Search

A labor tracking and yield mapping system for manual citrus harvesting was developed and evaluated in two commercial harvesting operations in Florida. The system consisted of button recognition and barcode units, integrated with a differential global positioning system (DGPS) unit mounted on a truck, which was used to handle fruit collection containers (tub or bin) in the citrus grove. When the truck operator stopped to load each full tub or bin, the button and barcode of the harvester who filled the container was activated. This action recorded the number or name assigned to the harvester and the DGPS location of the container. The button unit was more reliable and was designed with an audible (alarm) and a visible (light) feedback to provide the truck operator verification of the recorded information. The system tallied the number of containers each harvester had harvested for the day and allowed the entry of other information pertinent to the harvesting operation. The DGPS data were used to map the yields with ArcView software. Field tests showed the labor tracking system was reliable and user friendly.

End of Project Report The mapping of within-field crop yield variations offers a basis for the adjustment of input application rates to match yield potential. The aim of this project was to examine available yield mapping components and systems, and to assess the benefits that could accrue from yield mapping. The work also included site investigations of possible causes of yield reduction, and an evaluation of the potential increase in profitability from variable-rate N application. Yield mapping requires the continuous measurement of harvester position and grain yield. Accurate positioning was achieved with systems which received a remote correction signal via communications satellite. A grain flow meter in the clean grain elevator gave low errors in conditions where the deflector plate could be kept clean. The production of useful yield maps was sometimes limited by rainy weather leading to wet crops, lodging, and big variations in harvest conditions within the field. Difficulties are also presented by small, irregularly-shaped fields. The level of within-field variation in cereal yields (25% of the field area at least 21% below the mean) suggests that there is scope for identifying low-yielding areas and either taking corrective action or applying reduced inputs. An analysis of the results of N fertiliser trials suggests a potential gain from about £20/ha in sugar beet to less than £10/ha in spring barley if N application rate could be varied to apply the optimum amount to each area within a field. All the costs of application map development and variable-rate application would need to fall below these gains to justify the use of the technology. To allow further use to be made of yield maps, it is desirable that the factors responsible for yield reduction be identified, in particular those which recur each year. In the present study, soil chemical analysis and compaction affected yield in only a small proportion of cases. Differences between maps from successive years have been small. It is concluded that the immediate financial benefits from the use of this technology will be small. However, in a future where the recording, justification and minimisation of inputs such as pesticides and fertiliser will grow in significance, and where the cost of the equipment and software will continue to fall, yield mapping and variable-rate input application will find a useful role.

This work demonstrates the utility of sophisticated simulation tools in aiding agribusiness managers’ decision making. The system dynamics model developed here provides insight into the use of such models to evaluate potential adoption rates and diffusion patterns of yield mapping and monitoring technologies. The model allows for comparative analyses of the possible effects of different profit assumptions on adoption and diffusion.

Die Getreideernte in den Industrieländern wird heute ausschließlich mit dem Mähdrescher durchgeführt. Durch eine gleichzeitige Ertragskartierung werden die Voraussetzungen für eine teilflächenspezifische Feldbewirtschaftung geschaffen. Aus den geokartierten Ertragsdaten lassen sich Rückschlüsse über kleinräumige Unterscheide in den Wachstumsbedingungen innerhalb eines Feldes ziehen. In der vorliegenden Arbeit werden die theoretischen Grundlagen zur Ertragskartierung und ein Verfahren zum Betrieb des Mähdreschers mit geokodierten Teilflächendaten entwickelt und untersucht. Zu deren Realisierung werden Module für die Erfassung, Übertragung, Verarbeitung und Umsetzung geokodierter Teilflächendaten benötigt.

Build and test the Field Grid Sense system for yield mapping and machine control during harvesting. Secondly, use Field Grid Sense with chemical application equipment to demonstrate a workable in-field system. More specifically, the operation of the patented hardware/software Field Grid Sense (FGS) system will be tested in crop harvesting to demonstrate the system`s utility and to analyze the flexibility of operation under true field conditions. Additionally, FGS will again be used with chemical application equipment - equipment that needs modification to correct one or two slight shortcomings. This action will create improved systems and establish the worthiness, efficiency and necessity of chemical application equipment that is controlled and directed via the FGS package.

