binocular_disparitybinocular_disparityBinocularDisparityBinocularDisparitybinocular_disparity双目视差（算子）

名称

binocular_disparitybinocular_disparityBinocularDisparityBinocularDisparitybinocular_disparity — 使用相关技术计算校正图像对的视差。

签名

描述

binocular_disparitybinocular_disparityBinocularDisparityBinocularDisparityBinocularDisparitybinocular_disparity 通过相关性技术计算两幅校正图像间的像素级对应关系。与 binocular_distancebinocular_distanceBinocularDistanceBinocularDistanceBinocularDistancebinocular_distance 不同，其结果不会转换为距离值。

该算法需要一张参考图像 ImageRect1ImageRect1ImageRect1ImageRect1imageRect1image_rect_1 和一张搜索图像 ImageRect2ImageRect2ImageRect2ImageRect2imageRect2image_rect_2，两者必须经过校正处理，即对应的极线相互平行且位于相同的图像行上（）。若该假设不成立，可通过 calibrate_camerascalibrate_camerasCalibrateCamerasCalibrateCamerasCalibrateCamerascalibrate_cameras、gen_binocular_rectification_mapgen_binocular_rectification_mapGenBinocularRectificationMapGenBinocularRectificationMapGenBinocularRectificationMapgen_binocular_rectification_map 和 map_imagemap_imageMapImageMapImageMapImagemap_image 算子对图像进行校正。因此，对于参考图像 ImageRect1ImageRect1ImageRect1ImageRect1imageRect1image_rect_1 中的某个像素，通过在 ImageRect2ImageRect2ImageRect2ImageRect2imageRect2image_rect_2 中沿对应行搜索，并匹配尺寸为 MaskWidthMaskWidthMaskWidthMaskWidthmaskWidthmask_width × MaskHeightMaskHeightMaskHeightMaskHeightmaskHeightmask_height 的矩形窗口内的局部邻域，从而选定 ImageRect2ImageRect2ImageRect2ImageRect2imageRect2image_rect_2 中的对应像素。像素对应关系通过单通道视差图像 DisparityDisparityDisparityDisparitydisparitydisparity 中的 返回，该图像为参考图像 ImageRect1ImageRect1ImageRect1ImageRect1imageRect1image_rect_1 中的每个像素 指定 ImageRect2ImageRect2ImageRect2ImageRect2imageRect2image_rect_2 中合适的匹配像素 ，其计算依据为 。每个视差值的质量度量值存储于 ScoreScoreScoreScorescorescore 中，该值包含参考像素匹配函数 S 的最佳匹配结果。匹配过程直接采用原始未处理图像的灰度值。

所使用的匹配函数由参数 MethodMethodMethodMethodmethodmethod 定义，该参数分配三种不同的相关性：

• 'sad'"sad""sad""sad""sad""sad"：绝对差值之和 其中 。

• 'ssd'"ssd""ssd""ssd""ssd""ssd"：平方差之和 其中 。

• 'ncc'"ncc""ncc""ncc""ncc""ncc"：归一化互相关 其中 。

其中

、、、：两幅输入图像中对应像素的行和列坐标，

、：未处理输入图像的灰度值，

：相关性窗口的大小

：在宽度为 、高度为 的相关窗口内的均值。

请注意，'sad'"sad""sad""sad""sad""sad" 和 'ssd'"ssd""ssd""ssd""ssd""ssd" 方法直接比较掩模窗口内像素的灰度值，而 'ncc'"ncc""ncc""ncc""ncc""ncc" 方法则会补偿掩模窗口内的平均灰度值及其方差。因此，当两幅图像存在亮度和对比度差异时，应优先选用此方法。对于亮度和对比度相似的图像，则推荐使用 'sad'"sad""sad""sad""sad""sad" 和 'ssd'"ssd""ssd""ssd""ssd""ssd" 方法，因其内部计算复杂度较低而运行速度更快。

需要注意的是，对于上升的 ，相关性质量在 'sad'"sad""sad""sad""sad""sad" 和 'ssd'"ssd""ssd""ssd""ssd""ssd" 方法中呈下降趋势（最佳质量值为 0），而在 'ncc'"ncc""ncc""ncc""ncc""ncc" 方法中则呈上升趋势（最佳质量值为1.0）。

