Browsing by Subject "hakkuukonetieto"

Sort by: Order: Results:

Now showing items 1-3 of 3
  • Saukkola, Atte (Helsingin yliopisto, 2017)
    Kuviotason metsävaratietoa tarvitaan Suomessa yksityismetsätalouden suunnittelua ja metsäteollisuuden puunkorjuun suunnittelua varten. Hakkuukoneet mittaavat ja tallentavat kaadetuista rungoista useita läpimittoja sekä käyttöosan pituuden. Lisäksi jokaiselle kaadetulle puulle määritetään puulaji ja sijainti. Hakkuukonemittausten hyödyntäminen voisi tehostaa metsävaratiedon tuottamista ja parantaa sen laatua. Tällä hetkellä hakkuukoneella kerätyn puutiedon hyödyntämistä rajoittaa aineistojen saatavuus ja heikko yksittäisen kaadetun puun sijaintitarkkuus. Hakkuulaitteen paikannuksen vaikutusta yksittäisen puun sijainnin määrittämisen tarkkuuteen tutkittiin vertaamalla puun todellista sijaintia estimaatteihin, jotka perustuivat hakkuukoneen ja hakkuulaitteen sijainteihin puun kaatohetkellä. Hakkuulaitteella kerätyn puutiedon käyttöä aluepohjaisen puustotulkintamenetelmän aputietona tutkittiin käyttämällä kaukokartoitusaineistoihin perustuvien puustotunnusten ennustemallien opetusaineistona Uudenmaan alueelta kerättyyn hakkuukoneaineistoon muodostettuja ympyräkoealoja sekä maastokoealoja. Yksittäisten puiden paikannustavan, koealakokoon sekä tavallisten maastokoealojen määrä vaikutuksen selvittämiseksi puustotunnusten ennustamalleja laadittiin erilaisilla opetusaineistoilla. Vertailtavina tunnuksina olivat puustotunnusten keskineliövirheen neliöjuuret (RMSE) ja harhat, jotka laskettin leimikkotason puustotunnusten ennusteiden ja vertailukohtana käytettyjen leimikkotason hakkuukonemittausten erotuksista. RMSE-arvot ja harhat laskettiin myös suhteutettuna vertailuaineiston puustotunnusten keskiarvoihin. Tulosten perusteella hakkuulaitteen paikannus parantaa yksittäisen puun sijaintitarkkuutta 2,93 metriä, päästen yksittäisen puun tasolla 7,86 metristä 4,93 metrin sijaintitarkkuuteen. Säteeltään yhdeksän metrin koealoille lasketun pohjapinta-alan poistuman RMSE tarkentui 35 prosentista 23 prosenttiin, kun koealalta hakatut puut paikannettiin hakkuukoneen sijasta hakkuulaitteen sijaintiin. Kun hakkuulaitteen paikannuksen avulla kerätyistä puutiedoista muodostettiin opetuskoealoja, niin aluepohjaisella menetelmällä ennustettujen leimikkokohtaisen pohjapinta-alan, runkoluvun, kokonaistilavuuden ja tukkitilavuuden RMSE tarkentui 4,1 – 18,1 prosenttiyksikköä kun koealojen pinta-ala oli 254 tai 509 neliömetriä ja verrattavana olivat ennusteet, jotka perustuivat koealoihin, joissa yksittäiset puut oli paikannettu hakkuukoneen sijaintiin. Koealakoon ollessa 763 neliömetriä tai enemmän ei paikannusmenetelmällä enää ollut vaikutusta puustotulkinnan tarkkuuteen. Hakkuukoneaineistoon muodostettujen koealojen käyttö yhdessä tavallisten maastokoealojen kanssa tuotti kaikilla puustotunnuksilla pienimmän RMSE:n, mutta ero pelkkien maastokoealojen avulla tuotettuun tarkimpaan ennusteeseen oli vähäinen. Tarkimmat pohjapinta-alan, runkoluvun, tilavuuden ja tukkitilavuuden ennusteet saatiin joko hakkuukoneen sijaintia ja yli 509 neliömetrin koealakokoa käyttämällä tai hakkuulaitteen sijaintia ja alle 763 neliömetrin koealakokoa käyttämällä. Keskipituuden ja keskiläpimitan tarkimmat ennusteet saatiin hakkuukoneen sijaintia käyttämällä, mutta koealakoon vaikutus oli vähäinen. Hakkuukonekoealojen käyttö aluepohjaisen puustotulkinnan opetusaineistona tuotti kuviotasolla yliarvioita pohjapinta-alan, runkoluvun, tilavuuden ja tukkitilavuuden ennusteisiin, mutta hakkuulaitteen paikannus pienensi ennusteiden yliarviota 254 ja 509 neliömetrin koealakoilla.