The operation of the patented hardware/software Field Grid Sense (FGS) system is being tested in crop harvesting in order to demonstrate the systems`s utility and to analyze the flexibility of operation under true field conditions. FGS is also being used with chemical application equipment. This action will create improved systems and establish the worthiness, efficiency and necessity of chemical application equipment that is controlled and directed via the FGS package.

This is a quarterly progress report describing recent efforts aimed at building and testing the Field Grid Sense (FGS) system for yield mapping and machine control during harvesting. Secondly, the aim is to use Field Grid Sense with chemical application equipment to demonstrate a workable in-field system. More specifically, the operation of the patented hardware/software FGS system will be tested in crop harvesting to demonstrate the systems` utility and to analyze the flexibility of operation under true field conditions. Additionally, FGS will again be used with chemical application equipment -- equipment that needs modification to correct one or two slight shortcomings. This action will create improved systems and establish the worthiness, efficiency and necessity of chemical application equipment that is controlled and directed via the FGS package.

Many researchers, notably Schueller (1992) or Stafford et al. (1994), are working on the variability of cultivation (especially cereals) and inputs. In an agricultural parcel (supposed to be homogeneous) yield levels vary considerably from one place to another one. Variations from 1 to 4 are usual. The main measurement to represent this variation is yield mapping. This mapping is carried out at harvest time. This seems to be too late (no loop back to the corresponding cultivation is possible). Furthermore, some distortion may appear in this measure because of natural events (beating down of corn) or artificial events (harvest machinery tuning, ...). The object of the present project is to find an early indicator of the cultivation variability. This indicator should be correlated with yield and non destructive for wheatears. The crop used in the present study is wheat. The authors chose the number of wheatears as their indicator, and tried to put into place an automatic system for counting wheatears.

Despite considerable interest by Australian farmers in precision agriculture (PA), its uptake has been low. Analysis of the possible financial benefits of alternative management options that are based on the underlying patterns of observed spatial and temporal yield variability in a paddock could increase farmer confidence in adopting PA. The cost and difficulty in collecting harvester yield maps have meant that spatial yield data are generally not available in Australia. This study proposes a simple, economical and easy to use approach to generate simulated yield maps by using paddock-specific relationships between satellite normalised difference vegetation index (NDVI) and the farmer’s average paddock yield records. The concept behind the approach is illustrated using a limited dataset. For each of 12 paddocks in a property where a farmer’s paddock-level yield data were available for 3–5 years, the paddock-level yields showed a close to linear relationship with paddock-level NDVI across seasons. This estimated linear relationship for each paddock was used to simulate mean yields for the paddock at the subpaddock level at which NDVI data were available. For one paddock of 167 ha, for which 4 years of harvester yield data and 6 years of NDVI data were available, the map of simulated mean yield was compared with the map of harvester mean yield. The difference between the two maps, expressed as percentage deviation from the observed mean yield, was

Fertilizing in future will be based on local yield detection, local soil and weather conditions and on use of expert knowledge systems. Position detection will play the key role in a closed loop system for data collection and process control. Engineers have to solve the challenge of the nineties to develop and to implement reliable sensors, actuators and control equipments for practical use. High sophisticated technique will then be able to save costs and to protect environment.

Aims: The aims of this work were to see whether the traditional regionallybased view of terroir is supported by our new ability to use the tools of Precision Viticulture to acquire detailed measures of vineyard productivity, soil attributes and topography at high spatial resolution.Methods and Results: A range of sources of spatial data (yield mapping, remote sensing, digital elevation models), along with data derived from hand sampling of vines were used to investigate within-vineyard variability in vineyards in the Sunraysia and Padthaway regions of Australia. Zones of characteristic performance were identified within these vineyards. Sensory analysis of fruit and wines derived from these zones confirm that contrasting wines may derive from different areas within the same, uniformly-managed vineyard.Conclusions: The performance of vineyards is variable whether yield, fruit quality, wine quality, wine style or value is the measure of interest. The tools of Precision Viticulture enable both growers, winemakers and researchers to see that within vineyards deemed characteristic of a region, terroir is spatially variable at the within-vineyard scale.Significance and impact of the study: Whilst precision viticulture raises questions about the utility of the concept of terroir at regional scales, it has much to offer in promoting robust understanding of the impacts of soil and land attributes on grape and wine production, and thus, how management practices might be modified to gain greater control over fruit and wine quality. Accordingly, at least some of the elements of terroir can be considered to be manageable.

