BBOX is outside of WGS84 limits

[ihor-hrechka]

Hi!

I am trying to load BIL data from geoserver, but templateToURL(urlParam, x, y, level, provider) function attaches wrong bbox values to URL (see screenshot bellow). What can be the reason?

P.s. Cesium version 1.33 P.p.s. Let me know if you need more details!

Thank you in advance!

[kaktus40]

Hello, could you send your WMS capabilities document?

[ihor-hrechka]

Hello, thanks for the answer!

I can't send full xml, but I made two screenshots:

Geoserver GetMap capabilities:

And description of layer, which I'm trying to load:

Is it enough? Or you need more info?

[kaktus40]

It seems that all is correct. I'm sorry but I take vacation and I can't help you until 19/08/2017...

[ihor-hrechka]

Ok, I will wait! If new observations will come, I will let you know!

Have a good vacation!

[ihor-hrechka]

Hi! I hope you are back and your vacation was perfect! :)

I am still having an issue, described above.

Here my source code:

// create viewer
viewer = new Cesium.Viewer("map", {
        animation: false,
        timeline: false,
        homeButton: false,
        baseLayerPicker: false,
        navigationHelpButton: false,
        geocoder: false,
        infoBox: false,
        sceneModePicker: false,
        fullscreenButton: false,
        imageryProvider: new Cesium.createOpenStreetMapImageryProvider({
            url: "https://a.tile.openstreetmap.org/"
        }),
        contextOptions: {
            webgl: { preserveDrawingBuffer: true }
        }
});

// create terrain provider
var geoserverTerrainProvider = new Cesium.GeoserverTerrainProvider({
    url: geoServerUrl,
    layerName: "sf:sfdem"
});

// attach terrain provider
viewer.scene.globe.terrainProvider = geoserverTerrainProvider;

Request looks like this (bbox looks strange):

And response is ServiceExceptionReport:

<ServiceException>
      Error rendering coverage on the fast path  java.lang.NullPointerException
</ServiceException>

Do you have any idea what can be the reason?

Thanks!

[kaktus40]

It's very strange, from what I understand, your layer sf:dem has the next bounding box in CRS:84:

• south or minx : -103.871006...
• west or miny: 44.3702...
• north or maxx: -103.6293...
• east or maxy: 44.5016...

For record, the bbox for CRS:84 in WMS request is define in this order: '{south},{west},{north},{east}' so the biggest rectangle that can be requested should be bbox=-103.871006, 44.3702, -103.6293,44.5016 and you have a request with bbox= -181,-91,1.40,91

Could you add some line codes in the code of plugin? After line 368 (retour = Cesium.defined(scratchRectangle);), could you add the next lines and show me the result:

                            if(retour){
                                console.log('reference: west'+rectReference.west+'; south:'+rectReference.south+'; east:'+rectReference.east+'; north:'+rectReference.north);
                                console.log('calcul: west'+rectangleCalcul.west+'; south:'+rectangleCalcul.south+'; east:'+rectangleCalcul.east+'; north:'+rectangleCalcul.north);
                                console.log('result: west'+scratchRectangle.west+'; south:'+scratchRectangle.south+'; east:'+scratchRectangle.east+'; north:'+scratchRectangle.north);
                            }

[ihor-hrechka]

I added this lines here:

Result from console:

[ihor-hrechka]

Sorry, I closed issue by mistake.

[ihor-hrechka]

I also made one screenshot of templateToURL(urlParam, x, y, level, provider) values. Maybe it can be helpfull:

[kaktus40]

Well from what I see, all work well as I defined. I should add a max limit to rectangle according to the tilingScheme. This limit should be adaptable also.