相关窗口的大小由 和 表示，必须为奇数，并通过 MaskWidthMaskWidthMaskWidthMaskWidthmaskWidthmask_width 和 MaskHeightMaskHeightMaskHeightMaskHeightmaskHeightmask_height 参数传递。搜索空间受最小视差值 MinDisparityMinDisparityMinDisparityMinDisparityminDisparitymin_disparity 和最大视差值 MaxDisparityMaxDisparityMaxDisparityMaxDisparitymaxDisparitymax_disparity 的限制。由于图像边界之外的像素值未定义，因此 DisparityDisparityDisparityDisparitydisparitydisparity 和 ScoreScoreScoreScorescorescore 的域在图像边界处未设置：上下边界各保留 (MaskHeightMaskHeightMaskHeightMaskHeightmaskHeightmask_height-1)/2 的高度边界，左右边界各保留 (MaskWidthMaskWidthMaskWidthMaskWidthmaskWidthmask_width-1)/2 的宽度边界。基于相同原因，图像左右边界处的最大视差范围亦相应缩减。

由于匹配在处理纹理较差的区域时可靠性极低，可在 TextureThreshTextureThreshTextureThreshTextureThreshtextureThreshtexture_thresh 中定义相关窗口内灰度值的最小统计分布范围。该阈值同时应用于输入图像 ImageRect1ImageRect1ImageRect1ImageRect1imageRect1image_rect_1 和 ImageRect2ImageRect2ImageRect2ImageRect2imageRect2image_rect_2。此外， ScoreThreshScoreThreshScoreThreshScoreThreshscoreThreshscore_thresh 通过设定相关函数的最大值（'sad'"sad""sad""sad""sad""sad",'ssd'"ssd""ssd""ssd""ssd""ssd"）或最小值（'ncc'"ncc""ncc""ncc""ncc""ncc"）来保证匹配质量。将 FilterFilterFilterFilterfilterfilter 设为 'left_right_check'"left_right_check""left_right_check""left_right_check""left_right_check""left_right_check"可增强匹配结果的稳健性，因其同时依赖正向与反向匹配；而选择 'none'"none""none""none""none""none" 则关闭该功能。

用于提升 binocular_disparitybinocular_disparityBinocularDisparityBinocularDisparityBinocularDisparitybinocular_disparity 时间响应的金字塔层数由 NumLevelsNumLevelsNumLevelsNumLevelsnumLevelsnum_levels 参数决定。遵循粗到细的方案，先计算高阶视差图像，再将其分割为视差相似的矩形子图像，从而缩小下一低阶金字塔层的视差范围。纹理阈值（TextureThreshTextureThreshTextureThreshTextureThreshtextureThreshtexture_thresh）与分数阈值（ScoreThreshScoreThreshScoreThreshScoreThreshscoreThreshscore_thresh）在各层级应用，最终生成的 DisparityDisparityDisparityDisparitydisparitydisparity 与 ScoreScoreScoreScorescorescore 图像域，由所有单层处理结果的交集域构成。通常而言，当视差图像能被分割为更多均匀视差区域且指定视差范围越大时，金字塔结构的优势越显著。其弊端在于粗层级可能丢失关键纹理信息，导致视差值计算不足。

最后，参数 SubDisparitySubDisparitySubDisparitySubDisparitysubDisparitysub_disparity 的取值 'interpolation'"interpolation""interpolation""interpolation""interpolation""interpolation" 将执行视差值的亚像素级优化。若将该参数设为 'none'"none""none""none""none""none" 则关闭该功能。

执行信息

• 多线程类型：可重入（与非独占算子并行运行）。
• 多线程作用域：全局（可从任何线程调用）。
• 在内部数据级别上自动并行化。

参数

* Set internal and external stereo parameters.
* Note that, typically, these values are the result of a prior
* calibration.
gen_cam_par_area_scan_division (0.01, -665, 5.2e-006, 5.2e-006, \
                                622, 517, 1280, 1024, CamParam1)
gen_cam_par_area_scan_division (0.01, -731, 5.2e-006, 5.2e-006, \
                                654, 519, 1280, 1024, CamParam2)
create_pose (0.1535,-0.0037,0.0447,0.17,319.84,359.89, \
             'Rp+T', 'gba', 'point', RelPose)

