Talking about Digital Image Processing
Image processing is a very popular field in recent years, and related research and topics are much more. The recent patrol car race held by the school has also made me pay attention to this interesting field.
1.What is image processing?
Digital image processing, to put it bluntly, is a process of abstracting things. For example, we want to hold a photo contest. It seems difficult to define whether a photo looks good or not. Can we come up with a clear standard? Image processing can help us solve this problem. The image is no longer an image, but a container of information. It could be a lot of points, a matrix, or something else. The smart computer will help us analyze and return a digital message that can be processed, which is exactly what we want. What needs to be said here is that images in a broad sense are not necessarily visually meaningful. In fact, this is a very complicated process, but it can greatly facilitate our research.
2.Image processing application
1.Face recognition
First of all, the closest things to us are the mobile phone and the computer. Laziness promotes social progress. This is the truth. For example, for me personally, after the introduction of the face lock, it seems very complex to type the code in order to unlock my laptop; then, the access control system, fingerprint and iris recognition were very popular in the past, but these two methods are invasive in the process of collecting. By contrast, face recognition is much more convenient. Finally, the monitoring of criminals. all the city are monitored, which is an important means of tracing criminals. But letting the police officers check one by one is undoubtedly a vast project. If the system can automatically identify it, it will be very convenient.
2. Deep learning
“Image processing is a paradise for deep learning applications.” Deep learning can be said to be the most widely used image processing application, but I don’t know much about deep learning, so if you want to know more,please google by yourself.
3.Relevant software
1.LabVIEW+Matlab. LabVIEW’s data collection function is unusually strong, and it has a 60% market share in automation testing, but it doesn’t widely spread in China. It supports most cameras. Coupled with the NI Vision Vision Development Module, it is easy for it to implement many features. It can be easily developed in conjunction with the mathematical calculation tool Matlab, which is extremely powerful. But the disadvantage is that the library functions are not rich. MathWork’s Matlab software can be said to be an artifact of the algorithm, simple. Unparalleled computing power and simulation capabilities seem irreplaceable. Most of the world’s algorithm famous people will publish the corresponding Matlab source code when publishing the results, and then gradually rewrite it into other languages for practical application. Therefore, Matlab is very important in image processing learning.
2.VC++. This is the most widely used industry in the industry, and most camera manufacturers will also give VC development kits. And there are a lot of open source libraries, such as OpenGL, OpenCV, etc., which makes it very powerful. However, the VC series is slow to get started, the programming is complicated, and the bugs that cover full screen will make people feel dizzy.
So to sum up, Matlab is the best choice if you want to say simple image processing, or are still in the introductory stage.
As for the linear filtering, Gaussian fuzzy and other algorithms introduced, I don’t know much about it, I don’t want to swear here. The image processing‘s threshold is very low, but behind it is the ocean. Only by constantly exploring can we find out more.