According to the Bureau of Animal and Plant Health and the relevant conditions for importing agricultural products from various countries, all export agricultural products must ensure that their quality meets the standards set by the importing country. Otherwise, the entire batch of agricultural products will be destroyed, and pest quarantine is very important; especially for scale insects. We developed an equipment that combines AI technology for identifying scale insects to help shorten the quarantine time. To use this equipment, the fruit is put on the platform and it is then transmitted through the conveyor belt. When the platform stops at the check point under a camera (360 degree), the camera will take a shoot of the fruit and the edge machine with AI model that has been trained to identify the scale insects will identify if the insects are on the fruit. The developed system is able to provide a precise and efficient pest quarantine service.

According to statistical data , the global agricultural crop production has been reduced by more than 50% due to insect pests. The traditional way to detect pests is naked-eye observation, but naked eye observation inevitably causes errors in the identification of pests. Therefore, using the system we developed, which is consist of deep convolutional neural network, the accuracy of pest recognition can achieve as high as 95% or more. The developed system is able used as an automatic quarantine system during export, and it is able to save a lot of labor cost and reduce the loss caused by missed quarantine.

When the agricultural products are delivered to the importing country, the quarantine procedure is performed. Once pests are inspected which do not meet the relevant regulations, the entire batch of products will be destroyed or returned. In order to effectively reduce losses caused by quarantine, we has developed a machine for identifying pests, which can effectively catch the small pests to reduce omissions caused by manual inspection by human eyes. In the future, the developed system can be used to detect other kind of pests to enhance the quality of products.