Researchers explain the operational principles of ai worm monitoring equipment。
Sichuan rural daily news, full of media, romin
“pests that were previously tested were screened by manpower, and not only were they hard but also often missed. Now ai will be able to find out what's going on." on 15 october, the food and oil keeper wang yalon pointed to the five “mixed scouts” standing in the warehouse in the shallow silo of chengdu qingbai warehouse company。
These equipment, similar to the 360° rotation camera, are the “digital scouts” for the monitoring of ai worms, which allow traditional food reserves to “change their head”。
"digital sentinel" on duty
95% accuracy
Under the wet and hot climate of sichuan, more than 20 species of food-storage pests, such as corn elephants, red-coloured grains, and grain vermin, are growing rapidly, with a high incidence of single-carrier pests in the summer。
"the system is more accurate than the 20 or so species of food storage pests in the database, even incubated larvae. Ma, assistant researcher at the china food supply institute, detailed the principles of the equipment for the “mixed reconnaissance” unit, the aid integration depth learning algorithm module, which allows real-time pictures of the food surface, capture of pest dynamics, and web-based transmission of data to the aid system。
In the system interface, journalists see a clear picture of the type of pest, the number, the location where it occurs, the time of discovery, etc. “we can set the time, spacing and point of inspection on demand, cover the focal area of the custodian, and also function on holidays when no one is on duty.” according to ma yi, the system not only reduces the frequency with which custodians enter the warehouse more frequently, but also reduces the time for single-barrel infestation from half a day to 10 minutes, significantly improving monitoring efficiency。
It's called the "step."
Dynamic green prevention and control programmes
Pest disposal is another challenge. “the most frequent use of phosphorous aluminium fumigation in pests has previously been found to be insecticidal, but there is a risk that chemical agent residues may also affect the quality of food, and green pests have been in an urgent direction for the industry.”
“when pests are found, the system automatically links to the `silent' food storage model and generates customized disposal programmes.” in a presentation by lee, software engineer at the china food storage institute, the disposal programme gave priority to nitrogen-temperature conditioning, alternative cryogenic controls, and ancillary inner circulation winds。
In his presentation, li lian said that the large food storage model was an intelligent system designed specifically for the food storage industry, based on a nine-day, deepseek and tunjia production model base, integrating over 200,000 high-quality food storage knowledge modules, covering food storage processes, bug control, policy regulation, etc., and incorporating the continuous and iterative experience of industry experts. “models can combine food varieties, silo temperatures, and air-condensation to recommend the most appropriate green control to avoid blind use of medicines.” li xi said。
Laboratory hard core support
More than 20 living insects “train” ai
The high accuracy of ai is dependent on the technical clearance of scientific institutions. At the food insect laboratory of the china food storage institute, journalists saw the country's largest stock of food infestation species. More than 20 living insects, such as corn elephants and cereals, are raised in sterile hot and wet cultures。
"the living insects are the training materials of the ai system." ma's motto is that, in the culture room, the corn elephants inside are active between the corn particles, “we will take images of the insect eggs, larvae, larvae, adult insect stages, mark morphological characteristics, build a well-developed pest database, and allow ai to `know' every pest”
The laboratory also has the green control laboratory for food pests, the engineering laboratory and the pharmacology laboratory, of which the green control laboratory has the country's first visualized digital food pest storage basic research database platform, which has the capabilities of temperature control, wet control, light control and gas concentration control, all of which are centrally controlled by a medium-controlled computer and minimizes human error。
At present, this aid surveillance system has been included in the pilot list of science and technology results empowerment reforms for the first state-owned enterprises to be elected to the ncec, and received the china food and oil institute's science and technology award for 2024, which provides scientific support for food conservation erosion。
