Search the Community
Showing results for tags 'supervised classification'.
halo teman-teman gisarea indonesia kebetulan saya sedang mengolah data ALOS PRISM untuk tutupan lahan kota semarang. disini saya menggunakan software Erdas Imagine 2014, saya agak menemui kesulitan dalam prosesnya & semoga diantara teman2 disini bisa membantu memberikan masukan. "apakah tiap kita membuat signature untuk satu nama kelas harus dengan satu area of interest (AOI)? apakah bisa membuat satu kelas dengan banyak AOI? karena dari pengolahan data saya, tiap satu nama kelas harus satu AOI dan hasil supervisednya berantakan." terimakasih
G’day ‘GIS Area’ I’ve got a bit of a tricky conceptual problem and was looking to see if anyone had any thoughts. I have a vast area of low lying saltbush and grassy pastoral land (approximately 80km2) in which have been randomly placed several thousand lengths of white PVC piping protruding approximately a foot or so out of the ground more or less vertically (with no caps) above the majority of the grass and stubble. I’m attempting to use drones to scan the surface and get precise coordinates of every one of these lengths of pipe; however I haven’t been able to land on a method of doing this as yet. I’ve considered scanning the area with hyperspectral sensors and then performing a supervised classification of the data using the reflectance signature of the white PVC piping however the thin topographical cross-section of the pipes means too little of the PVC would be observable from a drone scanning the ground from directly above. The only thing I can think of to resolve this issue would be to tilt the sensor so that the resultant images capture the sides of the PVC rather than the tops and then adjusting the coordinates somehow. If anyone has any thoughts on this, or any other potential solution to my piping problem, I’d greatly appreciate the input. P.S. I'm quite a novice at remote sensing and airborne research so as jargon free a response as is practical would be appreciated also... Regards MattRNR
Salam Mastah dan sepuh semua..mohon bimbingannya. sy butuh tutorial detail klasifikasi terbimbing menggunakan ermapper. sdh browsing kemana2, setelah sy gunakan hasilnya tidak memuaskan.terima kasih
I have a question if anyone can answer, it'd be greatly appreciated! I have Erdas Imagine 10, Arcmap 10 and ENVI 5. I am interested in limiting classes to specific areas of the image during a supervised classification. For example, if I know that wetlands only occur in the south-west area of my image, is there a way to limit that class to only measure pixels in the south-west portion of my image? This way I am ensuring the wetlands class only appears where I know wetlands actually exist on the ground and the class can't appear anywhere else - only in the area of the image that I delineate, and so prevents class confusion in other parts of the map. I there a way to do this, is it possible with the software mentioned above?