A team of biologists from The University of Adelaide in Australia has developed an innovative scanning device capable of accurately measuring the potency of cannabis plants prior to harvest. This technology is particularly crucial for medical cannabis growers who must comply with strict regulations regarding the levels of Tetrahydrocannabinol (THC), the psychoactive compound responsible for the intoxicating effects of cannabis.
Understanding the THC content is essential not only for medical growers but also for industrial hemp farmers, whose crops must adhere to specific THC limits. “The capacity to predict cannabinoid profiles weeks before harvest has significant implications for cannabis production, enabling growers and breeders to enhance product quality, reduce costs, and ensure regulatory compliance,” said Dr. Aaron Phillips, the lead researcher of the study published in the journal Industrial Crops and Products.
Revolutionizing Cannabis Cultivation
The scanner allows growers to identify plants predicted to have optimal cannabinoid content, thus streamlining their cultivation efforts. This prevents the waste of resources on lower-quality plants and aids in determining the best times for harvesting, thereby maximizing yields while minimizing the overall growth cycle duration. Additionally, the technology can assist researchers in classifying and differentiating cannabis cultivars, which is valuable for breeders selecting diverse parent plants early in the development process.
To facilitate this process, the research team developed a leaf-scanning method that utilizes intact fan leaves. This approach eliminates the need for cutting samples and performing labor-intensive laboratory tests, such as high-performance liquid chromatography (HPLC) or gas chromatography coupled with mass spectrometry (GC-MS), which require hazardous chemicals. Instead, the scanner employs a technique known as fan leaf hyperspectral reflectance (FLHR), taking measurements across the plant’s canopy during both early and late flowering stages.
The scanner operates using specialized broadband halogen lighting and a spectroradiometer, which measures the wavelengths of light reflected from the leaves. This method enables the device to analyze the biochemical composition of the leaves without the need for invasive procedures. By capturing data across 2,151 wavelength bands from a specific area on a leaf, the researchers can generate reliable predictions of the final cannabinoid content in the mature plant.
Advanced Technology Meets Machine Learning
The research team has integrated machine learning models into their scanning technology, allowing for the detection of patterns within the spectral data that correlate with desirable cannabinoid concentrations. The machine learning model is trained using the spectral profiles of the leaves, along with actual cannabinoid concentrations observed in the plant’s flowers. To validate the model’s accuracy, the study employed a “leave-one-out” validation scheme, where the model was trained on data from nearly all plants in the experiment, testing it on the one plant it had never encountered. This rigorous process was repeated for each of the 70 plants involved in the study to ensure reliable performance under real-world conditions.
Looking ahead, the researchers plan to expand their technology to include additional cannabis genotypes and explore the earliest points in the growth cycle at which cannabinoid content can be accurately predicted. They are also collaborating with the German spectral sensing firm Compolytics to develop a compact device that resembles a supermarket barcode scanner, which will enhance the usability of their FLHR system.
Dr. Phillips noted that a future goal is to test their approach using drones to scan fields of hemp, enabling the identification of plants that exceed legal THC thresholds. This advancement could significantly streamline the management of cannabis crops, ensuring compliance with regulations while optimizing production.
