Abstract: In view of the recognition and rapid measurement of the water surface pollution area, a high-precision measurement scheme of the pollution area of the water surface is proposed to meet the pollution demand of the sea area management and the monitoring of the blind spot, by means of electronic measurement and image discrimination. According to the ranging principle of monocular vision based on machine vision, the UAV aerial photography is used to provide the corresponding parameters and the azimuth information of the polluted area for the measurement model. The image segmentation and stacking technique is used to identify the characteristics of the oil pollution area by artificial delineation boundary. Using the visual image processing method which is suitable for the area measurement, a human-computer interaction for measuring the pollution area of the water surface is designed to realize the real-time or later measurement. The experimental results show that the actual flight control measurement and software processing of UAV provide effective basis for water management. the polluted area is simulated, the accuracy of polluted area measurement is 10^-3 square meters, the measurement relative error is less than 5%, and the average relative error of vertical shooting is less than 1.3%.

Key words: machine vision, area measurement, image processing, human-computer interaction, UAV flight control