The Key Technical Research Of The Pastures Dynamic Monitoring Based On Computer Vision Measuring

Nowadays, scientific development and sustainable development strategies serves to be one of indispensable economic development policies of our country. The communism party's congress proposed to develop a kind of economy which is friendly and natural coexistence of man and nature------protecting ecological environment and changing farmland to forest and pasture. China is country which is abundant in grassland. Because of this, it would make enormous contribution to advantage of resources and potential of the development of grassland. As we didn't follow the rules required to make the development of grassland sustainable, it becomes more increasingly sandy and salty with the time, which largely restrains the animal husbandry. So the digital management of the grassland turns out to be a pressing problem. This paper is about key technology research of resource-saving digital management of Northeast livestock industry, that based on the National High Technology Research and Development Program (863 Program) project.It has long been know that only two key factor play significant role in the grassland dynamic surveillance----one is the height of plant, the other is wind speed. We utilized main principle of digital image processing and the computer vision to provide the parameters for digital management of grassland.With the development of computer hardware and software, using computer vision measurement method to measure certain elements in the scene has become increasingly important in computer vision topics.Computer vision measurement attracts far more attention and concern of scholars more. By using computer vision to reflect the key factor in the dynamic measurement of pasture:plant height measurement and wind speed. We made a new attempt.Primary methods commonly used for measuring the height can be put into several classifications:one is to obtain the three-dimensional data of object by building the three-dimensional construction through one or two images. The other is to measure the single image.But we always assume the prospect being measured is put vertically, if is not,and in oblique case, the accuracy would be greatly affected.In this paper, we made improvement for the single image measurement. The board papers are put in the scene, and then the coordinates can be acquired. The least square method has been used to calculate the plane homographs between the planes. Further more, calculate the projection matrix between the image plane and measured image plane coordinate system in the light of characteristics of homograph matrix and the special relationship of projection matrix. After that, obtain the high and low points of the to calculate the absolute low point truelow. Finally, get the final calculation of the absolute high Z-value,lownod truehigh and truelow time coordinates corresponding to the distance between the values. By using the Pythagorean Theorem, the more accurate value of height of foreground can be estimated This method solves the highly inclined foreground measurement problems and improve accuracy of the measurement. for the prospects of object height estimation.Traditional way among various kinds of other methods of measuring the wind speed mostly bases on the senor:thermal anemometer, three cups of point vortex sensor and the sensor cups optocoupler sensor, which have the advantage of real time and higher accuracy. Nevertheless, we would be able to test the wind speed by using computer vision to save the cost and expense, as a enormous number of existing cameras have been set up through the whole grassland. In this paper, we attempt to measure the speed of wind in a novel way.Firstly, the camera returns the continuous video in the regard to identification of wind wheel in the scene. Then computer reads the each frame in the video, binaries each frame, obtains the outline by gradientizing and acquires the maximum contour lines to fit the elliptical profile together with the center of the elliptic, length of shaft and width.Secondly, correct the tilt image to turn back to what they really are when it has been vertically shot in the camera. Just then, the oval-shaped image turns out to be a circle.What is more,there are some blur logo regions on the black identification of wind wheel result from the fact that a wind wheel is rotating all the time on the pasture, so we have to restore the image of motion blur to obtain a clearer black logo regions.Last but not least, we need to get the black identification area in the oval of image, calculating the coordinate of the center (nextX,nextY) which can be calculated and results in the angle rotated between the two frames. Because of this,we can get angular velocity,furthermore the wind speed.The method we proposed could be applied to test the wind speed in the good weather condition, such as pasture. The advantage of this method lies within using digital video processing technology which would be more easily and flexibly to be put into practice and save the cost. At mean time, this method serves to be used for any camera or any camcorder,any resolution of image.