Abstract: In order to study the influence of noise on detection capability of an avalanche photodiode(APD), hyperspectral images were used to calculate radiation power and the causes and characteristics of all kinds of noise in an APD detection system were analyzed. On this basis, the factors affecting the detection capability of the system were summarized. The relationship between signal-to-noise ratio and background light power in daytime and the relationship between multiplier factor and temperature were simulated. The spectral curves of hyperspectral images were used to calculate the power of background after atmospheric turbulence, atmospheric molecules and aerosol, and then the relationship between background radiation and APD noise after atmospheric transmission was obtained. The results show that when the bias voltage is 340V and background light is enhanced from 0.07mW to 0.35mW, the effective value of noise is enhanced from 108mW to 150mW. The effective value of APD noise will be enhanced with the enhancement of the power of natural background light under the same bias voltage. When background light is 0.35mW and bias voltage is enhanced from 320V to 340V, the effective value of noise is enhanced from 23mW to 150mW. Under the same background, the effective value of noise will increase with the increase of the bias voltage, and the growth rate will gradually accelerate. Background power in a specific region can be obtained from hyperspectral images. The influence of background radiation on the APD noise characteristics is also clarified.