分水岭算法（cv::watershed）。需要输入图像 + 标记图（不同整数标记不同种子区域，0 为未知），按梯度地形淹没求得分割边界。上游常用 opencv_threshold + opencv_connected_components 生成标记。

所属模块：OpenCV 图像分割 · 任务类型：opencv_watershed

端口

输入

端口数据类型必填
inImage / cv::Mat✓
markersImage / cv::Mat—

输出

端口数据类型必填
outcv::Mat✓

参数

参数类型默认值范围 / 选项说明
rect_xint-1[-1, 16384]
rect_yint-1[-1, 16384]
rect_widthint-1[-1, 16384]
rect_heightint-1[-1, 16384]

示例图

示例图为最小可运行流水线（读取 → 处理 → 写出），可直接拖入 GraphStudio 或保存为 .json 后用 --open 打开；data/test.png 替换为实际图像路径。

{
  "version": "2.0",
  "tasks": [
    {
      "id": "read1",
      "type": "opencv_image_read",
      "params": {
        "file_path": "data/test.png"
      }
    },
    {
      "id": "read2",
      "type": "opencv_image_read",
      "params": {
        "file_path": "data/test.png"
      }
    },
    {
      "id": "node",
      "type": "opencv_watershed",
      "params": {
        "rect_x": -1,
        "rect_y": -1,
        "rect_width": -1,
        "rect_height": -1
      }
    },
    {
      "id": "save",
      "type": "opencv_image_write",
      "params": {
        "file_path": "out_out.png"
      }
    }
  ],
  "edges": [
    {
      "from": "read1",
      "from_port": "out",
      "to": "node",
      "to_port": "in"
    },
    {
      "from": "read2",
      "from_port": "out",
      "to": "node",
      "to_port": "markers"
    },
    {
      "from": "node",
      "from_port": "out",
      "to": "save",
      "to_port": "in"
    }
  ]
}

本文是 GraphStudio 节点手册 系列文章之一；参数与行为以最新发布版为准。对应英文版见页面语言切换。