Skip to main content Skip to navigation

Aerial imaging to monitor blueberry shock disease

Volume 11 Issue 11

Dayna Loeffler1, Chris Benedict2, Chakradhar Mattupalli1

1 WSU Mount Vernon Northwestern Washington Research and Extension Center, Mount Vernon, WA 98273

2 WSU Whatcom County Extension

[Funding for this research was provided by the Northwest Agricultural Research Foundation]

 

Problem:

  • Blueberry shock virus (BlShV) is one of the important viruses affecting blueberry crop in the Pacific Northwest.
  • BlShV is pollen-borne and transmitted by honeybees. BlShV infected blueberry plants show blighted flowers and leaves that drop by early summer (Fig. 1). However, regrowth occurs so that by the end of the growing season the affected plants look almost normal bearing less fruit with yield losses. Shock affected plants show symptoms either two or three years in a row, but plants recover and do not show shock symptoms in subsequent years. Current management recommendation is for the disease to run its course.
  • Recently, there has been an emergence of recurring shock symptoms in previously shock-affected orchards. The cause and effects for this phenomenon is yet to be fully understood. Furthermore, there are several questions that need to be addressed such as: how plants showing recurring shock symptoms are spreading within a field, how are they recovering within a growing season and over multiple seasons, and if symptom expression varies with blueberry cultivars. Understanding disease distribution patterns is ultimately needed by growers for making sound disease management decisions.

Figure 1: Blueberry cv. Chandler plant showing blueberry shock disease symptoms

Objective:

  • To monitor blueberry shock disease progression using aerial imaging.

Field monitoring for diseases can be done manually but aerial images obtained using a small unmanned aircraft systems (colloquially called drones) can perform such tasks in a timely, cost-efficient manner, over larger fields, and across multiple growing seasons.

Approach:

  • A RGB (Red Green Blue) sensor mounted on a drone was used to obtain aerial images of three blueberry fields (‘Reka’, ‘Bluecrop’, and ‘Draper’) with a history of blueberry shock disease. Images were captured at three different time points in all three fields from mid-May to late June (9 total flights). Aerial images were processed using Pix4D software to generate an orthomosaic. Symptomatic plants were then geotagged based on ground truth data (Fig. 2).
  • Ground truthing was performed by assessing about 10,200 plants from all three fields at each sampling date. During the study period (mid-May to late June), plants (n=201) with symptoms resembling blueberry shock disease were identified, GPS coordinates recorded, and leaf samples were collected.
  • BlShV presence in the leaf samples was detected with ELISA (Enzyme Linked ImmunoSorbent Assay).

Figure 2: A representative aerial image from a small unmanned aircraft systems flight taken four weeks post bloom (late June) in ‘Draper’ blueberry field in Whatcom County. The GPS point (orange dot) denotes the location of a bush that tested positive for Blueberry Shock Virus (BlShV).

Preliminary Results:

  • BlShV was detected in all three fields with 86% of the leaf samples testing positive for the virus.
  • The number of new plants showing blueberry shock symptoms and that tested positive for BlShV decreased over time (Fig. 3). In a few cases, a plant that exhibited symptoms tested negative for BlShV.

Figure 3: A bar graph representing plants from three blueberry fields in Whatcom County that tested positive for Blueberry Shock Virus (BlShV).

  • It was far easier to detect symptomatic plants in aerial images on the last assessment date when more foliage was present and symptoms are better expressed than on the first date. The break in the consistent canopy (Fig. 2) was clearer on this later sampling date.

Conclusions & Future Work:

  • We generated baseline information to continue data collection in 2023. Because the location of diseased plants is geolocated at the individual plant level, we can monitor both the reoccurrence of symptoms across years and the within field spread over time.
  • Ground truthing and subsequent laboratory assessment validated aerial imaging data.
  • Utilizing a multi-spectral sensor may allow for the detection of diseased plants earlier in the growing season. In addition, the use of machine learning approaches could decrease the time needed to detect symptomatic plants from aerial images providing the ability to monitor more fields.