1.1.1. Product Description
SlantRange’s population analysis algorithm segments (isolates from the background and other objects in the field) individual plants in aerial multispectral images to calculate two important parameters for early season development: plant population density (stand count) and plant size.
SlantRange’s approach uses a direct counting method, where each plant in a pre-defined grid is counted and sized. The statistics (mean, standard deviation) are then reported for each grid spacing as previously defined. Figure 1 below shows two images of a corn field at approximately V2 growth stage. At left is a near infrared image and on the right is the same image represented in color with individual plant detections shown in bright green.
Figure 1: Near infrared image of V2 corn population (left) and the same image with SLANTRANGE’s plant detection algorithm highlighting individual plants (right)
The accuracy of SlantRange’s plant counting approach is a function of how well the plants can be detected and how well the area of the image can be calculated, and through previous trials has been demonstrated with errors less than 5% under the proper collection and plant growth stage conditions. Best-case collection conditions are when plants are large enough to be easily detected in the imagery yet small enough to not yet encroach on neighbors (typically V2-V3 for corn). Figure 2 below shows an example of a corn plot with high population density contrasted with a plot with low population density.
Figure 2: Examples of high plant population (left) and low plant population (right)
In addition to plant population density, SlantRange algorithms simultaneously measure the nadir-projected size of each plant.
1.1.2. Reporting Method
Mean and standard deviation of plants per unit area (e.g. plants per acre, plants per hectare, plants per square meter) over selected grid size
Mean and standard deviation of plant area (e.g. square cm) over selected grid size
Corn at V2 – V3