Latest Contributions by DonRaymond

Back again, Dan. I have the attached histogram out now. I am working with a smaller AOI/subset of the original raster. I have also delimited the lower bound to the min val in the array, and set the upper bound using a 'where' statement as you have suggested (1.25 st dev * mean). I assume if I set the lower bound to an artificial value with (mean - (1.25 * st dev)) and there are no data values it may create empty bins? I would prefer to have all bins populated. By the looks of the histogram above, it might also be beneficial to have a separate zero-only bin. Not sure if that can be implemented and still retain the other 20 bin ranges.    stds = [arr_min, arr_max]    new_Arr = np.where((testImage_arr < stds[0]) | (testImage_arr > stds[1]), 0, testImage_arr) Printing the new np.histogram array yields this: (array([226603, 939, 1265, 1468, 1548, 1474, 1537, 1492, 1463, 1341, 1365, 1291, 1203, 1159, 1079, 1041, 1027, 967, 889, 849]), array([ 0. , 0.00114019, 0.00228038, 0.00342057, 0.00456075, 0.00570094, 0.00684113, 0.00798132, 0.00912151, 0.0102617 , 0.01140189, 0.01254207, 0.01368226, 0.01482245, 0.01596264, 0.01710283, 0.01824302, 0.0193832 , 0.02052339, 0.02166358, 0.02280377])) I think this looks pretty good, but I now need to get only the counts out into a new array that I can plot. A final format might look something like this (generated in Excel using random numbers). Is there a simple way to grab the count portion only and convert that to plot as a new histogram? Thanks. ... View more

Sorry, I took that to be like a conditional statement, whereby everything within the brackets would be set to the image values in the array, and everything excluded would be set to -1 (null). In your array you seem to keep the values outside the required range and set everything inside the range to -1. Clearly I had the syntax/logic all wrong. Thanks. It would help if I read the entire post instead of missing your final sentence. Seems I would want this usage: new_Arr = np.where((testImage_arr < stds[0]) | (testImage_arr > stds[1]), -1, testImage_arr) ... View more

Thanks Dan. Yes, lots of reading still to do on Numpy. I must be missing something here because it still isn't working properly. #Set Environment import arcpy from arcpy.sa import * from arcpy import env import numpy as np import sys, os, string, glob import matplotlib.pyplot as plt arcpy.CheckOutExtension("spatial") arcpy.env.overwriteOutput = True #Identify a test raster testImage = arcpy.Raster(r"F:\GIS_Dev\RCM_VAPs_Validate\0Comparisons\RS2_SQ28_SLC\freeman\reports\gamma_dbl_clip.tif") #Convert Raster to Numpy Array - raster is 1677 columns x 7155 rows, floating point 32 bit testImage_arr = arcpy.RasterToNumPyArray(testImage,lowerLeft,1677,7155, nodata_to_value=0) #Set some variables for stats calculations sdFact = 2.5 nb_Classes = 20 #Calculate the min, max, mean and standard deviation of the array testImage_min = np.min(testImage_arr) testImage_max = np.max(testImage_arr) testImage_mean = np.mean(testImage_arr) testImage_std = np.std(testImage_arr) #Calculate a min and max based on sdFact, mean, and sdev arr_min = (testImage_mean - (sdFact/2) * testImage_std) arr_max = (testImage_mean + (sdFact/2) * testImage_std) arr_range = arr_max - arr_min print arr_range stds = [arr_min, arr_max] print stds new_Arr = np.where((testImage_arr > stds[0]) | (testImage_arr < stds[1]), testImage_arr, -1) plt.hist(new_Arr, bins=nb_Classes) plt.title("Histogram") plt.xlabel("Value") plt.ylabel("Count") plt.grid(True) plt.show() This yields the attached graph, which seems to include all values in the original raster, and not only the masked range. The histogram is compressed to the low end and we want to expose greater detail here by binning the desired range (arr_range = 0.00555599341169) across 20 bins. For the test raster:  min = 0 max = 0.349763 mean = 8.98602e-05 st dev = 0.0022224 ... View more

So I have the array from the input raster. Is there an efficient way to define the range I want based on mean and st dev? We would like only values that fall 1.25 st dev on either side of the mean. I found the numpy.clip function but it seems to reset data values above and below the specified min/max thresholds to those thresholds and what I need is to eliminate those values completely. I have extracted the mean and st dev from the original array as follows (sdFact previously defined as 2.5): #Calculate the min, max, mean and standard deviation of the array       testImage_min = np.min(testImage_arr)       testImage_max = np.max(testImage_arr)       testImage_mean = np.mean(testImage_arr)       testImage_std = np.std(testImage_arr)   #Calculate a min and max based on sdFact, mean, and sdev       arr_min = (testImage_mean - (sdFact/2) * testImage_std)     arr_max = (testImage_mean + (sdFact/2) * testImage_std) ... View more

Thanks for your help, Dan. I will give this a try. ... View more

For anyone interested, the following code solved this problem. mask = arcpy.GetParameterAsText(3) MinX, MinY, MaxX, MaxY = mask.split( " ") MinX = float( MinX) MinY = float( MinY) MaxX = float( MaxX) MaxY = float( MaxY)‍‍‍‍‍‍‍‍‍‍‍‍ The variables were then passed as follows: outExtent = Extent(MinX, MinY, MaxX, MaxY) outConstRaster = CreateConstantRaster(constantValue, "INTEGER",                                       cellSize, outExtent)‍‍‍‍‍‍‍‍‍‍‍‍ (Curtis Price‌ added code formatting) ... View more

Thanks very much for your help, Jake. I will give this a try today. ... View more