Grain yield maps must accurately display general yield patterns as well as details of local yield variation. Different geostatistical procedures for creating interpolated yield maps by integrating yield data with remotely sensed vegetation indices (VI) were evaluated. Yield monitor data and a multispectral satellite image at 4-m spatial resolution were collected in an irrigated maize (Zea mays L.) field and a rainfed soybean [Glycine max. (L.) Merr.] field. Ordinary kriging (OK), cokriging (CK), simple kriging with varying local means (SKLM), and kriging with external drift (KED) were compared using grain yield as primary variable and three different VI as secondary variables. At both sites, SKLM performed best in terms of the precision of grain yield maps and maps that depicted true yield patterns. Utilizing the most suitable vegetation index at each site and SKLM as interpolation method, the root mean squared error of yield predictions was increased by nearly 20% over OK. Methods such as KED or CK resulted in only small improvement of yield maps over those obtained with OK. Utilizing the satellite image decreased errors associated with yield monitor operations and allowed better prediction in areas where no reliable yield measurements were available. Due to the robust nature of modeling the relationship between primary and one or more secondary variables, the SKLM method has considerable potential for use in commercial precision-farming software. Future efforts on improving yield mapping should concentrate on obtaining improved yield monitor data and imagery and developing yield-sensitive VI for high-yielding crops.

Charles Sturt University has operated an airborne multispectral imaging system as a research support and management tool over south-eastern Australian crops since 1994. Our experiences have demonstrated the utility, timeliness and cost-effectiveness of qualitative multispectral imagery for monitoring and managing spatial variability in a range of agricultural crops, yet to date the technology remains underutilised in Australia. Images showing variations in the texture of soils in paddocks are a useful indicator of the location of different soil zones for soil sampling, and can assist in siting of treatment plots within paddocks. Multispectral imagery can be used for a synoptic assessment of early weed pressure in fallow paddocks or seedling crops. Locating variability in crop emergence and, later, canopy vigour and biomass, are all potentially means of undertaking precision farming without the capital investment associated with yield mapping. However, like any remote monitoring tool, follow-up ground-truthing must always be used to establish or confirm the causes of observed variability. The use of the technology as part of a greater data acquisition strategy is recommended.

I will show how yield mapping data, aerial photography data, and other agronomic data (fertility, soil parameters) can be integrated into a Geographical Information System (GIS) and give a “feel” of the value of these tools to look at crop production as a whole. The capability of GIS in handling and displaying several layers of georeferenced data leads naturally to a decisionmaking process quite similar to the one used in traditional photo interpretation of aerial imagery. This approach can be very valuable for farm managers and consultants in crop production.

Over the past few years, grain yield monitors have gained a significant hold in the market place. While the largest share of production agriculture acres are devoted to producing grain crops, high-value crops such as potatoes, tomatoes, sugarbeets, onions, and many others will benefit considerably by application of site-specific technology. Yield mapping is one of the tools that utilizes GPS technology and allows us to visualize our farms as an array of tiny parcels instead of one uniform aggregate. Yield mapping is simple, accurate measurement of yield at precise positions, the data from which is used to give us a visual report card of each parcel in that field. While yield mapping will not provide the entire basis of site-specific agriculture management, it begins to give a picture of how understanding spatial variation will revolutionize management of high-value crop production acres. The tools necessary to make yield measurements are now available. When combined with Differential GPS, the yield map becomes a powerful tool to identify atypical areas in the field. Without DGPS the process of identifying and treating areas within a field individually would be a nearly impossible task, and certainly cost-prohibitive. Identification of the spatial distribution of yield will contribute significantly to a grower's ability to make informed management decisions.

Measurements described in this paper were realized in order to find out whether there is some relationship between mass flow of plant material passing through a parallel plate capacitive sensor and its output signal. This possibility of mass flow determination could be useful for the aim of forage crops or sugar beet yield maps creation. A parallel plate capacitive throughput sensor was designed for that purpose. The capacitive sensor and the whole oscillating circuit were driven at 27 MHz frequency. The laboratory set-up consisted of a conveyer belt, carrying a measured quantity of material into sensor, equipped with an electronic measurement apparatus. Material passed through the sensor between its plates. Laboratory tests were carried out with forage crops and with sugar beet. Resulting coefficients of determination ranged from R2=0.95 to 0.97 for forage crops and around R2=0.98 for sugar beet. It was possible to conclude from our measurement that forage and sugar beet mass flow determination by means of the parallel plate capacitive sensor is a promising way. The results showed a strong linear relationship between the feed rates of plant material passing through the sensor and tested measuring capacitive sensor circuit output signal.

