2D and 3D shape inspection

Product shape is an important quality indicator. Wageningen Food & Biobased Research has developed various technologies to automatically and objectively determine and analyse product shape.

Shape is an important characteristic in determining the quality of a product. Consumers, for example, like specific types of vegetables and fruit, such as bell peppers, tomatoes, cucumbers and potatoes, to have a straight, symmetrical form. Deformities often indicate disease. In tomato seedlings, the shape of the seed leaves and heart leaves are often an important indicator for the future quality, as is the stem of the two-week old plant. In addition to the shape being a key indicator for quality, it is also relevant in the sorting process. Wageningen Food & Biobased Research has developed advanced technologies that enable companies to objectively determine and analyse the shape of their products. 

Objective shape inspection

Leveraging on our knowledge of 2D and 3D shape inspection, we developed a large number of applications for companies in the agro-food sector. Below are three examples:

  • Tomato seedlings: We developed a technology in which cameras observe tomato seedlings from various angles, based on which we create a 3D model of the plant. Many characteristics determine the quality of the plant, such as its volume, shape of the seed leaves and straightness of the stem. The plants are automatically sorted based upon these factors.
  • Bell peppers: Laser triangulation creates a precise 3D measurement of the bell peppers. We then determine shape characteristics such as the number of lobes, the depth of the constrictions and the level of symmetry.
  • Vine tomatoes: Based on 3D measurements, we determine whether the shape meets the standard and whether the uniformity of the vine is adequate. We also estimate the volume and weight of the individual tomatoes. The vines are then sorted automatically and cut by a robot.

3D reconstruction

Making a good 3D reconstruction is essential in order to establish a proper analysis of product shape. Depending on customer requirements we can use various sensors and technologies:

  • MARVIN: The MARVIN technology, developed by Wageningen Food & Biobased Research, makes it possible to look at objects from various perspectives using up to 12 cameras. Based on the object’s silhouette in every camera image, we can quickly create a reliable 3D reconstruction.
  • Laser triangulation: In this technology, we project a straight laser beam on an object. The line is monitored by a camera, and the height profile of the object is calculated by means of the deformation of the laser beam on the object. By running the object underneath a sensor on a transport belt, for example, we can determine the 3D shape of the entire object.
  • Stereo vision: This method is similar to how people observe depth with two eyes. Two cameras capture an image of an object after which the computer calculates a 3D reconstruction.