A free video tutorial from Jose Portilla
Head of Data Science, Pierian Data Inc.
4.6 instructor rating • 32 courses • 2,263,380 students
Learn more from the full coursePython for Computer Vision with OpenCV and Deep Learning
Learn the latest techniques in computer vision with Python , OpenCV , and Deep Learning!
14:03:33 of on-demand video • Updated March 2021
- Understand basics of NumPy
- Manipulate and open Images with NumPy
- Use OpenCV to work with image files
- Use Python and OpenCV to draw shapes on images and videos
- Perform image manipulation with OpenCV, including smoothing, blurring, thresholding, and morphological operations.
- Create Color Histograms with OpenCV
- Open and Stream video with Python and OpenCV
- Detect Objects, including corner, edge, and grid detection techniques with OpenCV and Python
- Create Face Detection Software
- Segment Images with the Watershed Algorithm
- Track Objects in Video
- Use Python and Deep Learning to build image classifiers
- Work with Tensorflow, Keras, and Python to train on your own custom images.
English [Auto] Welcome back in this lecture we're actually going to go over the assessment as a quick overview and then the next lecture will go over the solutions. This is a really cool assessment. We're going to combine a lot of ideas to actually create a real working application that blurs license plates. We're going to do in this lecture just do a quick overview of the notebook. So let's open up the Assessment Project. All right. So here is the object detection Assessment Project and don't worry we're not going over the solutions yet. And this one you're going to detect Russian license plates and then blur them out of images. So the Russians are pretty famous for having some of the most entertaining dash cam footage on the Internet. So let's try helping them out by actually blurring some of their license plate. So there's three ways to approach this project. One is you can just go for it just understand the idea that it's your job to blur Russian license plates and not even try to reference that notebook to see if you can build it on your own. Given some of the SML files there's a car plate that jpg file that you can use in order to try to build that function on that or the second way which is probably the most recommended way is to use this notebook as a Project Guide. We offer a guide of the main tasks you need to complete in order to actually create this application. The third way is if you want to take a step back and really treat this as a coatl on a project that's totally cool too. All you need to do is jump to the next lecture. Rugova the solutions that we can code along and build an understanding as I code the solutions out. In general our Project Guide all you need to do is import the usual libraries then you're going to read in this car plate that jpg file and I want you to create a function that displays the image in a larger scale you're really familiar with that by now and then this is the image we're going to be working with. So there is a screenshot from dash cam on YouTube. And this is a Russian license plate. So the ex-MIL file it's pre-trained in that data folder is for Russian license plates. It's not for American ones. That's why we're dealing with this particular image. So then you'll load this SML file and then I want you to create a function that takes in an image and draws a rectangle around what it detects to be a license plate. So keep in mind we're not blurring yet. All this functions should be doing is what we did in the previous lecture or at least really similar to it. And then just drawing a rectangle around it. So you'll have the template function. And once it runs it should be able to draw a red rectangle around it like this. So then the final task is to edit that function so it effectively actually blurs that detected region. So this is when you combine a lot of the image processing ideas that we previously discussed in other sections of the course. So there's some hints here basically steps on what you should do to actually detect and then blur it. A lot of it. The trick is just understanding and slicing. So definitely take a hard look or hard read at some of the hints I left for here. But at the end of the day we should be able to do is after running an image through the text and boilerplate it will automatically detect license plates and then blur them. So in this image it detects this license plate and I can read it it says something like 0 3 3 P-A. If I run it through my updated text and boilerplate it should eventually just be blurred to the point where I can no longer read it and notice it's just that section that's blurred the entire image is not blurred. This was on a rainy day. So it's a little blurry but definitely it's just a license plate that has been blurred with whatever blurry mechanism you're on. Probably a median blur is the easiest. OK. So best of luck on this and we'll see you at the next lecture where we go over the solutions.