I’m currently working on a pet project I call Night Camera iOS App. It aims at making it possible to shoot photos of appropriate quality in low light conditions. My use case is taking a photo of my sleeping kids. They are so cute when asleep but don’t like it when I use a flashlight on my phone.
Right now I managed to implement a great Intro to Python Image Processing in Computational Photography. It’s about enhancing a series of still dark images in Python using OpenCV. What I did is put that idea into an iOS app, getting the images in realtime using the videocamera. The implementation is in C++, hosted as a repo on Gitlab.
I’ll write a detailed post about the first version in a separate post. Now to results. They aren’t quite exciting now. To put it short:
![](https://about.stan.sidel.family/blog/project-log-nc-2019-09-04/images/brightened-single-shot_hu1e4e06f260eeeea98c6b20cbbd8ebeab_188298_0x600_resize_q75_box.jpg)
![](https://about.stan.sidel.family/blog/project-log-nc-2019-09-04/images/enhanced-image_huae0dbbc209cc52f60fa5028c17880d0a_85108_0x600_resize_q75_box.jpg)
![](https://about.stan.sidel.family/blog/project-log-nc-2019-09-04/images/stock-camera-result_hu3c4f7c90adf7b5437c3f397e701b86f6_43383_0x600_resize_q75_box.jpg)
The difference is largely due to the fact I used medium quality of the images from the videocamera. High-resolution images are very large and are processed quite slowly.
Next I’d like to try:
- Add processing properties controlls to experiment faster.
- Try to get higher image resolutions