Abstract: Abstract: Spray volume distribution in the three-dimensional (3D) space of nozzles is an essential interfering factor on spray drift and deposition of pesticide application, particularly on the atomization quality. Uniform distribution of spray can contribute to an obvious enhancement of pesticide efficacy, while reducing overuse and serious environmental contamination. However, the accurate measurement is still lacking in the real-time dynamic 3D distribution of spray volume, due mainly to long time consumption, and cumbersome procedure at present. In this study, a novel measurement for 3D spray volume distribution was developed using light detection and ranging (LIDAR) technology. Seven types of nozzles were tested, including the commonly-used nozzle of hollow cone, anti-drift hollow cone, flat fan, and anti-drift flat fan (HCI4002, TR8002, ITR8002, LU9002, IDK9002, LU12002, and IDK12002) in plant protection. The spray area of the nozzle was scanned using a 16-line laser LIDAR with the laser (Class 1) wavelength of 905 nm and the scanning range was -15°-15°. Specifically, the angular speed of horizontal rotation was 5 Hz, and the emission frequency was 320 Hz. The scanning lasted for 60 s, and all nozzles were tested with 3 replicates. The point cloud data was transferred to the laptop in form of packets in real time. MATLAB 2019b software was used to run the affine matrix and coordinate system transformation after data packet analysis for the droplet coordinates and spatial density. Meanwhile, the real value of spray volume distribution was measured in the spray section of 50 cm below the nozzle. Polyethylene (PE) centrifugal tubes with a volume of 50ml were arranged in a matrix to collect the droplets. Four kinds of fan nozzles were tested by a 5×15 collector matrix, and three kinds of hollow cone nozzles were tested by a 9×9 collector matrix. All nozzles were measured three times, and all tests lasted for 3 min, in order to collect enough droplets for a small weighing error. A neural network with 1 hidden layer (100 hidden neurons) and 1 output layer was used to fit the relationship between the traditional measurement and LIDAR scanning. The ratio between training, validation, and testing set was 70:15:15. The results showed that a high fitting precision was achieved in all seven kinds of nozzles for the correlation coefficient in the training set r≥0.995, validation set r≥0.935, testing set r≥0.877, and the correlation coefficient r≥0.990 for the flat fan nozzles. It proves that the LIDAR scanning can accurately and quantitatively analyze the spray volume distribution. The 3D spatial distribution of spray volume for all 7 nozzles was obtained after the spray area was layered and meshed, then to calculate the droplet density in each grid. A faster and easier procedure was made for the real-time 3D spray volume distribution, compared with the conventional one. The LIDAR technique can also be expected to provide an alternative way for atomization quality detection of sprayers, indoor and field test of spray drift, particularly on a rapid adjustment and online monitoring of operation quality in plant protection machinery in the field.