[PROJECT NAME] Intelligent Wood Identification
[PROJECT OVERVIEW]
To address the challenges such as high expertise requirements for identifying imported/exported wood and associated invasive pests, difficulty of instant on-site identification, and lengthy, costly laboratory identification, China Customs has developed a mobile intelligent wood identification device—the “WOOD AI” system. Leveraging AI image recognition and smartphone microscopic photography, it establishes a new regulatory model integrating intelligent identification, knowledge retrieval, and cloud-based support. Deployed at 43 ports and laboratories throughout China Customs territory, the system enables Customs officers to complete identification within 10 minutes with over 90% accuracy, reducing on-site processing time by 99%, shortening clearance duration, and lowering regulatory costs significantly.
[PROJECT DETAILS]
I. Key Approaches
1. Training of specialized high-accuracy AI models. Leveraging years of accumulated specimens of common imported/exported wood species and invasive pests, China Customs has built a training image database with more than 500,000 images. The identification models have been optimized according to the characteristics of target subjects, resulting in several highly accurate intelligent recognition models that will continue to improve through iterative upgrades.
2. Development of high-quality smartphone-compatible microscopes. Customized micro-lenses have been developed to enable accurate identification of wood species and invasive pests. When attached to smartphones, these lenses greatly enhance close-up imaging, capable of replacing professional microscopes while ensuring high-quality imaging for precise identification.
3. Building of portable intelligent identification mode. The system enables wood and pests recognition on mobile devices (phones, tablets, Customs individual inspection equipment), without reliance on specialized environments or hardware, greatly enhancing convenience and accelerating identification process.
4. Integration of database and expert collaboration. The system incorporates digital knowledge database and case database for wood and pests, along with an online expert support module to assist Customs inspection officers in resolving regulatory challenges.
II. Achievements
1. Enhanced Customs Inspection Efficiency. At pilot sites, the traditional 10-day process for wood and pests laboratory identification has been reduced to 10-minute AI recognition, reducing the identification time by approximately 99%.
2. Improved Customs Inspection Accuracy. The system has effectively achieved over 90% accuracy for wood and pests identification, minimizing human errors.
3. Reduced Administrative Costs. At pilot sites, over 50% of wood and pests can be rapidly identified and cleared, without sampling and laboratory identification procedures, thereby reducing both administrative costs and clearance duration.
III. Promotion and Application
Fully integrated in an 8kg portable toolkit, “WOOD AI” system features a bilingual (Chinese-English) interface, designed for global Customs inspection sites and laboratories. Officers without professional background can master operations in 10 minutes after brief instruction - making it ideal for developing countries seeking to enhance supervision capacity cost-effectively.
Currently, “WOOD AI” can identify 145 wood species (including 32 CITES-listed endangered species), 32 insect species, and 110 weed seeds species, with precise differentiation between softwood and hardwood (average accuracy: >90%). Furthermore, this system includes built-in knowledge dictionaries and expert support functions to assist Customs inspection officers in decision-making. China Customs continues algorithm research to expand species coverage and application scenarios, including enhanced identification of CITES-listed endangered species.