Machine vision based citrus yield mapping and fruit quality inspection system was developed. The system consisted of 3CCD camera, four halogen lamps, a DGPS receiver, an encoder and a laptop computer. A total of 3,653 images were taken during an experiment on the test bench at the Citrus Research and Education Center, Lake Alfred, Florida and 703 images were used for analysis. The system was also tested on a canopy shake and catch harvester at the Lykes grove located in Fort Basinger, FL. A total of 773 images were acquired as well. Fruit weight was measured in 14 test trials of image acquisitions during the test bench experiment as well as in two test trials of image acquisitions during a field trial with a commercial canopy shake and catch harvester. An image processing algorithm that could identify and inspect fruit qualities was developed. The number of fruit and total fruit areas were measured from the sets of color images using the developed algorithm. Finally number of citrus fruit that was found by the image processing algorithm during the test-bench experiment was compared against actual fruit weight. The coefficient of determination, R2, was 0.962 between them. For a validation purpose of the canopy shake and catch harvester experiment, number of fruit was counted manually from a total of 60 images. Human counting was compared with algorithm counting. The coefficient of determination, R2, was 0.891 between actual and estimated number of fruit.

The existence of spatial variability within fields points to considerable benefits if inputs to arable crop production are targeted according to locally determined requirement. The paper discusses yield mapping as a data source on variability and emphasises the need to integrate data on all sources of spatial variability with the management/agronomic expertise of the farmer to develop an application strategy for field inputs. A yield mapping system is described and examples of yield maps from one field in four successive years, together with data on soil spatial variability are presented. An exploratory study was made of appropriate statistical techniques to compare yield maps analytically. Consistency of yield variation in the test field from season to season and correlation with a limited sampling of soil nutrient parameters were low. However, significant correlation between soil series and yield was observed.

This paper evaluates the potential of very high resolution multispectral (Worldview-3) satellite imagery for mapping yield parameters in avocado and macadamia orchards. An evaluation of 18 structural and pigment based vegetation indices (VIs) derived from Worldview-3 imagery identified a positive relationship to nut/ fruit weight (kg/tree) R2>0.69 for macadamia and R2>0.68 for avocado; and nut/ fruit number (per tree) R2>0.6 for macadamia and R2>0.61 for avocado. Using the algorithms derived between the optimal VI and the measured parameter, yield and nut/ fruit number maps were derived for each block. In the absence of a commercial yield monitor, the resulting yield maps offer significant benefit to growers for improving orchard management, harvest scheduling, and forward selling decisions.

Aims: Yield monitors are becoming more common in North America. This research evaluates the precision and accuracy of a retro-fitted, commercially available grape yield monitor mid-season, for crop estimation and crop thinning applications, and at harvest for yield mapping.Methods and Results: Several grape yield monitors were mounted on the discharge conveyor belt of grape harvesters in both commercial and research vineyards in North America. Sensor response was compared to manual measurements at multiple masses, ranging from 20 kg to 28 Mg over the course of three seasons. Measurements were taken during crop thinning and estimation (mid-season) and at harvest. Results showed that the grape yield monitor performance was sufficient to generate good spatial maps of the relative variation in harvest yield and mid-season thinned yield. However, at harvest the sensor showed a shift in response between days of up to ±15%, such that the generation of absolute yield maps required a daily calibration against a known mass. Within a day (single harvest operation) the sensor response did not appear to drift. Mid-season applications required a different calibration to harvest applications.Conclusion: The yield sensor worked well for both mid-season and at harvest operations in North American vineyards but required a daily calibration to avoid drift issues. The mid-season yield calibrations were different between seasons; however, the harvest calibration factor was stable between seasons.Significance and Impact of study: The study showed that a commercial yield monitor with correct calibration was effective at even low fruit flow. This opens the possibility of using a harvest sensor mid-season to mechanically estimate fruit load from small point samples and to map the amount of fruit removed during fruit thinning operations. This will improve the quality of information available to viticulturist to understand fruit and crop load. The commercial yield monitor is suitable for use in North American vineyards.

