Dot Placement Analysis Using a Line Scan Camera and Rigid Body Rotation

Printer manufacturers design print heads and halftone patterns to result in specific types of image characteristics. Many image quality artifacts are a result of differences in actual dot placement from intended dot placement. Quantifying dot placement variations can be a complex endeavor requiring image capture at high magnification for adequate sampling. Large dot fields are imaged in pieces, which requires re-stitching of images prior to assessment, and extensive post processing to calculate actual dot positions compared with intended positions. This paper discusses a novel approach to dot placement analysis. This method includes a single, high magnification image, which is captured using a line-scan camera. Once the image is captured, the image is analyzed using principles of rigid body rotation to match a set of actual dot positions to a template of expected positions. Using this method, errors in dot placement can be assessed rapidly without errors that stem from image stitching. Introduction Controlling dot position is a key factor in controlling and effecting image quality in dot-based imaging systems. Often, halftone algorithms are developed to ensure minimal visibility. However, while the pattern is being rendered by the print engine, variations in imaging head speed or balance, jet-performance (in ink jet systems), paper handling, and other mechanical causes can disrupt the positions of the dots in the final pattern. In addition, interactions between marking and receiving media can further confound the rendering process by causing dot break-up or other effects. Dot placement errors not only effect the quality halftoned regions, but they also effect other features such as line fidelity (particularly that of fine lines) and text. Often the disparity between intended dot positions and actual dot positions is considerable. In some cases the positional errors can accumulate across a dot field. For example, in laser-based print engines, polygon mirror alignment issues can result in increasing errors in dot placement as one moves across the image in the cross process (or cross machine) direction. Figure 1a shows an example of a small portion of an intended dot pattern, while figure 1b shows an example of a the same dot pattern with some positional errors. Figure 1a: Dot pattern with dots in their intended positions. Figure 1b: Dot pattern with positional errors Figure 2a shows a plot of the intended dot positions (as depicted by their centroid locations) and the actual dot positions from the dot patterns shown in figure 1a and 1b. Figure 2a: Plot of dot positions Figure 2b shows a plot of the positions of the actual dots versus the intended dots as a function of the vector distances of their centroid locations (sqrt(X+Y)). Figure 2b: Plot of dot positions If the dot positions were identical, Y would equal X and the R value would be