|
<< 点击显示目录 >> 主页 mappVision帮助助手 > 机器视觉帮助 > mapp Vision > 用例 > Blob - 自学和后续对象搜索 |
•定义 NumSearchMax。
•添加新模型。
•使用示教机制放置十字光标，以确定模型所需的像素区域。
•定义 "执行兴趣区域"（Execution ROI），以确定搜索或教入过程的范围。
•多次教入模型可以获得更精确的数值。每次示教过程后的新值都会更新为模型参数的上下限。
•使用 "提交 "完成模型教入。
•使用 TestExecute 检查是否找到所有对象。
•可以手动优化模型参数，使模型在搜索时更加稳健。每次手动更改模型参数时都必须提交模型。
信息：
mapp Vision HMI 应用程序的吸管功能可用于确定灰度值的阈值。
•使用 RegionFeature 可使搜索更加稳健。
•优化执行兴趣区域（限制区域）。
信息：
对于获取的图像显示透视变形的对象，必须注意的是，对象越靠近图像边缘，灰度值和区域就会发生变化。在某些情况下，这会导致使用默认模型参数无法识别物体。
因此，建议将这些位于室外区域的物体也纳入教导过程。这样可以简化灰度值所需阈值的确定工作，从而在所有所需的图像区域内进行更稳健的搜索。
信息：
如果对象被图像边界或执行兴趣区域切断，则会通过参数 Clipped 显示出来。
•优化执行兴趣区域（限制区域）。
•不要使用多余的模型。每个模型的运行时间都会增加。
This section describes the basic approach to configuring vision function Blob for standard applications.
•Define NumSearchMax.
•Add a new model.
•Place the cross-hair pointer to determine the pixel regions necessary for the model using the teach-in mechanism.
•Define the Execution ROI in order to define the range of the search or teach-in process.
•Teaching-in a model multiple times can result in more accurate values. The new values after each teach-in process are updated as the upper and lower limits of the model parameters.
•Complete the model teach-in with "Submit".
•Use TestExecute to check whether all objects are found.
•Possibly optimize the model parameters manually to make the model more robust for the search. The model must be submitted every time a manual change is made to the model parameters.
Information:
The pipette function of the mapp Vision HMI application can be used to determine the threshold values for the grayscale value.
•Use RegionFeature to make a search more robust.
•Optimize the Execution ROI (restricted area).
Information:
In the case of objects whose acquired image shows a perspective distortion, it must be noted that the grayscale values and area change the further to the edge of the image the object is located. Under certain circumstances, this results in no recognition of the object with the default model parameters.
It is therefore recommended that these objects, which are located in the outdoor area, are also included in the teach-in process. This simplifies determination of the required threshold values for the grayscale value for a more robust search in all desired image areas.
Information:
If an object is cut off by the image border or the Execution ROI, this is indicated by parameter Clipped.
•Optimize the Execution ROI (restricted area).
•Do not use superfluous models. The runtime increases for each model.