[kaktus40]

Hello which is the version of geoserver. It seems there some issues with old versions Le 29 août 2017, à 08:48, IgorGrechka notifications@github.com<mailto:notifications@github.com> a écrit:

I also made one screenshot of templateToURL(urlParam, x, y, level, provider) values. Maybe it can be helpfull: [capture2]https://user-images.githubusercontent.com/1829604/29808021-2745ba5a-8c9f-11e7-9e17-a96033329730.PNG

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/kaktus40/Cesium-GeoserverTerrainProvider/issues/22#issuecomment-325571123, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AF0OfPeI77pCMVRjZ8GDk2y1tUaJSu4Jks5sc7Q7gaJpZM4OYKHJ.

[ihor-hrechka]

Hi! Please look at GeoServer's about page:

[kaktus40]

Is it possible for you to update the version of geoserver?From what I saw on internet, may be one piece of your dataset is bad... Le 30 août 2017, à 14:07, IgorGrechka notifications@github.com<mailto:notifications@github.com> a écrit:

Hi! Please look at GeoServer's about page: [capture]https://user-images.githubusercontent.com/1829604/29871551-d46115ae-8d94-11e7-9665-ae0ae45048c4.PNG

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/kaktus40/Cesium-GeoserverTerrainProvider/issues/22#issuecomment-325970473, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AF0OfAhiSVlq_isK5zCWLBUJ6tsDw5Frks5sdVBwgaJpZM4OYKHJ.

[ihor-hrechka]

Ok. I will check the possibility of updating the GeoServer to the latest stable version and let you know.

P.s. Maybe you know, where I can get correct dataset for testing?

[kaktus40]

Well in the end of readme.mdhttp://readme.md I give some links...

Envoyé par BlueMailhttp://www.bluemail.me/r?b=10455 Le 30 août 2017, à 15:36, IgorGrechka notifications@github.com<mailto:notifications@github.com> a écrit:

Ok. I will check the possibility of updating the GeoServer to the latest stable version and let you know.

P.s. Maybe you know, where I can get correct dataset for testing?

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/kaktus40/Cesium-GeoserverTerrainProvider/issues/22#issuecomment-325992192, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AF0OfAJeWfiGiHWmvsoNgrhOvkfGwuDuks5sdWVKgaJpZM4OYKHJ.

[ihor-hrechka]

Hi!

I updated GeoServer to version 2.11.2

Now the globe is loading (the half of the globe)

BBOX in first request is outside limits: but after starting rotating\zooming the globe it becomes correct:

But previously it was only one request after "GetCapabilities" and globe wasn't load. So I can say that update of GeoServer have resolved the issue.

Bit terrain looks like bellow:

Layer preview in GeoServer looks like this:

Can you tell me what is the reason of wrong terrain visualization and why Cesium loads exactly half of the globe?

Thank you!

[kaktus40]

hello, for the wrong terrain visualization, you should check https://github.com/kaktus40/Cesium-GeoserverTerrainProvider/issues/9#issuecomment-119314192 for example. By default the plugin works with data in Integer 16 bit format. Your data is on a float format (32 or 64 bit?). In this comment, you should find a solution.

For the half earth not rendered, it's also normal because cesium use function to determine if a tile is available. If a tile is not available, Cesium doesn't download the image corresponding. A workaround (for WMS) is to change this function so it always returns true!

[ihor-hrechka]

Hi! I checked the original image format and you are wright, it is in Float32 format. GDALinfo of image:

After I tryed to convert it using gdalwarp -ot Int16 -co "TILED=YES" -co "COMPRESS=DEFLATE" -r cubicspline "C:\sfdem_orig.tif" "C:\sfdem_conv.tif"

GDALinfo of converted image:

But the result is same:

If it required, I can upload original and converted images somewhere!

[kaktus40]

It seems you have a problem with the endiannes of your data (easily corrected in bil/DDS parameters in geoserver). You didn't have to convert your data in int 16 bit data. If you follow my comment, you just need to check the parameter in geoserver and implement postProcessArray when you use the plugin!

[ihor-hrechka]

Thanks a lot for your help!

I copied your formatArray implementation and changed Default Encoding and Byte Order in GeoServer layer settings. Now the picture looks fine:

As I understand - this "stair" is the end of elevation map in GeoServer. Am I wright?

[kaktus40]

Yes that right!