Index a 3D numpy array

Question

I have an image that has 7 bands with 6938x752 pixels in each band. I want to do some processing on it so am using a python module called RIOS that deals with the reading and writing of the image to let the user focus on the processing. It reads the image in as a numpy array in blocks so that the processing is more efficient than reading in the whole image.

It reads the image in as a numpy array with the shape (7, 200, 200). I want to process the data by pixel so that I have the information from each band for each pixel. Is there a way to index the array so that for I can process just the 7 values (one for each band) for each pixel within image?

The only little code I can provide here is the function I use to read in, process and write the data.

def MASKimage(info, inputs, outputs):

    inputs.image1 = inputs.image1.astype(numpy.float32) # Read the image in as an array with shape (7, 200, 200)

    process = * Do some processing on the array which is the 7 values pixel by pixel * # has shape (1,7)

    outputs.outimage = process

It is no clear what you are asking. Can you give an example of the type of processing you want to do. To get the corresponding pixel in each image you would do inputs.image1[:,0,0] for the 0th row 0th column and so on. — ebarr
– ebarr, Commented Feb 27, 2016 at 5:29

Gustavo Bezerra · Accepted Answer · 2016-02-27 15:26:47Z

1

You can try using np.transpose and np.reshape:

inputs.image1 = input.images1.transpose((1,2,0)).reshape(200*200,7)

You can just iterate through the image with a single loop to do your processing as each element represents a pixel with 7 bands.

answered Feb 27, 2016 at 15:26

Gustavo Bezerra

11.2k4 gold badges45 silver badges51 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

GeoMonkey Over a year ago

This is exactly what I needed! After some processing I am get a single pixel value. Imagine for example I want the mean of the 7 values. I want to write this to a new image but it requires a 3d array. Can you show me how to reshape this so I can write it back in the same shape as I read it (with 1 band and not 7)?

Collectives™ on Stack Overflow

Index a 3D numpy array

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related