I have converted a landsat tm to polygons using raster to polygon tools of arctoolboxs. But I get some irregular and square polygons. Can I get some square polygons and its size is equal to the pixel of tm imagery? Thank you.
Thank your answer. But the image include so many pixels. If I digitiz it, it will consume so time. In addition, I dont know how to binarize the image at each pixel of image. Can you give me some detail of procedure.
Maybe you can try creating a fishnet in data management\feature class, using the template of your landsat image and than transfer value from pixels to polygons using zonal statistic.
It sounds to me like you want each pixel of the TM imagery to be its own polygon, but if you have neighbouring cells with the same value these will be aggregated into the same polygon through the tool you are using. Can you tell us what the end goal is (i.e. why would someone want every pixel to be a cell) and we can better assist?
You can accomplish this task if you convert the raster (image) to a polygon. However, you need to ensure each grid cell has a unique value or adjoining cells with the same value will result in one polygon. It think you could get around this if you create a random (value) raster using the image as your master raster (snap to it) See "create random raster" in spatial analyst. Then convert the random raster to polygon. As others have suggested, this might not be the most efficient way to reach your end goal, but it should be possible with random raster.
Thanks for your answers, I have created a fishnet grid and random raster. Meanwhile, I known how to binarize raster. These answers are very helpful for me. The fishnet method can work. The random raster's value is floating type, but the raster to polygon only input integer raster dataset. So it can not be done.
Sorry I dont firstly tell the goal . My end goal is to get imperviousness from Landsat imagery for my research work. So I plan to create 1000 random square polygons, then overlay these polyons on the high resolution image (esri's world imagery) to digitize imperviousness. Do you think my research method is correct? Hope you can propose some helpful suggestions.
Thanks for you answers. In arcmap 9.3, I dont find image basemap about my study area. Even if I buy some high resolution image, I want to know how to create relation in band value and imperviousness(%).
Thank you. Under your suggestions, I visit esri's world imagery and find the basemap. The screen snap is attached. The image is so high resolution. Then I download the layer files, but it be opened in Arcglobe. Can I operate the imagery in arcmap? Is Its projection the UTM? How do my landsat tm image overay the world images?
you can try not to switch the option "simplify polygons" in tool raster to polygon, it should make only square polygons...or you can create a feature dataset in a geodatabase, the result should be the same.
Under your suggestions, when I selected the option, I got the attachment1. When I dont select the option, I got the attachment2. It should not make only square polygons( the black circle part). Elvio, if I have a irregular polygon layer and a square polygon layer, I want compute the percent of area within each polygon. What should I do in arcgis?