Michael Starke

Mr. Michael Starke

orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 August 2020 in Sustainability
Verified
Journal Article
Open access
PDF file
MSMichael Starke
Michael Starke
CDCédric Derron
Cédric Derron
FHFelix Heubaum
Felix Heubaum
MZMartin Ziesak
Martin Ziesak
In 2019, the machine manufacturer HSM presented a forwarder prototype for timber hauling in cut-to-length processes fitted with a new 10-wheel triple-bogie (TB) setup approach aimed at promoting sustainable forest management by reducing the ecological impact of forest operations, especially under soft-soil working conditions. The purpose of our study was to assess the resulting soil-protection effect emerging from additional wheel-contact surface area. For this, the rut development under known cumulative weight, related to the soil conditions of shear strength and moisture content, was recorded for later comparison. Additional terrestrial laser scanning (TLS) was used to generate a multi-temporal digital terrain model (DTM) in order to enhance the data sample, assess data quality, and facilitate visualization of the impact of local disturbance factors. In all TB configurations, a rut depth of 10 cm (5.8–7.2 cm) was not exceeded after the hauling of a reference amount of 90 m3 of timber (average soil shear strength reference of 67 kPa, volumetric water content (VMC) 43%). Compared to a reference dataset, all observed configurations ranked in the lowest-impact machine categories on related soil stability classes, and the configuration without bogie tracks revealed the highest machine weight to weight distribution trade-off potential.

2 September 2022 in International Journal of Forest Engineering
Verified
Journal Article
Open access
MSMichael Starke
Michael Starke
CGChris Geiger
Chris Geiger
When assessing forest road conditions, information about waterlogged areas on gravel roads brings high practical value when used as an indicator for road wear. Around these perimeters, lowered binding forces of the construction material reduce the stability of the road, which induces accelerated road damage. When a road is actively used to access a logging site under humid weather or thawing conditions, road wear can build up fast and make further use of the road critical. In this study, a deep learning algorithm was trained to test the detection of a combined observation of waterlogged appearances on forest roads from video and image data, collected from a passing vehicle's perspective. The training of a YOLO v5s model achieved an F1-score of 0.59 and shows the applicability of this approach with high confidence of detection. Evaluating further training characteristics such as precision, recall, and the object size-related detection confidence reveals challenges for a successful application in terms of undetected objects, variation of objects in the training step, the required amount of training data and the object distance focused.

2 September 2021 in Forests
Verified
Journal Article
Open access
PDF file
MSMichael Starke
Michael Starke
AKAnthonie Kunneke
Anthonie Kunneke
MZMartin Ziesak
Martin Ziesak
Forest roads are an important element in forest management as they provide infrastructure for different forest stakeholder groups. Over time, a variety of road assessment concepts for better planning were initiated. The monitoring of the surface cross-section profile of forest roads particularly offers the possibility to take early action in restoring a road segment and avoiding higher future costs. One vehicle-based monitoring system that relies on ultrasound sensors addresses this topic. With advantages in its dirt influence tolerance and high temporal resolution, but shortcomings in horizontal and vertical measuring accuracy, the system was tested against high resolution terrestrial laser scanner (TLS) data to find and assess working scenarios that fit the low- resolution measuring principle. In a related field test, we found low correct road geometry interpretation rates of 54.3% but rising to 91.2% under distinctive geometric properties. The further applied line- and segment-based method used to transform the TLS data to fit the road scanner measuring method allows the transfer of the road scanner evaluation principle to point-cloud or raster data of different origins.

