Browsing by Subject "Image analysis"

Now showing 1 - 3 of 3

Camera-guided weed hoeing in winter cereals with narrow row distance
(2020) Gerhards, Roland; Kollenda, Benjamin; Machleb, Jannis; Möller, Kurt; Butz, Andreas; Reiser, David; Griegentrog, Hans-Werner
Farmers are facing severe problems with weed competition in cereal crops. Grass-weeds and perennial weed species became more abundant in Europe mainly due to high percentages of cereal crops in cropping systems and reduced tillage practices combined with continuous applications of herbicides with the same mode of action. Several weed populations have evolved resistance to herbicides. Precision weed hoeing may help to overcome these problems. So far, weed hoeing in cereals was restricted to cropping practices with row distances of more than 200 mm. Hoeing in cereals with conventional row distances of 125–170 mm requires the development of automatic steering systems. The objective of this project was to develop a new automatic guidance system for inter-row hoeing using camera-based row detection and automatic side-shift control. Six field studies were conducted in winter wheat to investigate accuracy, weed control efficacy and crop yields of this new hoeing technology. A three-meter prototype and a 6-meter segmented hoe were built and tested at three different speeds in 150 mm seeded winter wheat. The maximum lateral offset from the row center was 22.53 mm for the 3 m wide hoe and 18.42 mm for the 6 m wide hoe. Camera-guided hoeing resulted in 72–96% inter-row and 21–91% intra-row weed control efficacy (WCE). Weed control was 7–15% higher at 8 km h−1 compared to 4 km h−1. WCE could be increased by 14–22% when hoeing was combined with weed harrowing. Grain yields after camera-guided hoeing at 8 km h−1 were 15–76% higher than the untreated control plots and amounted the same level as the weed-free herbicide plots. The study characterizes camera-guided hoeing in cereals as a robust and effective method of weed control.
DeepCob: precise and high-throughput analysis of maize cob geometry using deep learning with an application in genebank phenomics
(2021) Kienbaum, Lydia; Correa Abondano, Miguel; Blas, Raul; Schmid, Karl
Background: Maize cobs are an important component of crop yield that exhibit a high diversity in size, shape and color in native landraces and modern varieties. Various phenotyping approaches were developed to measure maize cob parameters in a high throughput fashion. More recently, deep learning methods like convolutional neural networks (CNNs) became available and were shown to be highly useful for high-throughput plant phenotyping. We aimed at comparing classical image segmentation with deep learning methods for maize cob image segmentation and phenotyping using a large image dataset of native maize landrace diversity from Peru. Results: Comparison of three image analysis methods showed that a Mask R-CNN trained on a diverse set of maize cob images was highly superior to classical image analysis using the Felzenszwalb-Huttenlocher algorithm and a Window-based CNN due to its robustness to image quality and object segmentation accuracy (r = 0.99). We integrated Mask R-CNN into a high-throughput pipeline to segment both maize cobs and rulers in images and perform an automated quantitative analysis of eight phenotypic traits, including diameter, length, ellipticity, asymmetry, aspect ratio and average values of red, green and blue color channels for cob color. Statistical analysis identified key training parameters for efficient iterative model updating. We also show that a small number of 10–20 images is sufficient to update the initial Mask R-CNN model to process new types of cob images. To demonstrate an application of the pipeline we analyzed phenotypic variation in 19,867 maize cobs extracted from 3449 images of 2484 accessions from the maize genebank of Peru to identify phenotypically homogeneous and heterogeneous genebank accessions using multivariate clustering. Conclusions: Single Mask R-CNN model and associated analysis pipeline are widely applicable tools for maize cob phenotyping in contexts like genebank phenomics or plant breeding.
