Final report on research project PS2010, proof of concept of automated mapping of weeds in arable fields

Many weeds occur in patches but farmers frequently spray whole fields to control the weeds in these patches. Given a geo-referenced weed map, technology exists to confine spraying to these patches. Adoption of patch spraying by arable farmers has, however, been negligible partly due to the difficulty of constructing weed maps. Building on previous DEFRA and HGCA projects, this proposal aims to develop and evaluate a machine vision system to automate the weed mapping process. The project thereby addresses the principal technical stumbling block to widespread adoption of site specific weed management (SSWM). The accuracy of weed identification by machine vision based on a single field survey may be inadequate to create herbicide application maps. We therefore propose to test the hypothesis that sufficiently accurate weed maps can be constructed by integrating information from geo-referenced images captured automatically at different times of the year during normal field activities. Accuracy of identification will also be increased by utilising a priori knowledge of weeds present in fields. To prove this concept, images will be captured from arable fields on two farms and processed offline to identify and map the weeds, focussing especially on black-grass, wild oats, barren brome, couch grass and cleavers. As advocated by Lutman et al. (2002), the approach uncouples the weed mapping and treatment processes and builds on the observation that patches of these weeds are quite stable in arable fields. There are three main aspects to the project. 1) Machine vision hardware. Hardware component parts of the system are one or more cameras connected to a single board computer (Concurrent Solutions LLC) and interfaced with an accurate Global Positioning System (GPS) supplied by Patchwork Technology. The camera(s) will take separate measurements for each of the three primary colours of visible light (red, green and blue) in each pixel. The basic proof of concept can be achieved in principle using a single camera system, but in practice systems with more than one camera may need to be installed so that larger fractions of each field can be photographed. Hardware will be reviewed regularly during the project in response to feedback from other work packages and updated as required. 2) Image capture and weed identification software. The machine vision system will be attached to toolbars of farm machinery so that images can be collected during different field operations. Images will be captured at different ground speeds, in different directions and at different crop growth stages as well as in different crop backgrounds. Having captured geo-referenced images in the field, image analysis software will be developed to identify weed species by Murray State and Reading Universities with advice from The Arable Group. A wide range of pattern recognition and in particular Bayesian Networks will be used to advance the state of the art in machine vision-based weed identification and mapping. Weed identification algorithms used by others are inadequate for this project as we intend to collect and correlate images collected at different growth stages. Plants grown for this purpose by Herbiseed will be used in the first instance. In addition, our image capture and analysis system will include plant characteristics such as leaf shape, size, vein structure, colour and textural pattern, some of which are not detectable by other machine vision systems or are omitted by their algorithms. Using such a list of features observable using our machine vision system, we will determine those that can be used to distinguish weed species of interest. 3) Weed mapping. Geo-referenced maps of weeds in arable fields (Reading University and Syngenta) will be produced with advice from The Arable Group and Patchwork Technology. Natural infestations will be mapped in the fields but we will also introduce specimen plants in pots to facilitate more rigorous system evaluation and testing. Manual weed maps of the same fields will be generated by Reading University, Syngenta and Peter Lutman so that the accuracy of automated mapping can be assessed. The principal hypothesis and concept to be tested is that by combining maps from several surveys, a weed map with acceptable accuracy for endusers can be produced. If the concept is proved and can be commercialised, systems could be retrofitted at low cost onto existing farm machinery. The outputs of the weed mapping software would then link with the precision farming options already built into many commercial sprayers, allowing their use for targeted, site-specific herbicide applications. Immediate economic benefits would, therefore, arise directly from reducing herbicide costs. SSWM will also reduce the overall pesticide load on the crop and so may reduce pesticide residues in food and drinking water, and reduce adverse impacts of pesticides on non-target species and beneficials. Farmers may even choose to leave unsprayed some non-injurious, environmentally-beneficial, low density weed infestations. These benefits fit very well with the anticipated legislation emerging in the new EU Thematic Strategy for Pesticides which will encourage more targeted use of pesticides and greater uptake of Integrated Crop (Pest) Management approaches, and also with the requirements of the Water Framework Directive to reduce levels of pesticides in water bodies. The greater precision of weed management offered by SSWM is therefore a key element in preparing arable farming systems for the future, where policy makers and consumers want to minimise pesticide use and the carbon footprint of farming while maintaining food production and security. The mapping technology could also be used on organic farms to identify areas of fields needing mechanical weed control thereby reducing both carbon footprints and also damage to crops by, for example, spring tines. Objective i. To develop a prototype machine vision system for automated image capture during agricultural field operations; ii. To prove the concept that images captured by the machine vision system over a series of field operations can be processed to identify and geo-reference specific weeds in the field; iii. To generate weed maps from the geo-referenced, weed plants/patches identified in objective (ii).

