Adaptive box filter is not working properly!

[yohochang11]

I try to use options of adaptive box filter to replace invalid pixels with loacl average but it is not working properly. The invalid pixels remain the same.
My setting is as below,
１）replace invalid pixels with loacl average
２）Threshold standard deviation = 1.5 Threshold difference = 0
3) Change min/max values 0.01 and 1 (the min and max values of the satellite image are 0 and 0.42)
4) Filter size is 7 by 7
The filter result is indetical to the input file, the invalid pixels of Landsat 7 SLC off image are about 2-3 pixels wide.

[tawoodard]

What are the values of the Invalid Pixels? Are they just numbers that are beyond 1.5 standard deviations difference from those around them? Or are they NaN values or some other type of invalid values?
I will probably need to look at your input data file, as my test on a random image with your parameters did produce some cells that changed. By the way, the option to Change minimum/maximum values did not change anything for me, other than the metadata. It doesn't scale the data or anything. I'm not sure what it would be needed for.
If your invalid values are NaN values, then NANFIX is what you would use.
If that is not the problem, then open a tech support ticket, so I can send you a DropBox link to send me your input file.
https://forums.clarklabs.org/hc/en-us/requests/new

[yohochang11]

I have uploaded the files to your link. The invalid pixel value is 0. It was a remote sensing course lab exercise for the last 10 plus years to spatial filter Landsat 7 SLC off images (images acquired after Earthexplorer website). It was OK if we use the version TerrSet 18.31 and before to adaptive filter those images. This year we try to process the data using TerrSet liberal GIS and we found that adaptive box filter was not working as we expected but TerrSet 18.31 could finish the job if we use adaptive filter box filter twice. Cheers! Yoho 東華大學環境學院自然資源與環境學系(所) 張有和副教授 97401 花蓮縣壽豐鄉志學村大學路二段一號 TEL: 03-8905183 FAX: 03-8903260 email:  ***@***.*** Dr. Yo-Ho Chang Department of Natural Resources and Environmental Studies College of Environmental Studies National Dong-Hwa University No. 1, Sec. 2, Da Hsueh Rd., Shoufeng, Hualien 97401, Taiwan, R.O.C. TEL: +886-38905183 FAX: +886-8903260 email:  ***@***.*** ***@***.***&> Tammy Woodard ***@***.***>於 2025年12月10日 週三，上午12:36寫道：

…

*tawoodard* left a comment (ClarkCGA/TerrSet#18) <#18 (comment)> What are the values of the Invalid Pixels? Are they just numbers that are beyond 1.5 standard deviations difference from those around them? Or are they NaN values or some other type of invalid values? I will probably need to look at your input data file, as my test on a random image with your parameters did produce some cells that changed. By the way, the option to Change minimum/maximum values did not change anything for me, other than the metadata. It doesn't scale the data or anything. I'm not sure what it would be needed for. If your invalid values are NaN values, then NANFIX is what you would use. If that is not the problem, then open a tech support ticket, so I can send you a DropBox link to send me your input file. https://forums.clarklabs.org/hc/en-us/requests/new — Reply to this email directly, view it on GitHub <#18 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKHWPYHCBDOFJSN76PXZXIL4A33BHAVCNFSM6AAAAACOOEGUX6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMMZTGE4DONBVGQ> . You are receiving this because you authored the thread.Message ID: ***@***.***>