Zhuangfang Yi 依庄防 comments

Results 37 comments of


                                            Zhuangfang Yi 依庄防

trafficstars

Sagemaker 'NoneType object' issue with data in 'walkthrough-classification-mxnet-sagemaker' example

Phewwww, I finally solved the problem and took me a whole morning today, @joshwapiano, and @mapmeld. You're right about S3 bucket and `prep_data(find_file( ... ))`, @mapmeld. I deleted `find_file( ......

OSError: cannot identify image file <_io.BytesIO object at 0x7fc87f2ec678>

@checongcong, I have python 3.6.3 installed in my python environment and used tensorflow 1.4.1 originally, but I also tried tensorflow 1.5.0, and both worked. Can you do two things over...

error happend when use "pip install label_maker"

@yjykitty, you might want to set up a [python development environment](https://gist.github.com/wronk/a902185f5f8ed018263d828e1027009b) instead of using anaconda for this particular task. Then you can install the latest version resterio and it won't...

error happend when use "pip install label_maker"

oh, i see what you meant @yjykitty. This error might come from one of Rasterio's dependencies, python-rasterstats, which I've seen before. Can you do `pip install --pre "rasterio>=1.0a12"` instead, it...

error happend when use "pip install label_maker"

@1126876778 and @yjykitty, have you solve the problem here? I will close up this issue if you have.

Tensorflow Inference over satellite image tiles

You're right here, @Geo-Trackers. However, a large object may cross two tiles or more tiles so can write a utility function to get rid of bounding boxes that overlap by...

Running inference over .tif in tf_od_predict.py

@Geo-Trackers, yes, you can run the inference over the image tile in tif format, and the [main script is this line](https://github.com/developmentseed/label-maker/blob/7b8953c6f7d6dce24fa3eed4a1b6179b60b7db4f/examples/utils/tf_od_predict.py#L93). You can replace `.jpg` with `.tif`.

Running inference over .tif in tf_od_predict.py

Yes, @Geo-Trackers. But it's not very straightforward. When you use Label Maker to generate `.tif` tile, each tile come with an ID, e.g. `207692-115930-18.tif`, you can use [Mercantile](https://github.com/mapbox/mercantile) e.g., `mercantile.bounds`...

Create example with non-API restricted imagery

Good idea, @drewbo, it could worth to look into [DG's Open Data Program](https://www.digitalglobe.com/opendata/) too.

Counting Total Number of Objects detected using tf_od_predict.py?

yes, @Geo-Trackers, go to these few lines from [here](https://github.com/developmentseed/label-maker/blob/a057f5d2759d7f21646f55a51bd9f4c332d399e3/examples/utils/tf_od_predict.py#L61). You can write a simple helper function to loop through and write down each bbox, the associated class, and the tile...