* Compute the mapping for rectified images.
gen_binocular_rectification_map (Map1, Map2, CamParam1, CamParam2, RelPose, \
                                 1, 'viewing_direction', 'bilinear', \
                                 CamParamRect1,CamParamRect2, \
                                 Cam1PoseRect1, Cam2PoseRect2,RelPoseRect)

* Compute the disparities in online images.
while (1)
  grab_image_async (Image1, AcqHandle1, -1)
  map_image (Image1, Map1, ImageRect1)

  grab_image_async (Image2, AcqHandle2, -1)
  map_image (Image2, Map2, ImageRect2)

  binocular_disparity(ImageRect1, ImageRect2, Disparity, Score, 'sad', \
                      11, 11, 20, -40, 20, 2, 25, 'left_right_check', \
                      'interpolation')
endwhile

* Set internal and external stereo parameters.
* Note that, typically, these values are the result of a prior
* calibration.
gen_cam_par_area_scan_division (0.01, -665, 5.2e-006, 5.2e-006, \
                                622, 517, 1280, 1024, CamParam1)
gen_cam_par_area_scan_division (0.01, -731, 5.2e-006, 5.2e-006, \
                                654, 519, 1280, 1024, CamParam2)
create_pose (0.1535,-0.0037,0.0447,0.17,319.84,359.89, \
             'Rp+T', 'gba', 'point', RelPose)

* Compute the mapping for rectified images.
gen_binocular_rectification_map (Map1, Map2, CamParam1, CamParam2, RelPose, \
                                 1, 'viewing_direction', 'bilinear', \
                                 CamParamRect1,CamParamRect2, \
                                 Cam1PoseRect1, Cam2PoseRect2,RelPoseRect)

* Compute the disparities in online images.
while (1)
  grab_image_async (Image1, AcqHandle1, -1)
  map_image (Image1, Map1, ImageRect1)

  grab_image_async (Image2, AcqHandle2, -1)
  map_image (Image2, Map2, ImageRect2)

  binocular_disparity(ImageRect1, ImageRect2, Disparity, Score, 'sad', \
                      11, 11, 20, -40, 20, 2, 25, 'left_right_check', \
                      'interpolation')
endwhile

* Set internal and external stereo parameters.
* Note that, typically, these values are the result of a prior
* calibration.
gen_cam_par_area_scan_division (0.01, -665, 5.2e-006, 5.2e-006, \
                                622, 517, 1280, 1024, CamParam1)
gen_cam_par_area_scan_division (0.01, -731, 5.2e-006, 5.2e-006, \
                                654, 519, 1280, 1024, CamParam2)
create_pose (0.1535,-0.0037,0.0447,0.17,319.84,359.89, \
             'Rp+T', 'gba', 'point', RelPose)

* Compute the mapping for rectified images.
gen_binocular_rectification_map (Map1, Map2, CamParam1, CamParam2, RelPose, \
                                 1, 'viewing_direction', 'bilinear', \
                                 CamParamRect1,CamParamRect2, \
                                 Cam1PoseRect1, Cam2PoseRect2,RelPoseRect)

* Compute the disparities in online images.
while (1)
  grab_image_async (Image1, AcqHandle1, -1)
  map_image (Image1, Map1, ImageRect1)

  grab_image_async (Image2, AcqHandle2, -1)
  map_image (Image2, Map2, ImageRect2)

  binocular_disparity(ImageRect1, ImageRect2, Disparity, Score, 'sad', \
                      11, 11, 20, -40, 20, 2, 25, 'left_right_check', \
                      'interpolation')
endwhile

* Set internal and external stereo parameters.
* Note that, typically, these values are the result of a prior
* calibration.
gen_cam_par_area_scan_division (0.01, -665, 5.2e-006, 5.2e-006, \
                                622, 517, 1280, 1024, CamParam1)
gen_cam_par_area_scan_division (0.01, -731, 5.2e-006, 5.2e-006, \
                                654, 519, 1280, 1024, CamParam2)
create_pose (0.1535,-0.0037,0.0447,0.17,319.84,359.89, \
             'Rp+T', 'gba', 'point', RelPose)

