search

terraref / reference-data

Coordination of Data Products and Standards for TERRA reference data
https://terraref.org
BSD 3-Clause "New" or "Revised" License
9 stars 2 forks source link

VNIR jpg output and livestream view have green line #250

Open NewcombMaria opened 6 years ago

NewcombMaria commented 6 years ago

@remotesensinglab and @Paheding we need your assistance on an issue with the VNIR output that we've observed since the camera has been installed. The jpg files that are included in the output of every measurement, as well as the Headwall software 'waterfall' live-view that we monitor during the scan, both show a green line. The line extends East-West in the image, more on the North side. By looking at the raw data, are you able to determine which wavebands are involved? Or provide any other feedback on the issue?

Here's an example of the jpg output, a zoomed in snapshot of part of the jpg output from Septmeber 3, approximately 3pm, with the band of sun on the south side, and the green line visible more on the north side. vnir_sep3_approx3pm

Paheding commented 6 years ago

@NewcombMaria We can definitely check this issue. Could you please provide a full directory that we can download this VNIR data?

NewcombMaria commented 6 years ago

@Paheding the stripe is visible on all the VNIR images, including 22-Aug exposure test images (targets deployed) and short scans on 31-Aug, 1-Sep, and 2-Sep (targets not deployed).

Paheding commented 6 years ago

@NewcombMaria Thanks, we will do some investigation on those dates data.

NewcombMaria commented 6 years ago

Copying email discussion from 20-Sept:

From: Newcomb, Maria S - (newcombm) Sent: Thursday, September 20, 2018 11:51 AM Subject: Re: 7100019 - VNIR Hi Vasit, All of the VNIR data collected since Sept 14, including Sept 14, has non-uniformity correction enabled. And all data since VNIR installation up to Sept 14 is without the correction. Thanks, Maria

From: Robert Strand robert.strand@lemnatec.com Sent: Thursday, September 20, 2018 12:41:44 PM Subject: 7100019 - VNIR Justin and Nadia,

Since we didn’t have a meeting, I wanted to give a quick recap of the VNIR situation.

After receiving the VNIR analysis from Vasit regarding what appeared to be a bad pixel, I contacted Headwall to get their take on the situation. Their immediate reaction was that it was related to their non-uniformity correction. This correction is enabled from within their Hyperspec III software. Upon inspection, the correction was not active. After activating the correction, the anomaly does not seem to appear in the output.

I think it would be helpful if Vasit could take a look at a corrected scan and give his approval.

We typically don’t touch many settings in Hyperspec III… mostly because there is no need. Additionally, there is little documentation.

I suspect that Headwall disabled the correction as part of their post-repair checkout of the system and simply didn’t re-enable it before shipment. Bob

Paheding commented 6 years ago

@NewcombMaria Thanks. This is great to know. We will do some evaluation on the corrected scan.

Paheding commented 6 years ago

@NewcombMaria We have checked the several images before and after making the correction. According to some analysis on the several sensitive bands to the stripping noise, we found that the stripping noise is not present on the images after the correction. Thus we may conclude that the noise issue is due to the unenabled non-uniformity correction.

An example of the comparison on images before and after making the correction is shown below. From these images, it can be observed that stripping noise is disappeared after non-uniformity correction (image 2018-09-15), especially it is more notable on band 805 (wavelength 914 nm).

image