Computer Vision in Agriculture

Improve Quality & Yield by Detecting Plant & Produce Defects with Akridata

Crop disease and processing inaccuracies cost agriculture companies billions of dollars each year. Use Akridata Inspection Studio to build, test, and deploy high-accuracy computer vision models that increase both yield and margins.

13% of food is lost between harvest and retail across the globe, resulting in unnecessary food shortages as well as wasted water, energy, land, labor, and money.

Manual, labor-intensive crop and produce inspection processes simply aren’t able to catch most disease and quality issues before it’s too late. But producers are struggling to implement effective computer vision systems and often lack the image data needed to train their inspection models.

SOLUTION

Precision Growing, Sorting & Packing

Goal

Improve the precision and reliability of automated vision systems to optimize land use, improve produce grading, and reduce packaging errors.

Problem

Small pockets of crop disease in a large field — or blemished berries on a sorting line — are tough to detect by human eye and end up affecting quality and yield. But existing rule-based vision systems are either too inflexible or inconsistent to make a big impact.

Solution

Leverage Akridata’s Inspection Studio to build and deploy reliable crop and produce inspection models that are based on state-of-the-art deep learning and work with your existing imaging hardware.

See how Akridata can help.

HOW IT WORKS

Integrated Plant/Produce Inspection System

Image Data Collection

As a hardware-free solution, Inspection Studio serves as an added intelligence layer to help you make better inspection decisions. The software collects data from the image database associated with your current vision system and can be easily customized to match your specific environment.

Advanced Device Inspection

Using deep learning AI models and batch-based analysis, Inspection Studio detects even the most subtle variations in crop appearance, catching disease earlier in the growing process. The model can also improve sorting system accuracy, and verify the integrity and appearance of each finished package.

Defect Categorization and Decision

Inspection Studio determines whether identified irregularities are acceptable cosmetic variations or unacceptable abnormalities that could negatively impact taste and appearance. Model data can also help identify the root cause of processing defects in order to implement corrective procedures.

Continuous Deployment and Monitoring

The vision model continues to update and improve itself with every inspected image, delivering consistently reliable results without the fatigue factor of human inspection. It reduces manufacturing bottlenecks and can be easily scaled across multiple production facilities.

Need More?

Akridata Visual Data Platform

Don’t have an in-house Data Science team for produce defect detection?

Akridata Edge provides ready-to-use agriculture models that have been rigorously tested and refined using millions of plant images, ensuring accurate defect detection.