Management of plant communities on set-aside land and its effects on earthworm communities

Plant communities of set-aside agricultural land in a European project were managed in order to enhance plant succession towards weed-resistant, mid-successional grassland. Here, we ask if the management of a plant community affects the earthworm community. Field experiments were established in four countries, the Netherlands, Sweden, the UK, and the Czech Republic. High (15 plant species) and low diversity (four plant species) seed mixtures were sown as management practice, with natural colonization as control treatment in a randomized block design. The response of the earthworrns to the management was studied after three summers since establishment of the sites. Samples were also taken from plots with continued agricultural practices included in the experimental design and from a site with a late successional plant community representing the target plant community. The numbers and biomass of individuals were higher in the set-aside plots than in the agricultural treatment in two countries out of four. The numbers of individuals at one site (The Netherlands) was higher in the naturally colonized plots than in the sowing treatments, otherwise there were no differences between the treatments. Species diversity was lower in the agricultural plots in one country. The species composition had changed from the initial community of the agricultural field, but was still different from a late successional target community. The worm biomass was positively related to legume biomass in Sweden and to grass biomass in the UK. (C) 2005 Elsevier SAS. All rights reserved.

The influence of winter oilseed rape (Brassica napus ssp oleifera var. biennis) canopy size on grass weed growth and grass weed seed return

Four experiments conducted over three seasons (2002-05) at the Crops Research Unit, University of Reading, investigated effects of canopy management of autumn sown oilseed rape (Brassica napus L. ssp. oleifera var. biennis (DC.) Metzg.) on competition with grass weeds. Emphasis was placed on the effect of the crop on the weeds. Rape canopy size was manipulated using sowing date, seed rate and the application of autumn fertilizer. Lolium multiflorum Lam., L. x boucheanum Kunth and Alopecurus myosuroides Huds. were sown as indicative grass weeds. The effects of sowing date, seed rate and autumn nitrogen on crop competitive ability were correlated with rape biomass and fractional interception of photosynthetically active radiation (PAR) by the rape floral layer, to the extent that by spring there was good evidence of crop: weed replacement. An increase in seed rate up to the highest plant densities tested increased both rape biomass and competitiveness, e.g. in 2002/3, L. multiflorum head density was reduced from 539 to 245 heads/m(2) and spikelet density from 13 170 to 5960 spikelets/m(2) when rape plant density was increased from 16 to 81 plants/m(2). Spikelets/head of Lolium spp. was little affected by rape seed rate, but the length of heads of A. myosuroides was reduced by 9 % when plant density was increased from 29-51 plants/m(2). Autumn nitrogen increased rape biomass and reduced L. multiflorum head density (415 and 336 heads/m(2) without and with autumn nitrogen, respectively) and spikelet density (9990 and 8220 spikelets/m(2) without and with autumn nitrogen, respectively). The number of spikelets/head was not significantly affected by autumn nitrogen. Early sowing could increase biomass and competitiveness, but poor crop establishment sometimes overrode the effect. Where crop and weed establishment was similar for both sowing dates, a 2-week delay (i.e. early September to mid-September) increased L. multiflorum head density from 226 to 633 heads/m(2) and spikelet density from 5780 to 15 060 spikelets/m(2).