Service units used in precision agriculture are able to improve processes such as harvesting, sowing, agrochemical application, and manure spreading. This two-part work presents, a path tracking controller based on an algebraic approach for an articulated service unit, suitable for embedded applications, and its implementation to a hierarchical navigation strategy to aid a manual harvesting process. The path tracking controller approach can be scaled to several trailers attached to the service unit. For harvesting, the service unit drives within an olive grove environment following the previously developed path and a trailer is used as a mobile hopper where olives, collected by human labour, are deposited. The service unit also registers and geo-references the amount of olives (mass) collected for the subsequent creation of yield maps. The developed navigation strategy improved the time associated with harvesting olives by approximately 42–45%. The mathematical formulation of the problem, some real time experimental results, the creation of a yield map and the statistical analysis that validated the method are included. Fil: Auat Cheein, Fernando Alfredo. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Universidad Técnica Federico Santa María; Chile Fil: Scaglia, Gustavo Juan Eduardo. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Universidad Nacional de San Juan. Facultad de Ingeniería. Instituto de Ingeniería Química; Argentina Fil: Torres Torriti, Miguel. Pontificia Universidad Católica de Chile; Chile Fil: Guivant, José. University of New South Wales; Australia Fil: Prado, Alvaro Javier. Universidad Técnica Federico Santa María; Chile Fil: Arnò, Jaume. Universidad de Lleida; España Fil: Escolà, Alexandre. Universidad de Lleida; España Fil: Rosell Polo, Joan R.. Universidad de Lleida; España

A machine vision system utilizing color vision was investigated as a means to identify citrus fruits and to estimate yield information of the citrus grove in real-time. Images were acquired for 98 citrus trees in a commercial grove located near Orlando, Florida. The trees were distributed over 48 plots evenly. Images were taken in stationary mode using a machine vision system consisting of a color analog camera, a DGPS receiver, and an encoder. Non-overlapping images of the citrus trees were taken by determining the field of view of the camera and using an encoder to measure the traveled distance to locate the next position for acquiring an image. The threshold of segmentation of the images to recognize citrus fruits was estimated from the pixel distribution in the HSI color plane. A computer vision algorithm to enhance and extract information from the images was developed. The total time for processing an image was 119.5 ms, excluding image acquisition time. The image processing algorithm was tested on 329 validation images and the R2 value between the number of fruits counted by the fruit counting algorithm and the average number of fruits counted manually was 0.79. Images belonging to a same plot were grouped together and the number of fruits estimated by the fruit counting algorithm was summed up to give the number of fruits/plot estimates. Leaving out outliers and incomplete data, the remaining 44 plots were divided into calibration and validation data sets and a model was developed for citrus yield using the calibration data set. The R2 value between the number of fruits/plot counted by the yield prediction model and the number of fruits/plot counted by hand harvesting for the validation data set was 0.53.

Kroulik M., Masek J., Kviz Z., Jost B., Prosek V. , 2011. Sensors connection for yield determination on round balers with variable chamber . Res. Agr. Eng., 57: 51–55. The main aim of this article is to present a technical solution for straw and forage yield mapping when us ing round balers with variable chamber for harvest. The yield measurement is based on monitoring of instantaneous position of a tension roller mechanism for press chamber circular belt. Wheat straw was har vested – baled, during our trial measurements. The acreage of the trial field was 12 ha. Calibration of the measuring system showed a strong dependence of the tension roller position on the amount of pressed straw (R 2 = 0.99). Geostatistical evaluation confirms a spatial relationship of measured data sets with a moderate spatial dependence. Finally, yield map of straw was created.

Crop yield data are a key component of precision agriculture and are critical for both development and evaluation of precision management strategies. Ideally, software that generates grain yield maps from raw yield monitor data should automatically correct errors associated with machine and operating characteristics. Perhaps the most basic correction required is to properly compensate for the time lag (or position lag) between the cutting of the crop from the field and the grain flow measurement by the flow sensor in the combine. Past research has suggested several approaches to automatically determine delay time, but for various reasons these have not been implemented in mapping software. In this article, we present a new, computationally efficient method that can accurately determine delay time for individual fields using the image processing method of phase correlation. The phase correlation delay identification (PCDI) method was evaluated using a number of yield maps with varying degrees of harvest complexity, and results were compared to a geostatistical method. The PCDI method produced accurate estimates of delay time in approximately 90% of test datasets and provided a way to evaluate the reliability of the estimate. Additionally, the PCDI method was more computationally efficient than previous methods. Results of this study will increase the feasibility of including automatic delay time compensation in yield mapping software.

