When it is detected that a solid immersion lens comes into contact with the semiconductor device, the lens is caused to vibrate by a vibration generator unit. Next, a reflected light image from the lens is input to calculate a reflected light quantity of the reflected light image, and it is judged whether a ratio of the reflected light quantity to an incident light quantity is not greater than a threshold value. When the ratio is greater than the threshold value, it is judged that optical close contact between the lens and the semiconductor device is not achieved, and the lens is again caused to vibrate. When the ratio is not greater than the threshold value, it is judged that optical close contact between the lens and the semiconductor device is achieved, and an observed image of the semiconductor device is acquired.