The effects of seed mix and management on the abundance of desirable and pernicious unsown species in arable buffer strip communities

Buffer strips are refuges for a variety of plants providing resources, such as pollen, nectar and seeds, for higher trophic levels, including invertebrates, mammals and birds. Margins can also harbour plant species that are potentially injurious to the adjacent arable crop (undesirable species). Sowing perennial species in non-cropped buffer strips can reduce weed incidence, but limits the abundance of annuals with the potential to support wider biodiversity (desirable species). We investigated the responses of unsown plant species present in buffer strips established with three different seed mixes managed annually with three contrasting management regimes (cutting, sward scarification and selective graminicide). Sward scarification had the strongest influence on the unsown desirable (e.g. Sonchus spp.) and unsown pernicious (e.g. Elytrigia repens) species, and was generally associated with higher cover values of these species. However, abundances of several desirable weed species, in particular Poa annua, were not promoted by scarification. The treatments of cutting and graminicide tended to have negative impacts on the unsown species, except for Cirsium vulgare, which increased with graminicide application. Differences in unsown species cover between seed mixes were minimal, although the grass-only mix was more susceptible to establishment by C. vulgare and Galium aparine than the two grass and forb mixes. Annual scarification can enable desirable annuals and sown perennials to co-exist, however, this practice can also promote pernicious species, and so is unlikely to be widely adopted as a management tool in its current form.

Management of grass weeds in overwintering stubbles

With uncertainty concerning the future of set-aside, over-wintering stubble is an attractive management option within the agri-environment scheme. Over-wintering stubbles could be included as part of rotational set-aside, benefiting farmland biodiversity. However, there is little research on managing stubbles to maximise weed seed loss, so farmers may be reluctant to adopt this option for fear of increased weed infestation. The purpose of this investigation is to develop effective management of over-wintering stubbles to minimise pernicious grass weeds in sequential crops, whilst maintaining beneficial species diversity. Research has focused on four annual grass-weeds (Alopecurus myosuroides, Anisantha sterilis, Bromus commutatus and Lolium multiflorum) of increased occurrence and/or resistance to herbicides. Hitherto, work has concentrated on the effects of stubble manipulation on weed seed germination and mortality, in particular by straw spreading or removal after harvest. The dynamics of artificially inoculated weed populations were monitored from harvest until early spring. Results obtained indicate that where straw is retained on the soil surface, it provides a favourable microclimate for seed depletion of Anisantha sterilis and Bromus commutatus through germination. Conversely, greater depletion of Alopecurus myosuroides and Lolium multiflorum seed occurred from stubbles in which a straw layer was absent. Seed recovery work provided evidence that most seeds remaining ungerminated throughout the trial period were still viable, but a large proportion of the seeds sown were unaccounted for. As these species are not generally favoured as a food source, the as yet unknown fate of these seeds has implications for subsequent grass-weed infestations.

The influence of winter oilseed rape (Brassica napus ssp oleifera var. biennis) canopy size on grass weed growth and grass weed seed return

Four experiments conducted over three seasons (2002-05) at the Crops Research Unit, University of Reading, investigated effects of canopy management of autumn sown oilseed rape (Brassica napus L. ssp. oleifera var. biennis (DC.) Metzg.) on competition with grass weeds. Emphasis was placed on the effect of the crop on the weeds. Rape canopy size was manipulated using sowing date, seed rate and the application of autumn fertilizer. Lolium multiflorum Lam., L. x boucheanum Kunth and Alopecurus myosuroides Huds. were sown as indicative grass weeds. The effects of sowing date, seed rate and autumn nitrogen on crop competitive ability were correlated with rape biomass and fractional interception of photosynthetically active radiation (PAR) by the rape floral layer, to the extent that by spring there was good evidence of crop: weed replacement. An increase in seed rate up to the highest plant densities tested increased both rape biomass and competitiveness, e.g. in 2002/3, L. multiflorum head density was reduced from 539 to 245 heads/m(2) and spikelet density from 13 170 to 5960 spikelets/m(2) when rape plant density was increased from 16 to 81 plants/m(2). Spikelets/head of Lolium spp. was little affected by rape seed rate, but the length of heads of A. myosuroides was reduced by 9 % when plant density was increased from 29-51 plants/m(2). Autumn nitrogen increased rape biomass and reduced L. multiflorum head density (415 and 336 heads/m(2) without and with autumn nitrogen, respectively) and spikelet density (9990 and 8220 spikelets/m(2) without and with autumn nitrogen, respectively). The number of spikelets/head was not significantly affected by autumn nitrogen. Early sowing could increase biomass and competitiveness, but poor crop establishment sometimes overrode the effect. Where crop and weed establishment was similar for both sowing dates, a 2-week delay (i.e. early September to mid-September) increased L. multiflorum head density from 226 to 633 heads/m(2) and spikelet density from 5780 to 15 060 spikelets/m(2).

