Said Google’s artificial intelligence tall then talk about how to help the farmers sorting cucumber & Ji Zhenyu Tencent since Google (micro-blog) for machine learning Tensorflow tools at the end of last year announced the open source since the worldwide developers using Tensorflow application development on artificial intelligence. Although the concept of artificial intelligence sounds very tall, but in fact the use of artificial intelligence programs can help solve some of the most common things in life. For example, in Japan, a car industry engineer named Makoto Koike, the use of deep learning methods for the planting of cucumber parents set up a system of automatic sorting cucumber. Makoto’s parents are engaged in the cucumber cultivation industry, has its own large scale Cucumber Planting farms. But in the enjoyment of the joy of cucumber harvest at the same time, a pile of trouble also. Due to the length, thickness, shape, color, texture of the cucumber are different, the need for artificial classification of cucumber quality, according to the level of quality, the price is also different. Makoto and its parents in Cucumber park this is a tedious, time-consuming and laborious work, fine, in the cucumber harvest season, the average mother Makoto daily sorting work to spend more than 8 hours of cucumber. There are many years of experience in the automotive industry, see the car production line on a variety of automated processes Makoto thus thought, why not set up a cucumber automatic sorting system to save the trouble of artificial. But the sorting of cucumber is not an easy job, even from the point of view of the person, to skilled this technology, but also need to spend a long training period, not to mention the machine. In Japan, there is no uniform industry standard of vegetables category, but by each farm decide the crop produced quality standard, Makoto will be the quality of their parents’ farm output of cucumber is divided into 9 grades, Makoto said, his mother also recently to learn how to accurate sorting technology of cucumber. Considering the complexity and accuracy of the sorting of Cucumber in accordance with the requirements of the length, shape, texture, color and other elements are divided into 9 grades of cucumber, Makoto decided to use deep learning to build this set of automatic system, he said that his inspiration from Google for a period of time before the fame AlphaGo Makoto, that is AlphaGo’s success inspired him learning technology to carry on the automatic sorting system is built up by cucumber depth. Specifically, the Makoto system uses Google’s open source system Tensorflow, through image recognition, and combined with the hardware controller, the final realization of the sorting process. The automatic sorting system of cucumber deep learning system processes the process from image recognition, based on the training data set to learn computer, to know what is the most important image in the "elements", and then according to the order of importance, so as to realize the classification. After the completion of the cucumber automatic sorting system, but this method is still facing many challenges in the actual application. First, the system requires a sufficient amount of training data, in order to train this!相关的主题文章:

Comments are closed.
a b

Recent Posts