org.openpnp.vision.pipeline.stages.MatchPartTemplate

org.openpnp.vision.pipeline.stages.MatchPartTemplate

功能

用于在图像中匹配单个零件并校正其方向。它接收一个候选的旋转矩形（RotatedRect）和一张模板图像，通过模板匹配确定零件的精确角度（在 0°、90°、180°、270° 四个离散方向中选择最佳匹配），并输出校正后的旋转矩形。

参数

例子

importcv2importnumpy as np def generate_all_images():# 1. 模板图像（正面红色矩形，宽200高100）template=np.zeros((100,200,3),dtype=np.uint8)cv2.rectangle(template,(0,0),(200,100),(0,0,255),-1)cv2.imwrite("template_rect.png",template)print("生成模板:template_rect.png")#2.测试图像（蓝色背景640x480） img=np.full((480,640,3),(255,0,0),dtype=np.uint8)# 红色矩形1：中心(150,150)，尺寸150x80，角度30°（完全在图像内，左边界不超出） rect1=((150,150),(150,80),30)box1=cv2.boxPoints(rect1)box1=np.int32(box1)cv2.fillPoly(img,[box1],(0,0,255))# 红色矩形2：中心 (480, 280)，尺寸 150x80，角度 120°（完全在图像内，不超出右边界和下边界）rect2=((480,280),(150,80),120)box2=cv2.boxPoints(rect2)box2=np.int32(box2)cv2.fillPoly(img,[box2],(0,0,255))# 绿色三角形（干扰物，不与红色矩形重叠，完全在图像内）# 放置在右下角空白区域，红色矩形2右边界约624，下边界约450，三角形选在(600, 440)附近pts=np.array([[580,430],[620,460],[560,460]], np.int32)cv2.fillPoly(img,[pts],(0,255,0))cv2.imwrite("test_multi_parts.png", img)print("生成测试图像: test_multi_parts.png (两个红色矩形+一个绿色三角形，均不重叠且完全在图像内)")if__name__=="__main__":generate_all_images()

<cv-pipeline><stages><cv-stageclass="org.openpnp.vision.pipeline.stages.ImageRead"name="readTest"enabled="true"file="D:\3rd\openpnp_prj\openpnp-official\openpnp-test-images\my_test\test_multi_parts.png"color-space="Bgr"handle-as-captured="false"/><cv-stageclass="org.openpnp.vision.pipeline.stages.ConvertColor"name="testBgr2Hsv"enabled="true"conversion="Bgr2HsvFull"/><cv-stageclass="org.openpnp.vision.pipeline.stages.MaskHsv"name="extractRedTest"enabled="true"auto="false"fraction-to-mask="0.0"hue-min="235"hue-max="21"saturation-min="50"saturation-max="255"value-min="50"value-max="255"soft-edge="0"soft-factor="1.0"invert="true"binary-mask="true"property-name="MaskHsv"/><cv-stageclass="org.openpnp.vision.pipeline.stages.FindContours"name="contoursTest"enabled="true"retrieval-mode="External"approximation-method="Simple"/><cv-stageclass="org.openpnp.vision.pipeline.stages.MinAreaRectContours"name="candidateRects"enabled="true"contours-stage-name="contoursTest"/><cv-stageclass="org.openpnp.vision.pipeline.stages.FilterRects"name="filterSize"enabled="true"width-max="200.0"width-min="80.0"length-max="120.0"length-min="50.0"aspect-ratio-max="0.0"aspect-ratio-min="0.0"enable-logging="true"rotated-rects-stage-name="candidateRects"/><cv-stageclass="org.openpnp.vision.pipeline.stages.SelectSingleRect"name="selectRect"enabled="true"position="0"rotated-rects-stage-name="filterSize"/><cv-stageclass="org.openpnp.vision.pipeline.stages.ImageRead"name="readTemplateColor"enabled="true"file="D:\3rd\openpnp_prj\openpnp-official\openpnp-test-images\my_test\template_rect.png"color-space="Bgr"handle-as-captured="false"/><cv-stageclass="org.openpnp.vision.pipeline.stages.ConvertColor"name="templateBgr2Hsv"enabled="true"conversion="Bgr2HsvFull"/><cv-stageclass="org.openpnp.vision.pipeline.stages.MaskHsv"name="templateBinary"enabled="true"auto="false"fraction-to-mask="0.0"hue-min="235"hue-max="21"saturation-min="0"saturation-max="255"value-min="0"value-max="255"soft-edge="0"soft-factor="1.0"invert="false"binary-mask="true"property-name="MaskHsv"/><cv-stageclass="org.openpnp.vision.pipeline.stages.ImageRecall"name="recallTest"enabled="true"image-stage-name="readTest"/><cv-stageclass="org.openpnp.vision.pipeline.stages.ConvertColor"name="testGray"enabled="true"conversion="Bgr2Gray"/><cv-stageclass="org.openpnp.vision.pipeline.stages.MatchPartTemplate"name="matchPart"enabled="true"log="true"template-stage-name="templateBinary"model-stage-name="selectRect"threshold="0.1"/><cv-stageclass="org.openpnp.vision.pipeline.stages.ImageRecall"name="recallOriginal"enabled="true"image-stage-name="readTest"/><cv-stageclass="org.openpnp.vision.pipeline.stages.DrawRotatedRects"name="drawCorrected"enabled="true"rotated-rects-stage-name="matchPart"thickness="4"draw-rect-center="false"rect-center-radius="20"show-orientation="false"><colorr="255"g="255"b="255"a="255"/></cv-stage><cv-stageclass="org.openpnp.vision.pipeline.stages.ImageWrite"name="save"enabled="true"file="output_matched_single.png"/></stages></cv-pipeline>

效果

END