* Compute the mapping for rectified images.
gen_binocular_rectification_map (Map1, Map2, CamParam1, CamParam2, RelPose, \
                                 1, 'viewing_direction', 'bilinear', \
                                 CamParamRect1,CamParamRect2, \
                                 Cam1PoseRect1, Cam2PoseRect2,RelPoseRect)

* Compute the disparities in online images.
while (1)
  grab_image_async (Image1, AcqHandle1, -1)
  map_image (Image1, Map1, ImageRect1)

  grab_image_async (Image2, AcqHandle2, -1)
  map_image (Image2, Map2, ImageRect2)

  binocular_disparity(ImageRect1, ImageRect2, Disparity, Score, 'sad', \
                      11, 11, 20, -40, 20, 2, 25, 'left_right_check', \
                      'interpolation')
endwhile

* Set internal and external stereo parameters.
* Note that, typically, these values are the result of a prior
* calibration.
gen_cam_par_area_scan_division (0.01, -665, 5.2e-006, 5.2e-006, \
                                622, 517, 1280, 1024, CamParam1)
gen_cam_par_area_scan_division (0.01, -731, 5.2e-006, 5.2e-006, \
                                654, 519, 1280, 1024, CamParam2)
create_pose (0.1535,-0.0037,0.0447,0.17,319.84,359.89, \
             'Rp+T', 'gba', 'point', RelPose)

* Compute the mapping for rectified images.
gen_binocular_rectification_map (Map1, Map2, CamParam1, CamParam2, RelPose, \
                                 1, 'viewing_direction', 'bilinear', \
                                 CamParamRect1,CamParamRect2, \
                                 Cam1PoseRect1, Cam2PoseRect2,RelPoseRect)

* Compute the disparities in online images.
while (1)
  grab_image_async (Image1, AcqHandle1, -1)
  map_image (Image1, Map1, ImageRect1)

  grab_image_async (Image2, AcqHandle2, -1)
  map_image (Image2, Map2, ImageRect2)

  binocular_disparity(ImageRect1, ImageRect2, Disparity, Score, 'sad', \
                      11, 11, 20, -40, 20, 2, 25, 'left_right_check', \
                      'interpolation')
endwhile

结果

binocular_disparitybinocular_disparityBinocularDisparityBinocularDisparityBinocularDisparitybinocular_disparity 在所有参数值正确时返回 2 ( H_MSG_TRUE )。如有必要，则抛出异常。

可能的前趋

map_imagemap_imageMapImageMapImageMapImagemap_image

可能的后继

thresholdthresholdThresholdThresholdThresholdthreshold, disparity_to_distancedisparity_to_distanceDisparityToDistanceDisparityToDistanceDisparityToDistancedisparity_to_distance, disparity_image_to_xyzdisparity_image_to_xyzDisparityImageToXyzDisparityImageToXyzDisparityImageToXyzdisparity_image_to_xyz

替代

binocular_disparity_mgbinocular_disparity_mgBinocularDisparityMgBinocularDisparityMgBinocularDisparityMgbinocular_disparity_mg, binocular_disparity_msbinocular_disparity_msBinocularDisparityMsBinocularDisparityMsBinocularDisparityMsbinocular_disparity_ms, binocular_distancebinocular_distanceBinocularDistanceBinocularDistanceBinocularDistancebinocular_distance, binocular_distance_mgbinocular_distance_mgBinocularDistanceMgBinocularDistanceMgBinocularDistanceMgbinocular_distance_mg, binocular_distance_msbinocular_distance_msBinocularDistanceMsBinocularDistanceMsBinocularDistanceMsbinocular_distance_ms

另见

map_imagemap_imageMapImageMapImageMapImagemap_image, gen_binocular_rectification_mapgen_binocular_rectification_mapGenBinocularRectificationMapGenBinocularRectificationMapGenBinocularRectificationMapgen_binocular_rectification_map, binocular_calibrationbinocular_calibrationBinocularCalibrationBinocularCalibrationBinocularCalibrationbinocular_calibration

模块

三维计量