Untersuchungen zum Herbizidstress in Zuckerrüben mit drei feldtauglichen optischen Sensoren und Methoden der Bildanalyse
(2014) Roeb, Johannes
Weed management in sugar beets is based on the repeated use of herbicide mixtures after crop emergence. Due to the limited selectivity of active ingredients, herbicide treatments not only control the weeds but will reduce the growth of sugar beets also. Yield losses due to herbicide stress are expected to range between 5-15%. Using optical sensors is a nondestructive method to assess changes in reflection, leaf fluorescence or chlorophyll fluorescence kinetics induced by herbicides. To evaluate the applicability of three optical sensors for assessing herbicide stress and to measure the influence of herbicides and herbicide mixtures on sugar beets, a pot experiment was performed at the University of Hohenheim, Germany. Sugar beets were grown under natural light and temperature conditions and treated with the active ingredients metamitron, phenmedipham, desmedipham, ethofumesate, triflusulfuron-methyl and dimethenamid-P in their commercially available formulation and practical dosage. In total five single herbicides as well as five different herbicide mixtures were applied in the cotyledon stage (EC 10), the two-leaf stage (EC 12) or the four- to six-leaf stage (EC 14/16) of sugar beets. Stress reactions were monitored with three optical sensors: Images of a digital camera (Canon EOS 1000D) were used to determine leaf coverage area, plant shape and leaf color. Measurements were performed about every second day and a growth index had been calculated. A multispectral fluorometer (FORCE-A MULTIPLEX®) was used to detect the blue-green fluorescence, red fluorescence and far-red fluorescence and to calculate fluorescence indices. A portable imaging sensor of chlorophyll fluorescence kinetics (WALZ IMAGING PAM) was used on a daily basis to determine changes in the maximum quantum efficacy (Fv/Fm) induced by herbicide treatments. Four respectively two weeks after the treatment sugar beets were harvested for dry matter analysis. For each of the herbicide treatments and for each of the three application dates five Mitscherlich pots were used for replication. Each pot had about four sugar beet plants. Based on digital imaging it was possible to measure leaf coverage area and determine growth depressions induced by herbicide treatments containing mixtures of the active ingredients phenmedipham, desmedipham and ethofumesate. Herbicide mixtures with more active ingredients increased the stress reaction of sugar beets. Differentiation between untreated plants and sugar beets treated with different herbicides or mixtures was possible a few days after application. Results were correlated with dry matter. Changes in plant shape parameters indicated a delayed development of herbicide treated plants. Higher red content of leaf color was attributed to a relative loss of chlorophyll. Measurements with the FORCE-A MULTIPLEX® fluorometer detected an increase in red and far-red fluorescence but not blue-green fluorescence within 1-2 days after treatments with the aforementioned mixture of active ingredients. About the same trends were found at all application dates. Most fluorescence values were affected by growth stage and leaf area of sugar beets. Thus, although differences between treated and untreated plants were strong, it was not possible to discriminate between stress reactions on different herbicide treatments. Based on the maximum quantum efficacy (Fv/Fm) measured with the WALZ IMAGING-PAM chlorophyll fluorescence sensor previous studies, describing the time course of stress reaction on application of PSII-inhibitors in sugar beets were confirmed. After a strong decrease of Fv/Fm within one day, recovery to the initial value was observed within ten days. Quantification of herbicide stress induced by PSII-inhibitors was possible due to different intensities and durations of the stress reaction. Photochemical stress response to treatments with metamitron or chloridazon was lower than with products containing phenmedipham or desmedipham. Stress reaction on herbicide mixtures not only depended on content of PSII-inhibitors but also on formulation. Weather conditions were more important than the sugar beet development stage considering the stress reaction. Observations from previous studies, indicating an increase in herbicide stress after precipitation and at low temperatures, were also confirmed in this study. Differences in stress reactions of cotyledons and first true leaves can be explained by a higher uptake of herbicides in young tissues. The influence of other herbicides, mixtures, dosages and formulations on herbicide stress in sugar beets has to be further investigated. Moreover the complex interrelations between sugar beet development stage, weather conditions and stress reaction could only be investigated in systematic field trials. For the measurement of stress reactions on herbicides the described optical sensors and methods can be used, each having different advantages and disadvantages.