In addition to the inspection tools, users can also configure pre-processing filters on the image to improve the result of the inspections. The different filters are described below:
The Open filter can be used to separate white objects on dark backgrounds from each other, useful in cases where multiple segments are being identified as a single entity. It is the opposite of the Close filter, and is equivalent to an Erode followed by a Dilate.
The Close filter can be used to connect white objects on dark backgrounds to each other. A common use case is to apply this filter to dot-matrix-printed text such that they become connected as a single character for easier segmentation. It is the opposite of the Open filter, and is equivalent to a Dilate followed by an Erode.
The Erode filter will shrink, or decrease, the size of a white object on a dark background. It has similar applications to the Open filter, except the Open filter will apply a dilation as well, returning the object to the original size. This is the opposite of the Dilate filter.
The Dilate filter will grow, or increase, the size of a white object on a dark background. It has similar applications to the Close filter, except the Close filter will apply an erosion as well, returning the object to the original size. This is the opposite of the Erode filter.
The Invert filter will invert the polarity of each pixel within the image. In other words, it will turn dark pixels into light pixels and vice versa.
The Color filter will allow you to highlight pixels within a specific range in your image by converting pixels within the color range selected into black. If you select the “binarize” option, it will also convert all pixels outside of your specific range into white pixels.
The Threshold filter behaviors similar to the color filter but instead of selecting a range, you just need to set a threshold level. All pixels darker than the threshold will be converted into black pixels and pixels below the threshold will be converted to white pixels. By default, the “Automatic” mode of this threshold will be set but setting the algorithm to “Manual” will allow you to set the threshold for more configurability.
The Blur filter removes high frequency content from the image in order to reduce noise within the image. In other words, it introduces some blurriness to the image so that it smoothes out some of the features of interest. Four blur algorithms are offered: Box, Gaussian, Median, and Bilateral.
The Edge filter highlights edges in the image using a Sobel Filter. This allows for edges of objects in the frame to be more clear for inspection tools.