Information on temporal and spatial variation in weed seedling populations within agricultural fields is very important for weed population assessment and management. Primarily, spatial information allows a potential reduction in herbicide use, when post-emergent herbicides are only applied to field sections with high weed infestation levels. This paper presents a system for sitespecific weed control in sugar beet, maize, winter wheat, winter barley, winter rape and spring barley. The system includes on-line weed detection using digital image analysis, computer-based decision making and GlobalPositioning System-controlled patch spraying. In a 2year study, herbicide use with this map-based approach was reduced in winter cereals by 6-81% for herbicides against broad leaved weeds and 20-79% for grass weed herbicides. Highest savings were achieved in cereals followed by sugar beet, maize and winter rape. The efficacy of weed control varied from 85% to 98%, indicating that site-specific weed management will not result in higher infestation levels in the following crops. Keywords: weed distribution, site-specific weed control, patch spraying, on-line weed detection, decision support systems, crop rotation. GERHARDS R & OEBEL H (2006) Practical experiences with a system for site-specific weed control in arable crops using real-time image analysis and GPS-controlled patch spraying. Weed Research 46, 185-193.

Information on temporal and spatial variation in weed seedling populations within agricultural fields is very important for weed population assessment and management. Most of all, it allows a potential reduction in herbicide use, when post-emergence herbicides are only applied to field sections with weed infestation levels higher than the economic weed threshold; a review of such work is provided. This paper presents a system for site-specific weed control in sugarbeet (Beta vulgaris L.), maize (Zea mays L.), winter wheat (Triticum aestivum L.) and winter barley (Hordeum vulgare L.), including online weed detection using digital image analysis, computer-based decision making and global positioning systems (GPS)-controlled patch spraying. In a 4-year study, herbicide use with this map-based approach was reduced in winter cereals by 60% for herbicides against broad-leaved weeds and 90% for grass weed herbicides. In sugarbeet and maize, average savings for grass weed herbicides were 78% in maize and 36% in sugarbeet. For herbicides against broad-leaved weeds, 11% were saved in maize and 41% in sugarbeet.

No abstract

Intensive surveys were conducted in 2 fields in eastern Nebraska to determine the spatial stability of common sunflower, velvetleaf, green and yellow foxtail, and hemp dogbane over 4 yr (1992 to 1995). The 1st field was planted to soybean in 1992 and corn in 1993, 1994, and 1995. The 2nd field was planted to corn in 1992 and 1994 and soybean in 1993 and 1995. Weed density was sampled prior to postemergence herbicide application at approximately 800 locations per year in each field on a regular 7 m grid. The same locations were sampled every year. Weed density at locations between the sample sites was determined by linear triangulation interpolation. Weed seedling distribution was significantly aggregated, with large weed-free areas in both fields. Common sunflower, velvetleaf, and hemp dogbane patches were very persistent in diameter in the east-west and north-south directions and in location and area over 4 yr in the 1st field. Foxtail distribution and density continuously increased in each of the 4 yr in the first field and decreased in the 2nd field. A geographic information system was used to overlay maps from each year for a species. This showed that 36% of the sampled area was continuously free of common sunflower, 62.5% was free of hemp dogbane, and 11.5% was free of velvetleaf in the 1st field, but only 1% was free of velvetleaf in the 2nd field. The persistence of broadleaf weed patches suggests that weed seedling distributions mapped in one year are good predictors of future seedling distributions. Improved and more efficient sampling methods are needed.

No abstract

Site-specific weed management is the part of precision agriculture (PA) that tries to effectively control weed infestations with the least economical and environmental burdens. This can be achieved with the aid of ground-based or near-range sensors in combination with decision rules and precise application technologies. Near-range sensor technologies, developed for mounting on a vehicle, have been emerging for PA applications during the last three decades. These technologies focus on identifying plants and measuring their physiological status with the aid of their spectral and morphological characteristics. Cameras, spectrometers, fluorometers and distance sensors are the most prominent sensors for PA applications. The objective of this article is to describe-ground based sensors that have the potential to be used for weed detection and measurement of weed infestation level. An overview of current sensor systems is presented, describing their concepts, results that have been achieved, already utilized commercial systems and problems that persist. A perspective for the development of these sensors is given.

In this study, whole-plant bioassays were performed on 72 Apera spica-venti populations that have survived application of acetolactate synthase (ALS)-inhibiting herbicides in recent years. Molecular genetic analysis of the ALS gene revealed a Thr mutation at Pro 197 within 67 populations. Sequencing of the whole ALS gene from wild-type and resistant plants not carrying the above-mentioned mutation revealed the presence of a Leu mutation at Trp 574 within two populations and an Asn mutation at Pro 197 within two populations. As the Pro 197 -Asn amino acid substitution is reported for the first time in a field-selected weed population, a Cleaved Amplified Polymorphic Sequences (CAPS) marker was developed for its quick detection. In addition, one novel mutation was found within a population that coded for a His substitution at Arg 377 . Enzyme assays confirmed a significant reduction in inhibition of ALS activity compared with the wild type. This population showed resistance to sulfonylureas (SUs) and crossresistance to sulfonylaminocarbonyltriazolinones (SCTs) and triazolopyrimidines (TPs) within the whole-plant bioassays. ALS protein sequence alignments from weedy and cultural plants revealed that the Arg377 is highly conserved among known wild-type enzymes. In agreement with existing literature concerning the structure and mechanisms of inhibition of plant ALS, this mutation is probably involved in target-site resistance to ALS inhibitors. Our results suggest that further single-nucleotide polymorphisms impairing proper herbicide performance might be selected within field populations in the near future, making the shortand long-range evolution of target-site resistance difficult to predict depending solely on herbicide use history.

The increasing public concern about food security and the stricter rules applied worldwide concerning herbicide use in the agri-food chain, reduce consumer acceptance of chemical plant protection. Site-Specific Weed Management can be achieved by applying a treatment only on the weed patches. Crop plants and weeds identification is a necessary component for various aspects of precision farming in order to perform on the spot herbicide spraying or robotic weeding and precision mechanical weed control. During the last years, a lot of different methods have been proposed, yet more improvements need to be made on this problem, concerning speed, robustness, and accuracy of the algorithms and the recognition systems. Digital cameras and Artificial Neural Networks (ANNs) have been rapidly developed in the past few years, providing new methods and tools also in agriculture and weed management. In the current work, images gathered by an RGB camera of Zea mays, Helianthus annuus, Solanum tuberosum, Alopecurus myosuroides, Amaranthus retroflexus, Avena fatua, Chenopodium album, Lamium purpureum, Matricaria chamomila, Setaria spp., Solanum nigrum and Stellaria media were provided to train Convolutional Neural Networks (CNNs). Three different CNNs, namely VGG16, ResNet–50, and Xception, were adapted and trained on a pool of 93,000 images. The training images consisted of images with plant material with only one species per image. A Top-1 accuracy between 77% and 98% was obtained in plant detection and weed species discrimination, on the testing of the images.