation_penalty_p2 192
# <0 - 127> confidence_score_map[7] array configuation
# To map the confidence values between 0 ~ 127 to 0 ~ 7,
# We need to specifiy 7 values in the increasing order as follows:
# confidence_score_map_0, confidence_score_map_1, ..., confidence_score_map_5, confidence_score_map_6.
#
# confidence values, c will be map to 0 ~ 7 as follows:
# c <= confidence_score_map_0 -> 0
# confidence_score_map_0 < c <= confidence_score_map_1 -> 1
# confidence_score_map_1 < c <= confidence_score_map_2 -> 2
# confidence_score_map_2 < c <= confidence_score_map_3 -> 3
# confidence_score_map_3 < c <= confidence_score_map_4 -> 4
# confidence_score_map_4 < c <= confidence_score_map_5 -> 5
# confidence_score_map_5 < c <= confidence_score_map_6 -> 6
# confidence_score_map_6 < c -> 7
# User can specify confidence_score_map_0 ~ confidence_score_map_6
# Should make sure that
# confidence_score_map_0 >= 0
# confidence_score_map_6 <= 127
confidence_score_map_0 0
confidence_score_map_1 1
confidence_score_map_2 2
confidence_score_map_3 3
confidence_score_map_4 5
confidence_score_map_5 9
confidence_score_map_6 14
期待您的回复! 谢谢!