10 November 2021 in Croatian journal of forest engineering
Verified
Journal Article
Open access
MSMichael Starke
Michael Starke
CGChris Geiger
Chris Geiger
When investigating the forwarding process within the timber supply chain, insufficient data often inhibits long-term studies or make real-time optimisation of the logistics process difficult. Information sources to compensate for this lack of data either depend on other processing steps or they need additional, costly hardware, such as conventional crane scales. An innovative weight-detection concept using information provided by a commonly available hydraulic pressure sensor may make the introduction of a low-cost weight information system possible. In this system, load weight is estimated by an artificial neural network (ANN) based on machine data such as the hydraulic pressure of the inner boom cylinder and the grapple position.In our study, this type of crane scale was set up and tested under real working conditions, implemented as a cloud application. The weight scale ANN algorithm was therefore modified for robustness and executed on data collected with a commonly available telematics module. To evaluate the system, also with regard to larger sample sizes, both direct weight-reference measurements and additional volume-reference measurements were made. For the second, locally valid weight-volume conversion factors for mainly Norway spruce (Picea abies, 906 kg m-3, standard error of means (SEM) of 13.6 kg m-3), including mean density change over the observation time (–0.16% per day), were determined and used as supportive weight-to-volume conversion factor.Although the accuracy of the weight scale was lower than in previous laboratory tests, the system showed acceptable error behaviour for different observation purposes. The twice-observed SEM of 1.5% for the single loading movements (n=95, root-mean-square error (RMSE) of 15.3% for direct weight reference; n=440, RMSE=33.2% for volume reference) enables long-term observations considering the average value, but the high RMSE reveals problems with regard to the single value information.The full forwarder load accuracy, as unit of interest, was observed with an RMSE of 10.6% (n=41), considering a calculated weight-volume conversion as reference value. An SEM of 5.1% for already five observations with direct weight reference provides a good starting point for work-progress observation support.

30 May 2024 in European Journal of Forest Research
Verified
Journal Article
Open access
PDF file
FSFrancesco Sforza
Francesco Sforza
MSMichael Starke
Michael Starke
PDPatrick Dietsch
Patrick Dietsch
PTPeter Thür
Peter Thür
ELEmanuele Lingua
Emanuele Lingua
MZMartin Ziesak
Martin Ziesak
The efficiency of forest logging operations can be strongly affected by the layout of the harvesting pattern, which is usually based on silvicultural constraints and technical feasibility. Specifically, individual tree volume and the spatial distribution of trees significantly impact the overall harvesting performance. Spatial optimization of tree selection at the forest stand level may improve timber harvest efficiency by maximizing key performance indicators, such as the economic benefit, under given operational and silvicultural constraints. In this study, we applied two harvesting operation-optimization approaches based on integer programming for uphill cable yarding operations in mountain areas, including tree selection and load maximization. The first approach involves tree selection based on single tree harvest, while the second one performs tree selection based on tree clusters harvest per work cycle. As input elements a productivity model, derived by time-motion study with a Mounty MT50-2 and individual tree parameters extracted from high-resolution airborne laser scanning data, were prepared. Single tree information was further rated by financial value, and subsequently combined with the productivity model, allowing a detailed breakdown of operational costs. The results showed that optimizing the tree selection while respecting the allowable cut timber volume established in the harvesting plan can improve the efficiency of forest operations. The cluster approach was shown to be more efficient in terms of economic benefit compared to the actual selection, with an increase of 24.94%. However, the single tree approach resulted in a decrease of economic benefit compared to the actual selection, with a decrease of 22.85%.

1 September 2023 in Schweizerische Zeitschrift fur Forstwesen
Verified
Journal Article
Closed access
PDF file
MSMichael Starke
Michael Starke
PDPatrick Dietsch
Patrick Dietsch
MZMartin Ziesak
Martin Ziesak
Seilkransimulation 4.0: Die HAFL erarbeitet ein System zur Unterstützung der Planung und Durchführung von Holzerntemassnahmen mit Seilkrananlagen. Die Simulation stellt in einer tabellarischen Zusammenstellung die Kosten und die Leistungsparameter des Hiebes dar und ermöglicht eine 3-D-Animation über den gesamten Simulationszeitraum des Arbeitsablaufs.