In this program, an image must be transformed to a binary image by thresholding before analysis, as is necessary for other image analyses. The thresh-old value is currently set manually for every image in ImageJ. Consequently, when analyzing numerous images, this protocol is inefficient despite rapid measurement for a single image. Additionally, it is not yet user-friendly because setting of the appropriate threshold values for image processing (i.e., creating the binary images) is expected to be difficult for unskilled users. Therefore, an automatic and reliable thresholding method should be developed for estimating root length in ImageJ.
The objective of this study was to compare the 16 thresholding methodsofimageprocessingforbinaryimagestoestimatericeroot length automatically using ImageJ. Then the datasets of root length estimated by ImageJ were compared with those by WinRHIZO.