FarmiSpace understands that knowing the distribution and yield of crops is crucial. It affects how governments can more effectively allocate water resources, manage market conditions to avoid supply-demand imbalances, and monitor land usage. For traders, nothing is more important than knowing the current season's yield to make purchasing and selling decisions at more attractive prices.

Therefore, FarmiSpace utilizes satellite data combined with our AI technology to reveal local land usage types and identify crops planted in agricultural areas. Specifically, we focus on rice cultivation, providing yield prediction.

Land Usage and Crop Identification

We categorize land usage types into agricultural land, fish ponds, urban areas, water reservoirs, forests, rivers, and channels. By analyzing different bands of reflected light, including infrared and near-infrared, our AI model can identify large-scale land areas.

Subsequently, we further identify agricultural land by analyzing the changing reflection patterns over time, which differ based on the growth stages of various crops. Our AI model classifies the distribution and area of crops planted in agricultural land, including rice, sugarcane, citrus fruits, and others.

Results in Taiwan

Rice is the primary staple crop in East Asia. Dapi, Yunlin County, Taiwan, has the highest rice cultivation area in the country, highlighting its importance. Understanding the location and cultivation area aids government management of farmers and water resource allocation. FarmiSpace conducted identification in Dapi for the first quarter of 2023, covering a total area of 4,434 hectares. After data annotation and model training, the accuracy of land usage and crop identification reached 85%.

The AI identification results revealed a rice cultivation area of 3,085 hectares, accounting for 69.6% of the total area. The yield prediction is 26,970 metric tons, with an estimated value of $30.5 million USD.

Results in Indonesia

In addition to the case in Dapi, FarmiSpace's identification technology can be applied globally. In Karawang, West Java Province, Indonesia, an important agricultural production area, FarmiSpace also conducted crop identification.

The total identified area in Karawang was 192,075 hectares, with rice cultivation covering 106,660 hectares, accounting for 55.5% of the Karawang area. The predicted yield is 552,716.3 metric tons, with an estimated value of $475.3 million USD.

Applicable Globally for Any Crop

Utilizing satellite imagery, our algorithm can identify any location on Earth, supporting the recognition of various crops beyond rice.

The distribution and yield of crops enable stakeholders to estimate crop production in a region, allowing them to:

Understand current land usage without physical inspection.
Determine the location and area of crop fields to prevent market imbalances.
Estimate potential crop yields for import-export planning.
Allocate water resources more effectively to prevent damage or waste.

FarmiSpace's AI technology supports diverse crop and land usage identification tasks, helping stakeholders grasp land usage and plan procurement operations in advance for smoother operations.

The application of FarmiSpace X AI technology will continue to expand in the future, providing innovative agricultural solutions.