Since the emergence of the big data concept, more and more industries have already had their own big data applications, but there are not many scenarios that can be applied, let alone deep machine learning.
Since 2011, after BigData was hyped up, more and more industries have their own big data applications, including not only data collection, but also data storage, and even data governance, but no matter what they do The degree of data management deviates from the big data application scenario, which is low-value or even worthless, not to mention the AI-machine learning-deep learning based on big data.
Next, let’s talk about the scenarios of agricultural big data. I will start with the two lines of agricultural segmentation and audience. The agricultural sub-fields are also the development of breeding (pretend to include animal husbandry), planting and forestry, and the three sub-fields have a certain overlap. The audience (stakeholders) starts from the government, producers, traders, fresh food e-commerce (including community group purchases). As for the aforementioned food distribution companies, agent sellers, agent buying, wholesalers, etc., they are all merged in the traders. It is here, and agricultural scientific research institutions are not involved this time. Plus a big data on agricultural product prices, this concept has been mentioned in recent years.
Unify the cognition: agricultural big data not only refers to the big data visualization large screens of various governments (it is said to have emergency command and decision support functions), but also includes data collection, storage and management. And big data does not necessarily have to be deployed on a private cloud/public cloud. It is also possible to deploy it on a physical server, so don’t be tricked into buying cloud services.
And agricultural big data not only includes agriculture, but the three rural areas are not divided into families. In the common large-scale visualization screen, there may also be some rural and peasant data. Don’t look outside!
Breeding industry (including animal husbandry)
Breeding includes animal husbandry, and I will not talk about animal husbandry alone.
The agricultural industry is smart agriculture with a higher degree of intelligence, or it can be said to be a higher degree of digitization. As for the concept of smart agriculture and digital agriculture, The agricultural industry can not only use big data and the Internet of Things, but also use low-level AI, which is still a little far away from machine learning and deep learning.
Producers in the agricultural industry
At present, the agriculture industry is the most suitable sub-industry that uses the Internet of Things. Whether it is pig raising, cattle raising, chicken raising, and agriculture, there is the Internet of Things to provide a bunch of IoT data, combined with some production planning, feeding execution and other management The data builds a data collection system, and can conduct big data analysis on the breeding process and make response measures. If insufficient oxygen in the fish pond is identified, guidance measures will be initiated. The video recognizes that the cow is not moving, the temperature is high, and the medicine is applied. Chickens are fed water quantitatively to reduce infection, etc. There are still many application scenarios.These scenarios can be done without using big data, but they are better.
Government on the breeding industry
There are also smart animal husbandry big data platforms in various provinces, cities and counties. The basic agricultural government big data is the management of various agricultural resources, and it can also be displayed using agricultural resource geographic information. For agricultural + animal husbandry, only general agricultural resources are pasture areas. The location, scale, contact phone number of a breeding farm in a certain area, and the location of an epidemic prevention station in a certain area show the current grassland area. Of course, when there is not enough grass in the pastoral area, it can also have a commanding role.
In case of emergency, it can also be used for command during natural disasters, such as the desertification of grassland for animal husbandry. Some scenes can be combined with drones.Drones and shaking can also become data sources for the aquaculture industry.
The planting industry has also said before that big data on vegetable planting is still a pseudo-scene. Only the identification of pests and diseases combined with AI is still a valuable scenario.
Other planting scenes also have value, such as precious herbs and flowers.
There is also a thermal imaging for viewing pests and growth forecasts (remote sensing can also do several categories).
For producers in the planting industry
For producers, they can identify current problems through images of pests and diseases, and apply fertilizers or pesticides in a targeted manner. However, not all categories have models now. The more accurate algorithms are tomatoes, grapes, lettuce, etc., such as broccoli, peas and other models that are not too mature. Of course, these are only suitable for novices in agriculture or to try new varieties, and veterans can understand by themselves, and are more accurate than models and experts.
Government on plantation
For the government, the planting industry currently uses a lot of scenes, and there are also large-scale greenhouses with fixed subsidies.
The agricultural emergency management caused by the identification of diseases and insect pests in the county, but basically the system is built there, I have not seen it, because it is really large, such as locusts, which generally cannot be solved immediately.
In the planting industry, the government’s agricultural big data applications include land surveying, land calculation, arable land red line and land circulation, etc. You can check the details yourself.
In fact, forestry is currently used in more scenarios at the government level, and producers are generally regulated by state-owned enterprises.
The currently visible application scenario of forestry big data is the display of forestry resources. Forest fire prevention, anti-theft and anti-poaching scenarios have a lot of value. It can also integrate security equipment (such as Hikvision), Internet of Things equipment and the actual work of the Forest Public Security Bureau. , Combined with big data, of course, it can be done without big data, but it is a little troublesome to process IoT data.
In addition, there are pre-warning, in-process handling of forest fires and geological disasters, post-event evaluation and knowledge accumulation. The essence of big data is data accumulation and algorithm evolution, that is, knowledge graph, which can reason about knowledge and generate knowledge by itself.
Fresh food e-commerce
Fresh food e-commerce’s application of big data is basically the old e-commerce routine: real-time data analysis, such as sales unit price, amount, total amount, etc.; smart recommendation for precision marketing (others like to call a thousand people in front), recommendation Give users the fresh food that users like; big data is familiar; cost estimation, use big data to estimate costs, guide purchase and pricing, and marketing; demand forecast, guide the purchase and processing process.
Of course, the above-mentioned large companies for fresh food e-commerce, too small regional, small-scale fresh food group purchase manufacturers, should not waste resources on this.
Finally, in fact, many big data applications for fresh food e-commerce are also needed by traders, but agricultural products traders basically do not have IT capabilities, so many SaaS service providers are needed to provide big data application services for traders Up.
Price big data
Among all agricultural big data, the first application is agricultural resource big data and price big data, which are all used for government supervision and used by other small manufacturers to make their own scenarios.
But for price big data, there is a fatal flaw that is the magnitude and representativeness of the price quantity. At present, only big data at a certain point in time is collected from the country’s 100+ agricultural batch markets, so the reliability of price data is not high. of.
Unlike fresh food e-commerce, there are a lot of terminal-oriented retail price data and wholesale purchase price data, and some analysis can be done, but these data are not external and not open source.
There are also some agricultural big data IT vendors that build regional e-commerce platforms one by one, and take regional agricultural big data projects one by one to obtain price data; then build their own bulk trading platforms (some single products) to obtain price data; Combined with the country’s agricultural product testing prices; purchasing part of the price data, and doing quality processing; the final business model is to sell big data on the external sales price, you may see here, everyone can guess who this company is.
In fact, it has to be identified according to the actual application scenarios. The agricultural big data applications of aquaculture, plantation, forestry, fresh food e-commerce, etc. can also be dig deep. I have the opportunity to dig deeper and show everyone.
The foundation of big data application is: the data is true, accurate, and representative; there are application scenarios, not your own YY application; using hadoop and traditional analysis software, even excel, also need to be reviewed.