Iranian Scientists Utilize AI for Early Detection of Plant Diseases

In this research, they have tried to provide an efficient and fast solution for diagnosing diseases by combining the appearance of plant leaves and computer analysis without the need for an expert to be physically present on the site.

In this research, first, images of diseased and healthy plant leaves were collected and noise in the image was removed by using software filters. Then, the leaf-related part of the image was separated from the image and important features like its color, texture, and shape were analyzed. The color of the leaves was examined in different color spaces and parameters like the degree of redness or yellowness were extracted. The texture of the leaves was also examined using specific image patterns to identify the appearance differences between healthy and diseased leaves.

Also, the overall size and shape of the leaves were also examined to detect structural changes caused by disease. Finally, all this information was fed into classification algorithms to automatically identify whether the leaf belonged to a healthy or diseased plant.

The findings of this study show that the designed system was able to identify diseases in 9 different plant species, including apple, corn, grape, potato, tomato, cherry, peach, pepper and strawberry, with high accuracy. Based on tests conducted on a dataset of over 87,000 images, the accuracy of this method was about 98 percent using an algorithm called ‘random forest’ and nearly 95% using another algorithm called ‘support vector machine’.

In a relevant development in May, a technological company in Iran had also succeeded in manufacturing a device capable of quickly identifying the concentration of bacteria in solutions and measuring the quality of engine oil by using artificial intelligence.

“Our company's product is the Optomac (Bactometer) device, which can detect the concentration of bacteria in solutions; of course, bacteria that are in suspension and visible,” Yasin Sama’yee, the executive director of the company stationed at the incubator center of the Islamic Azad University’s Isfahan branch, told ANA on the sidelines of Asr-e Omid event in the Northwestern city of Tabriz.

“The customers of this device include drinking water production companies, food, research companies, universities and university laboratories,” he added.

“This device is small and portable and does not require many accessories; it just connects to electricity and starts working,” Sama’yee said.

He explained that with the help of artificial intelligence, this device performs detection in real time, while the same process in a laboratory using the traditional method may take several days or a week.

4155/v