Abstract: A conveyor-based system and method for assessing perishable consumable product quality include at least one conveyor for moving products and one or more image capture devices. A processor-based control circuit estimates the time a product will be at the center of an image capture device's field of view, considering conveyor speed, product type, size, and shape. At this estimated time, the control circuit causes the image capture device to capture images from various perspectives. These captured images facilitate efficient and precise quality assessment, including defect detection, identification, and severity evaluation using machine learning models and reference image data.

Abstract: A system for calibrating the focus of image capture devices positioned to capture images of products is disclosed. The system includes a product support surface, at least one image capture device, and a control circuit. The control circuit receives an input indicating a product is positioned within the image capture device's field of view, identifies the product's type and/or size, and then adjusts the image capture device's lens to a focus complementary to the identified product type and/or size. A method is also disclosed, involving supporting a product, capturing an image, detecting its presence, identifying its type/size, and adjusting the lens focus for optimal image capture. This enables the acquisition of maximally focused images for efficient and precise quality assessment of perishable consumer products.

Abstract: Systems and methods for assessing the quality of a plurality of perishable, consumable products include at least one conveyor. At least one image capture device may be positioned proximate the conveyor to continuously capture at least one image of the product moving on the conveyor from at least one perspective. A control circuit obtains and processes the image. In response to determining that the image contains a depiction of the product, the control circuit further processes the image to identify the product and detect one or more defects on a surface of the identified product. The control circuit may match the depiction of the product to a stored reference model. The circuit determines a size of the defect, translates this size into a defect severity level, and correlates the severity level to a predetermined threshold defect severity level. A notification is output indicating if the product is of acceptable quality.