Systems and methods for plant species identification

Abstract

A computer-implemented method, computer program product and computer system (100) for identifying weeds in a farmland using a dual task convolutional neural network (120) having a topology of: an intermediate module (121) that performs a classification task associated with a first loss function (LF1); and a semantic segmentation module (122) that performs a segmentation task associated with a different second loss function (LF2). The intermediate module and the segmentation module are trained together in consideration of the first and second loss functions (LF1, LF2). The system performs a method comprising receiving a test input (91) comprising an image of a weed plant showing a crop plant of a crop species in a farmland and showing one or more weed species in the crop plant; predicting the presence of one or more weed species (11, 12, 13) present in the respective tiles; outputting the corresponding intermediate feature map to a segmentation module as the output of a classification task; generating a mask for each weed species as a segmentation output of the second task by extracting multi-scale features and context information from the intermediate feature map and serially connecting the extracted information to perform semantic segmentation; and generating a final image (92) indicating, for each pixel, whether the pixel belongs to a particular weed species, and if so, which weed species the pixel belongs to.

Description

technical field [0001] The present invention relates generally to electronic data processing, and more particularly to image processing methods, computer program products and systems for weed identification in agricultural fields. Background technique [0002] The presence of weed communities in farmland has a negative impact (see H. van Heemst, "The influence of weed competition on crop yield", Agricultural Systems, Vol. 18, No. 2, pp. 81-93, 1985). A weed in the context of this document refers to any plant that grows in a field other than the crops grown in the field. Research has identified two main reasons: competition and plant health issues. Certain plant species, such as weeds, compete with crops for soil, nutrients and sunlight, resulting in slow and reduced crop growth. In addition, some weeds are hosts of pests and diseases. To do this, farmers use herbicides to eliminate or control weed populations. [0003] The table below includes examples of some weeds and ...

AI Technical Summary

Problems solved by technology

The pixel accuracy obtained with this state of the art approach has not reached a satisfactory level to support farmers in optimizing their activities to protect their fields

Images