Detection of immature green citrus fruit is important during the early life cycle of citrus fruit. It allows growers to manage citrus groves more efficiently and maximize yields by identifying expected fruit yields well in advance before harvesting. It also helps the growers prepare harvesting equipment and pickers for the harvesting operation. A novel technique was developed for detecting immature green citrus fruit from an outdoor color image and counting number of fruits. This technique is unique in that it is the first known attempt towards exploring it on green citrus fruits. A set of 71 images containing immature green citrus fruit was acquired in an experimental citrus grove at the University of Florida, Gainesville, Florida, USA. An algorithm was developed using a set of 11 training images by calculating the fast Fourier transform leakage values for fruit and leaves. A threshold value was obtained by comparing the percent leakage of fruit and other objects. The algorithm was tested on a set of 60 validation images. The correct total fruit count for a validation set came out to be 120, whereas the actual number of fruit was 146. The overall correct detection rate was 82.2 %. The proposed algorithm can be further improved to help growers manage their grove more efficiently.

Hyperspectral imagery deals with large volumes of data due to hundreds of spectral bands used in the images. Although hyperspectral images with higher spectral resolution usually carry more information, the processing of the images requires a significant amount of memory and is usually very slow. In addition, hyperspectral images contain considerable amount of redundant information, which does not help or even hinder the algorithm in making the correct decision. Band selection is an approach to both reduce the dimensionality of hyperspectral images and save calculation time for further applications, such as detection of blueberry fruit with different maturity stages. Hyperspectral images of blueberry fruit were taken in a commercial blueberry field. Mature fruit, intermediate fruit, young fruit and background were the four classes to be studied. A supervised band selection method was proposed using Kullback-Leibler divergence (KLD). Wider bands were made by combining 20 hyperspectral bands so that the selected bands could be used in a blueberry yield mapping system using a lower-cost multispectral camera. Based on the analysis, six combined bands were selected: 543.1 – 572.6 nm, 627.4 – 658.8 nm, 663.6 – 695.2 nm, 725.4 – 757.4 nm, 773.5 – 805.6 nm and 838 – 870.5 nm. The test result showed that the proposed band selection method worked well for the task of blueberry growth stages detection.

Sensors, communication systems and geo-reference units are required to achieve an optimized management of agricultural inputs with respect to the economic and environmental aspects of olive groves. In this study, three commercial olive harvesters were tracked in Spain and Chile using remote and autonomous equipment to determine their time efficiency and field capacity. An experimental methodology for analyzing the data to determine the field capacity and efficiency is proposed, which, along with a conventional methodology, was used to analyze the data to determine field capacity and efficiency. The results of both methodologies are compared to validate the suitability of the experimental methodology. Furthermore, a yield monitor was developed and evaluate using one of the tested olive harvesters. The results show that yield monitoring of olives is possible, but further research is needed to archieve a more reliable methodology.

One hundred and eighty eight blueberry fruit and leaf samples were obtained from a commercial blueberry field in Waldo, Florida in June, 2010. Spectral reflectance was measured in the ultraviolet (UV), visible and near-infrared (NIR) ranges (200 nm-2500 nm) with an increment of 1 nm for each sample. Five different categories (mature fruit, intermediate fruit, immature fruit, light-green leaf and dark-green leaf) were used for classification model construction and validation. Significant differences in reflectance among the five categories were found in the visible and NIR region. Especially, the mature fruit had much lower reflectance in both regions, which shows great potential for distinguishing mature fruit from other categories. Based on the spectral characteristics of each category, fourteen normalized vegetation indices were developed for further statistical analysis to find significant bands for classifying different fruit maturity status as well as leaves. Principal component analysis (PCA), classification regression tree and multinomial logistic regression were conducted to develop prediction models for distinguishing different classes. The multinomial logistic regression model with three independent variables, which are the combinations of reflectance at six wavelengths (500, 525, 550, 575, 680, and 750 nm) performed the best, with prediction accuracy of 100%. The six wavelengths thus can be used for developing an easy-to-use and low cost fruit maturity sensor for a blueberry yield mapping system.