图像处理技术在现实中的运用 - 好文

电子科技大学格拉斯哥学院 2017贺宸，同组同学无

前言：这是在听完学院有关图像处理之后的一些感想，将技术用在现实当中。

关键词：图像处理，交通

主体：
在校学习期间，学院为了让新生接触更多领域，了解前沿科技，设立新生研讨课。其中，有一节课专门介绍了图像处理。
图像处理(image
processing)，是用计算机对图像进行分析，以达到所需结果的技术。又称影像处理。图像处理一般指数字图像处理。数字图像是指用工业相机、摄像机、扫描仪等设备经过拍摄得到的一个大的二维数组，该数组的元素称为像素，其值称为灰度值。图像处理技术一般包括图像压缩，增强和复原，匹配、描述和识别3个部分。

在讲座上，老师展示了几个图像处理技术在实际生活中的运用。第一个展示是一张很高像素的照片，再利用图像处理技术之后，一些关键的像素被选择，这部分被选择的像素重新组合，变成了一张缩略图——就是一张缩小版的原图。这是将图像处理技术用于生成预览图，在浏览一些大图的时候，用户可以根据预览图，选择是否查看原图。这种应用可以为用户节省流量。还有一个运用，是在小车，无人飞行器上装载摄像头，利用图像处理技术处理摄像头拍摄的画面。就可以得到稳定的图像。若有两个摄像头，在处理之后，两张图像可以较好的拼接在一起，晃动的图像可以通过拼接，使得拍摄物体的主体部分稳定不晃动且无明显拼接痕迹。在演示视频中，由无人飞行器拍摄的主楼，在屏幕上稳定的显示了出来。在拍摄体积较大的物体时，一个摄像设备无法完整的记录下影像，可以用几台小摄像设备，利用图像处理的技术，得到完整的像。

我觉得，这种技术可以运用于道路监控。我父亲是一位警察，就我故乡的情况来看，目前城市中的监控大都是在交通要道上设立摄像头，对往来车辆监控。监控系统发展至今，其拍摄画面像素非常高，图像里大大小小的车牌，经过放大后，上面的数字和字母可以看的非常清楚。这种系统固然优秀，但是，在有些偏远地区，或者是在高速上，很难做到在很长的一段道路上设立监控摄像头，高成本低收益。可是，当高速公路上出现拥挤情况时，警察也非常需要实时的画面，以便突发事件发生后采取应对措施。在这种情况下，无人飞行器和图像处理技术可以起到至关重要的作用。高速公路上，拥挤地段的位置和地点都是不确定，随时都会发生变化的。在警署指挥中心的人员，可以利用无人机，拍摄现场的实时画面。经过图像处理技术，将拍摄到高分辨率的图像简单加工，变成较为模糊的图像，通过网络发送至指挥中心的系统里。经过加工的低分辨率视频不会使用太多流量，当需要看到高清画面，如查看车牌，分析纠纷，界定责任的时候，可以查看原录像。另外，如需要一整段公路的交通情况，可以将几台摄像机拍摄的图像进行处理，合并，得到一个整体的路段情况。
During the school study period, the college set up a new seminar for the new
students to contact more fields, understand cutting-edge technology. Among
them, there is a class devoted to image processing. Image processing is a
technique of analyzing an image with a computer to achieve a desired result.
Also known as image processing. Image processing generally refers to digital
image processing. A digital image is a large two-dimensional array taken with
an industrial camera, camera, scanner, etc. The elements of the array are
called pixels and their values are called gray values. Image processing
techniques generally include image compression, enhancement and restoration,
matching, description and recognition of three parts. In the lecture, the
teacher showed the application of several image processing techniques in real
life. The first display is a very high-resolution photo. After using image
processing technology, some key pixels are selected. This part of the selected
pixels is recombined into a thumbnail - a reduced version. Original picture.
This is to use image processing technology to generate preview images. When
browsing some large images, users can choose whether to view the original image
according to the preview image. This kind of application can save traffic for
users. Another application is to load a camera on a car or an unmanned aerial
vehicle, and use image processing technology to process the picture taken by
the camera. A stable image can be obtained. If there are two cameras, after
processing, the two images can be better spliced together, and the swaying
images can be spliced so that the main part of the object is stable without
shaking and no obvious splicing marks. In the demo video, the main building
taken by the unmanned aerial vehicle is displayed on the screen stably. When
shooting a large object, a camera device cannot record the image completely.
You can use several small camera devices to obtain a complete image by using
image processing technology. I think this technology can be applied to road
monitoring. My father is a policeman. As far as the situation in my hometown is
concerned, most of the current monitoring in the city is to set up a camera on
the main road to monitor the traffic. Since the development of the monitoring
system, the pixels of the shooting screen are very high. The large and small
license plates in the image can be seen clearly after the enlarged numbers and
letters. This kind of system is excellent, but in some remote areas, or at high
speeds, it is difficult to set up surveillance cameras on a long road, high
cost and low profit. However, when congestion occurs on the highway, the police
also need a real-time picture to take countermeasures after an emergency. In
this case, unmanned aerial vehicles and image processing techniques can play a
crucial role. On the highway, the location and location of the crowded areas
are uncertain and will change at any time. The personnel at the police command
center can use the drone to take a live view of the scene. Through the image
processing technology, the high-resolution image is simply processed into a
more blurred image and sent to the command center system via the network.
Processed low-resolution video does not use too much traffic. When you need to
see high-definition images, such as viewing license plates, analyzing disputes,
and defining responsibilities, you can view the original video. In addition, if
you need a whole road traffic situation, you can process and combine the images
taken by several cameras to get an overall road segment.

结语：
这是个人的一些畅想，未经过学术性的调研。

热门工具换一换