An investigation to enhance understanding of the stimulation of weed seedling emergence by soil disturbance

Enhanced understanding of soil disturbance effects on weed seedling recruitment will help guide improved management approaches. Field experiments were conducted at 16 site-years at 10 research farms across Europe and North America to (i) quantify superficial soil disturbance (SSD) effects on Chenopodium album emergence and (ii) clarify adaptive emergence behaviour in frequently disturbed environments. Each site-year contained factorial combinations of two seed populations (local and common, with the common population studied at all site-years) and six SSD timings [0, 50, 100, 150, 200 day-degrees (d°C, base temperature 3°C) after first emergence from undisturbed soil]. Analytical units in this study were emergence flushes. Flush magnitudes (maximum weekly emergence per count flush) and flush frequencies (flushes year 1) were compared between disturbed and undisturbed seedbanks. One year after burial, SSD promoted seedling emergence relative to undisturbed seedbanks by increasing flush magnitude rather than increasing flush frequency. Two years after burial, SSD promoted emergence through increased flush magnitude and flush frequency. The promotional effects of SSD on emergence were strongest within 500 d°C following SSD; however, low levels of SSDinduced emergence were detected as late as 3000 d°C following SSD. Accordingly, stale seedbed practices that eliminate weed seedlings should occur within 500 d°C of disturbance, because few seedlings emerge after this time. However, implementation of stale seedbed practices will probably cause slight increases in weed population densities throughout the year. Compared with the common population, local populations exhibited reduced variance in total emergence measured within sites and across SSD treatments, suggesting that C. album adaptation to local pedo-climatic conditions involves increased consistency in SSD-induced emergence.

Towards more precise sugar beet management based on geostatistical analysis of spatial variability within fields

Within-field variation in sugar beet yield and quality was investigated in three commercial sugar beet fields in the east of England to identify the main associated variables and to examine the possibility of predicting yield early in the season with a view to spatially variable management of sugar beet crops. Irregular grid sampling with some purposively-located nested samples was applied. It revealed the spatial variability in each sugar beet field efficiently. In geostatistical analyses, most variograms were isotropic with moderate to strong spatial dependency indicating a significant spatial variation in sugar beet yield and associated growth and environmental variables in all directions within each field. The Kriged maps showed spatial patterns of yield variability within each field and visual association with the maps of other variables. This was confirmed by redundancy analyses and Pearson correlation coefficients. The main variables associated with yield variability were soil type, organic matter, soil moisture, weed density and canopy temperature. Kriged maps of final yield variability were strongly related to that in crop canopy cover, LAI and intercepted solar radiation early in the growing season, and the yield maps of previous crops. Therefore, yield maps of previous crops together with early assessment of sugar beet growth may make an early prediction of within-field variability in sugar beet yield possible. The Broom’s Barn sugar beet model failed to account for the spatial variability in sugar yield, but the simulation was greatly improved when corrected for early canopy development cover and when the simulated yield was adjusted for weeds and plant population. Further research to optimize inputs to maximise sugar yield should target the irrigation and fertilizing of areas within fields with low canopy cover early in the season.