System and method for identifying lengths of particles

A method for identifying lengths of one or more particles in a flow of harvested material comprising the one or more particles, the method comprising:

• using an agricultural harvester having at least one work assembly for performing one or both of harvesting a crop or processing harvested material of the crop, wherein the at least one work assembly is adjusted using one or more machine parameters;
• transporting the harvested material as a flow of the harvested material through the agricultural harvester while the agricultural harvester is operating;
• obtaining, using at least one camera on the agricultural harvester, one or more images of the flow of the harvested material;
• analyzing, in an analytical routine using a computing unit, the one or more images of the flow of the harvested material to derive particle lengths of the one or more particles of the flow of the harvested material contained in the one or more images,
• wherein the particle lengths derived in the analytical routine are an excess length that is at least greater than a predetermined length,
• wherein the analytical routine is based on a machine learning method trained to find the one or more particles with the excess length,
• wherein analyzing the one or more images comprises:
• identifying, in a respective image of the one or more images, in a proposal step one or more proposed regions in the respective image that contain potential particles with the excess length, with the one or more proposed regions being less than an entirety of the respective image so that a remainder of the respective image is not included in the one or more proposed regions, and
• classifying, in a classification step, only in the one or more proposed regions but not in the remainder of the respective image, to determine resulting regions from the proposed regions that contain the one or more particles with the excess length,
• wherein the computing unit ascertains, based on the analytical routine, an amount or number or quantity of particles with the excess length;
• wherein the computing unit controls, based on the number or the quantity of the particles with the excess length, one or more aspects of the agricultural harvester;
• wherein the analytical routine is based on a trained neural network as the machine learning method;
• wherein the proposal step uses a feature extractor based on a convolutional network directly or indirectly on the images to generate a feature space;
• wherein the proposal step uses anchor boxes on one, some or each pixel of the feature space;
• wherein the proposal step uses a classifying neural network on the anchor boxes in order to identify the one or more proposed regions from the anchor boxes;
• and
• wherein the anchor boxes are trained together with the neural network in a training step by using predetermined anchor boxes as a starting point for training the anchor boxes.