Automatic Fruits / Flowers Sorting

Our computer vision system helps to sort yield by a number of visual features, such as shape, size, color, texture, surface quality, maturity, and symmetry.

Automation in agriculture plays a vital role today. Many routine tasks are going to be fully transferred to machines. One such routine task is fruit sorting. A modern imaging system based on different types of cameras together with machine learning and computer vision technologies can extract dozens of features, allowing the quality of the item to be determined and sorting it accordingly. After fine tuning, the computer vision machine works incredibly fast and provides obvious benefits to farmers.

We use the following fruit/vegetables/flowers in our solutions:

• Potato
• Apple
• Tomato
• Orange
• Strawberry
• Pineapple
• Mango
• Flowers

Our fruit-grading system includes four base steps that use different approaches: image acquisition, image pre-processing, feature extraction, and finally, sorting and grading. The most sophisticated task is feature extraction. We use artificial neural networks models for feature extraction and this model differs from one fruit to another. BitRefine group also performs tests on different types of cameras, such as hyperspectral or thermal-imaging technologies, which can be very helpful for a number of detection and evaluative tasks.