  • Mustola, Marjo (Helsingin yliopisto, 2021)
    The loss of forest biodiversity is a global issue. In Finland, there are many measures aiming at preserving the forest biodiversity, for example, protection of the forest habitats defined in the Forest Act and in the Nature Conservation Act; protection of threatened species; the nature management methods in commercially utilized forests and the forest certification. Information of forest resources is collected mainly by remote sensing methods, but also field inventories are carried out, especially when preservation of the areas of high biodiversity value needs to be verified and monitored. Metsäteho Oy has developed an automated method for delineating harvested stands based on harvester location data. The method can be a beneficial tool for providing up-to-date forest resource information. The objective of this study was to research, if harvester location data can be utilized in verifying preservation of areas of high biodiversity value, and in recognizing potential areas, and how that could be implemented in practice. The harvester data used in this study was collected from geographically diverse areas in Finland, and the data contains stem-wise coordinates of harvester while cutting the tree. The delineations of operated areas were generated from harvester location data using the automated method developed by Metsäteho Oy. After the stand delineation was generated, the automated method was utilized in recognizing non-harvested areas left inside and between the harvested stands. Both harvested stands and non-harvested areas were compared to open forest data (including the data of the protected habitats according to the Forest Act and METSO Programme; other protected areas; habitats of threatened species; the Topographic database) using spatial data analysis. The aim was to investigate, if the known areas of high biodiversity value were delimited outside of the harvested stands or if they were left non-harvested within the harvested stand area. In addition, the aim was to research why the automatically recognized non-harvested areas were left without harvesting, and if the non-harvested areas could be potential areas of high biodiversity value. After the spatial data analysis was completed, also field surveys were carried out. Based on the spatial data analysis and the field surveys, the known areas of high biodiversity value were mainly delimited out-side of the harvested stands. The cases in which they were left without harvesting within the harvested stands, were possible to recognize through spatial data analysis. According to the spatial data analysis, part of the automatically recognized non-harvested areas were potential areas of high biodiversity value. Recognized non-harvested areas can be utilized also in recognizing retention tree groups with certain limitations. According to the results, the recognition method for high biodiversity value areas, based on harvester location data, can be utilized when verifying preservation of the high biodiversity value areas, and also other areas that are recorded in spatial data. Based on the observations of this study, it is possible to develop an automated recognition method for high biodiversity value areas, when spatial analysis of datasets in vector format is automated. The positioning accuracy of harvester and the automated method for delineating harvested stands are still causing some challenges when interpreting the results. Also, timeliness and accuracy of the available data of high biodiversity value areas affect on the results of the automated method. To combine different data sources effectively, a data platform is needed in order to use the automated method fluently. The recognition method for high biodiversity value areas can be utilized, for example, when reporting the quality of harvesting work. In addition, the method can be utilized in targeting and minimizing the amount of field inventories when verifying new areas of high biodiversity value. The method enables collecting information and automated monitoring of how the nature management has been integrated into forest management operations in practice. That information contributes to the utilization of forests in economically and ecologically sustainable way.
  • Hakala, Mikko (Helsingin yliopisto, 2021)
    Up-to-date forest inventory benefits the entire forest industry, all the way from forest owners to buyers of raw wood. The forest inventory gathered through remote sensing data and field sample plots by both National Forest Inventory (NFI) and Finnish Forest Centre supports large-scale strategic planning of forestry management and creates a foundation for forest planning as well as up-to-date forest inventory. Operative planning and up-to-date forest inventory also require information about recent cuttings. Finnish Forest Centre has deemed it necessary to develop tools to monitor the realized cuttings on an annual basis. The aim is that data from annual forest operations and cuttings could be transferred into updated forest inventory as soon as possible. The main focus of this thesis was on a method developed by Metsäteho Oy, whereby a stand delineation is automatically created for each forest stand based on harvester data (Melkas ym. 2020). Stand delineation carried out on the basis of stem-specific harvester location data would enable to constantly update the forest inventory in conjunction with logging operations. Stand delineation is an important information because stand area is routinely used as a coefficient in the estimation of stand-specific logging accumulation (Belbo & Talbot 2020). The harvester data was gathered between December 2017 and June 2018 and comprised approximately 3,000 harvested objects and 5,316,214 locations all over continental Finland. The stem-spesific location data recorded by harvester is used in automated stand delineation. Using triangulation, the location data of the stands was combined into a network, and a buffer zone was created for the resulting polygon to reduce the contribution of errors in GNSS navigation while also reflecting the reach of the harvester boom. The use of harvester location data also made it possible to automatically create a strip road network, which in turn allowed to calculate stand-specific strip road variables. Compared to aerial photography references, automated delineation yielded reliable stand delineations when carried out with three most common logging methods and when the stand area was at least .75 hectares. The automated stand was on average three per cent larger than the reference stand manually created from digital aerial photographs. Compared to reference stands, the relative areas of the automated stands were as follows: 1.044 for the first thinning; 1.020 for later thinnings; 1.034 for clear cutting; and 1.031 for all of these harvesting methods combined. There was little variation between the various harvesting methods, and the correlation between automated stand areas and references increased with the size of the stand. For stands with an area more than 1.5 hectares the relative difference in areas was, on average, only around one per cent. Another aim was the validation of automated strip road calculation. On the basis of harvester locations, a strip road network was created, where the to-and-fro movement of the harvester was ignored. Next, the automatically created strip road network was used to calculate the average spacings between strip roads (in metres) and strip road density (in metres/hectare). This was done comprehensively for each stand. In addition, the strip road variables were calculated by emulating sample plot measurements carried out by the Finnish Forest Centre in the evaluation of the quality of harvesting sites objects. Both results were realistic when compared to best practices in forest management. On average, the spacing between strip roads in thinning areas was 20.7 metres and 17.1 metres in clear cuttings. To sum up, there was a reliable correlation between automated stand delineation and reference stands both in terms of area and location; thus, it would be viable to integrate the automatically delineated stands as part of reliable and up-to-date forest inventory. The results of strip road calculation are applicable to validate the implementation of the recommendations set for strip road networks.